Method for treating a cancer patient based on atomic therapeutic indexes and non-radiation therapy

Abstract

The present invention relates to the generation of an Atomic Therapeutic Indicator (ATI) for a test sample by the quantification of manganese; in voxels of a 3D region of the sample, wherein the 3D region is topographically defined by co-ordinates XYZ. The ATI is used to assess the radio-responsiveness i.e. sensitivity or resistance to radiation treatment, of a cancer i.e. a tumor/neoplasm. In a preferred embodiment, the present invention relates to a method of generating the ATI, assessing the radio-responsiveness of a tumor/neoplasm based on the ATI and, based on the assessment, either treating or not treating the tumor with radiation. The present invention also relates to a method of determining if a cancer is likely to reoccur post radiation treatment comprising quantifying the level of manganese in voxels of a 3D region of a test sample from the cancer and determining the frequency of high metallomic regions (HMRs) in the cancer, wherein a high frequency of HMRs is indicative that the cancer is likely to reoccur and a low frequency of HMRs is indicative that the cancer is unlikely to reoccur; and associated methods of treatment. The invention further relates to a method of determining the radio-responsiveness of a melanoma, the method comprising determining the level of melanin in a test sample from the melanoma, wherein the lower the level of melanin the more sensitive the melanoma is to radiation and the higher the level of melanin the more resistant the melanoma is to radiation; and associated methods of treatment.

Claims

1. A method of treating a patient determined to have a cancer resistant to radiation treatment with a cancer treatment not including radiation treatment comprising: (A) identifying a cancer patient who is resistant to radiation treatment by: (i) obtaining a tumour sample from the patient, wherein the tumour sample corresponds to a total sample volume of at least about 100010005 cubic microns; (ii) quantifying the level of manganese in the tumour sample by generating an Atomic Therapeutic Index (ATI), wherein the ATI corresponds to the median level of manganese as related to a pre-defined volume of the tumour sample expressed in calibrated counts per second (CC/S) as measured by laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS); wherein the pre-defined volume corresponds to about 35355 cubic microns; and (iii) comparing the generated ATI to a predetermined ATI threshold, wherein the predetermined ATI threshold corresponds to about 2000 CC/S for a volume of about 35355 cubic microns; wherein the cancer is determined to be resistant to radiation treatment when the ATI of the tumour sample is higher than the predetermined ATI threshold; and (B) administering a cancer treatment to the patient with cancer determined to be resistant to radiation treatment, wherein the cancer treatment is not radiation therapy, and wherein the cancer is prostate cancer, breast cancer, seminoma, lymphoma, small cell lung cancer, brain cancer, mesothelioma, or melanoma.

2. The method of claim 1, wherein said cancer treatment is chemotherapy, immunotherapy, hormone therapy, or surgery.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1: Photographic image of a standard 5 micron, H&E stained tissue section from a formalin-fixed paraffin-embedded (FFPE) block from a stage Ib malignant melanoma of the neck of a 48 year old male, with the TNM staging criteria of T2aN0M0, where T2 represents invasion of the muscularis propria; N0 represents no lymph node metastasis, and M0 represents no distant metastasis. Note that the top right-hand side of this section consists of pale amelanotic tumour cells, while the darker regions on the left-hand side and bottom of the Figure are composed of cells containing darkly staining melanin, and with even larger patches of melanin, both intra- and extra-cellular, as black spots in the lower part of the image.

(2) FIG. 2: Illustration of three different manganese-based binding entities. It should be noted that a single .sup.55Mn.sup.2+ does not bind all three entities at once (this figure is for illustrative purposes onlytaken from Slade and Radman, Microbiology and Molecular Biology Reviews, 75, 133-191, 2011.)

(3) FIG. 3A. Photographic image of an H&E stained 5 micron tissue section from a block of a normal human cortex from a 50 year old human female showing the characteristics of normal cortical tissue. It is illustrative of the position of a three track ablation (taken from the larger multi-track tissue ablation) that was carried out on an adjacent unstained section from the same block, but on a different slide. The H&E stained area shown in the Figure corresponds to the equivalent unstained area analysed by LA-ICP-MS on a different slide. FIG. 3B. The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn, and .sup.56Fe are shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 70 voxels per track. FIG. 3C-FIG. 3E. Graphical representation of the raw numerical values (from FIG. 3B), wherein each track consisted of 70 voxels of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows.

(4) FIG. 4A. Photographic image of an H&E stained 5 micron tissue section from a block of a 19 year old human female with a brain neoplasm classified as glioblastoma multiforme. The H&E stained area shown in the Figure corresponds to the equivalent unstained area analysed by LA-ICP-MS on a different slide. FIG. 4B. The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn, and .sup.56Fe are shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 68 voxels per track. FIG. 4C-FIG. 4E. Graphical representation of the raw numerical values (from FIG. 4B) of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows from the human female with glioblastoma multiforme. The Y axis shows calibrated counts per second (CC/S) of that metal ion, while the X axis shows the voxel number of the three ablation tracks.

(5) FIG. 5A. Photographic image of an H&E stained 5 micron tissue section from a different region of the same block of the 19 year old human female with glioblastoma multiforme shown in FIG. 4A-FIG. 4E. FIG. 5B. The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe are shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 69 voxels per track. FIG. 5C-FIG. 5E. Graphical representation of raw numerical values (from FIG. 5B) of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows from the human female with glioblastoma multiforme. The Y axis shows calibrated counts per second (CC/S) of that metal ion, while the X axis shows the voxel number of the three ablation tracks.

(6) FIG. 6A. Photographic image of an H&E stained 5 micron tissue section from a 60 year old human male with malignant mesothelioma. FIG. 6B. The calibrated signal of each of four metals .sup.55Mn, .sup.66Zn, .sup.56Fe and .sup.63Cu are shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 67 voxels per track. FIG. 6C-FIG. 6F. Graphical representation of the raw numerical values (from FIG. 6B) of calibrated counts per second for the four metals, .sup.55Mn, .sup.66Zn, .sup.56Fe and .sup.63Cu in adjacent voxels in each of three ablation rows from the 60 year old human male with malignant mesothelioma. The Y axis shows calibrated counts per second (CC/S) of that metal ion, while the X axis shows the voxel number of the three ablation tracks.

(7) FIG. 7A. Photographic image of an H&E stained 5 micron tissue section from a 50 year old male with malignant melanoma of the esophagus. FIG. 7B. The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe are shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 68 voxels per track. FIG. 7C-FIG. 7E. Graphical representation of the raw numerical values (from FIG. 7B) of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows from the 50 year old human male with malignant melanoma of the esophagus. The Y axis shows calibrated counts per second (CC/S) of that metal ion, while the X axis shows the voxel number of the three ablation tracks.

(8) FIG. 8A. Photographic image of an H&E stained 5 micron tissue section from a 57 year old male with diffuse B-cell lymphoma. FIG. 8B. The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe are shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 79 voxels per track. FIG. 8C-FIG. 8E. Graphical representation of the raw numerical values (from FIG. 8B) of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows from the 57 year old male with diffuse B-cell lymphoma in the testis. The Y axis shows calibrated counts per second (CC/S) of that metal ion, while the X axis shows the voxel number of the three ablation tracks.

(9) FIG. 9A. Photographic image of a representative H&E stained 5 micron tissue section from a 38 year old male with a small cell undifferentiated malignant carcinoma of the lung. FIG. 9B. The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe are shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 65 voxels per track. FIG. 9C-FIG. 9E. Graphical representation of the raw numerical values (from FIG. 9B) of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows from the 38 year old male with a small cell undifferentiated malignant carcinoma of the lung. The Y axis shows calibrated counts per second (CC/S) of that metal ion, while the X axis shows the voxel number of the three ablation tracks.

(10) FIG. 10A. Photographic image of a representative H&E stained 5 micron tissue section from a 52 year old male with seminoma. FIG. 10B. The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe are shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 79 voxels per track. FIG. 10C-FIG. 10E. Graphical representation of the raw numerical values (from FIG. 10B) of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows from the 52 year old male with seminoma. The Y axis shows calibrated counts per second (CC/S) of that metal ion, while the X axis shows the voxel number of the three ablation tracks.

(11) FIG. 11A to FIG. 11F. Graphical representation of total median manganese contents, (expressed as calibrated counts per second, CC/S of .sup.55Mn), in tumours from patients with seminoma (55) (FIG. 11A), lymphoma (10) (FIG. 11B), small cell carcinoma of the lung (20) (FIG. 11C), melanoma (64) (FIG. 11D), brain (glioblastoma and astrocytoma) (25) (FIG. 11E), and mesothelioma (10) (FIG. 11F). Each square represents a single patient except in the melanomas, where two of the 64 patients each had two major lineages within their tumours, and each lineage was represented separately in the histogram with a square.

(12) FIG. 12: Two-dimensional representation of laser ablation tracks across the top part of the tissue section shown in FIG. 1 of one of the melanoma patients whose tumour had two major lineages within it. Scale; right hand side, (dark grey), 0-15,000; middle, (whitish), 15,000-45,000; left hand side, (pale grey), 45,000-150,000 calibrated counts per second.

(13) FIG. 13A to FIG. 13D. Graphical representation of an analysis of all four metals, with FIG. 13A, FIG. 13B, FIG. 13C, and FIG. 13D showing .sup.55Mn, .sup.66Zn, .sup.56Fe and .sup.63Cu respectively of the tumour shown in FIG. 1.

(14) FIG. 14A. Photographic image of an H&E stained tissue section of a melanoma from the left forefinger of a 54 year old male: stage IV, T4N0M1. FIG. 14B. Two-dimensional representation of ablation tracks across the tumour. Scale; pale grey (main bulk .sup.55Mn value of the tumour) 0-2,000 CC/S; whitish flat areas, generically termed herein as High Metallomic Regions (HMRs), which in this particular example correspond to >6,500 calibrated counts per second.

(15) FIG. 15A. Photographic image of a standard 5 micron H&E stained tissue section from a formalin-fixed paraffin-embedded block from a melanoma on the left forefinger of the 54 year old male analyzed in FIG. 14-FIG. 14B illustrating the tissue and cellular morphology. FIG. 15B. The metallomic content of the sample shown in A was measured. Part B is a two-dimensional representation of .sup.55Mn levels in individual voxels of this section after LA-ICP-MS. Two clusters of contiguous voxels of high .sup.55Mn content, denoted HMRs (.sup.55Mn) and shown as black areas, become visible after visual inspection of the data of the voxel matrix. FIG. 15C. Two-dimensional representation of .sup.55Mn levels in individual voxels of the section of A and B were further analyzed by applying a threshold of 2 the median value per voxel of the whole area, (plus the machine background value). Black squares show those voxels that are above this particular (2 median) threshold. FIG. 15D. Two dimensional representation of .sup.55Mn levels in each individual voxels of the section of A and B were further analyzed by applying a threshold of the median value per voxel of the whole area, plus the standard deviation (St.Dev.) of the voxel values of the whole area, plus the machine background value. Black squares show those voxels that are above this particular (median+St.Dev) threshold.

(16) FIG. 16A: Photographic image of the same standard 5 micron H&E stained tissue section from the tumour of the melanoma patient from FIG. 15A. FIG. 16B: Two-dimensional reperesentation of the same voxel matrix of .sup.55Mn values shown in FIG. 15C showing the two clusters of contiguous voxels that constitute the HMRs (.sup.55Mn). FIG. 16C: .sup.66Zn voxel values in the identical voxels to those shown in the FIG. 16B and using the same threshold criterion (2 median) for .sup.66Zn as for .sup.55Mn.

(17) FIG. 17A to FIG. 17H. Two-dimensional representation of the distribution and emergence of HMRs (.sup.55Mn) in voxel matrices generated via LA-ICP-MS using different threshold criteria. The four left hand side panels (FIG. 17A:T1, FIG. 17B:T2, FIG. 17C:T3 and FIG. 17D:T4) show data from laser ablation of a 3131 voxel area from a standard unstained 5 micron tissue section of a small cell undifferentiated carcinoma of the lung from a 51 year old female patient. The four right hand side panels (FIG. 17E:T1, FIG. 17F:T2, FIG. 17G:T3 and FIG. 17H:T4) show data from laser ablation of a 3131 voxel area of the 54 year old melanoma patient described in FIG. 14A-FIG. 14B, FIG. 15A-FIG. 15D, and FIG. 16A-FIG. 16C. The .sup.55Mn threshold values applied to the voxel matrix data are; T1, 0.5 median; T2, 1 median; T3, 1.5 median and T4, 2 median. Dark voxels in each panel are those in which .sup.55Mn values exceed the threshold for that panel. Single non-contiguous voxels above threshold are designated as singletons, while two contiguous voxels are designated as doublets.

(18) FIG. 18A to FIG. 18H. Two-dimensional representation of the distribution and no emergence of HMRs(.sup.66Zn) in voxel matrices generated via LA-ICP-MS for the same samples from the two patients of FIG. 17A-FIG. 17H. The four left hand side panels (FIG. 18A:T1, FIG. 18B:T2, FIG. 18C:T3 and FIG. 18D:T4) show data from laser ablation of the same 3131 voxel area as in FIG. 17A-FIG. 17H which was simultaneously analyzed for .sup.66Zn, .sup.56Fe and .sup.63Cu. The data for .sup.66Zn are shown. The four right hand side panels (FIG. 18E:T1, FIG. 18F:T2, FIG. 18G:T3 and FIG. 18H:T4) show data from laser ablation of the same 3131 voxel area as in FIG. 17A-FIG. 17H. The .sup.66Zn threshold values applied to the voxel matrix data are as used previously; T1, 0.5 median; T2, 1 median; T3, 1.5 median and T4, 2 median. Dark voxels in each panel are those in which .sup.66Zn values exceed the threshold for that panel.

(19) FIG. 19A-FIG. 19B: Histograms showing median .sup.55Mn contents expressed as Calibrated Counts per Second (CC/S) of .sup.55Mn in laser ablated samples from tumours of 20 patients with small cell carcinomas of the lung. Each grey square represents the median value calculated from a total area of approximately 1,800 voxels from the tumour of a single patient which was sampled. FIG. 19A: The top histogram illustrates 14 patients with tumours without HMRs(.sup.55Mn). FIG. 19B: The lower histogram shows tumours with HMRs(.sup.55Mn) from six patients denoted t1 through t6 which were determined at the standard T4 threshold of 2 median. Black squares are the median values of the HMRs(.sup.55Mn) found in the tumour from each of patients t1 to t6 determined using the standard T4 threshold of 2 median. The HMR(.sup.55Mn) values are joined via a dotted line to the bulk median .sup.55Mn value of that tumour for each of patients t1 to t6. The bins in the histogram are 200 CC/S units.

(20) FIG. 20A-FIG. 20H: Schematic representation of the size, shape and content of a selected sample of contiguous voxel configurations in a simulated 2D tumour landscape. Voxels containing cancerous cells above a designated threshold are shown in black. FIG. 20A illustrates the position of all eight possible voxel configurations when only voxel doublets are examined. FIG. 20B and FIG. 20E illustrate one of the many possible contiguous voxel configurations that satisfy the criterion of an 88 above threshold HMR(.sup.AM), where .sup.AM represents Any Metal. FIG. 20C illustrates a voxel configuration that is only 77 voxels which would fall below an 88 minimal threshold. FIG. 20D and FIG. 20F illustrate voxel configurations that would be characteristic of lymphatic vessels or ducts. FIG. 20G illustrates a voxel configuration that is characteristic of single file movements of cancerous cells, or of significant machine stutter. FIG. 20H is characteristic of the multiple voxel configurations seen in some melanomas where the voxels above threshold can be due either to cells with high levels of any metal, or melanin granules that bind any metal.

(21) FIG. 21: Histogram showing the median .sup.55Mn contents expressed as Calibrated Counts per Second (CC/S) of .sup.55Mn in laser ablated samples from a total area of approximately 1800 voxels of tumours of 10 patients with diffuse B cell lymphoma. Each grey square represents the median value calculated from the tumour of a single patient. The bins in the histogram are 200 CC/S units.

(22) FIG. 22A-FIG. 22B: Histograms of median .sup.55Mn contents expressed as Calibrated Counts per Second (CC/S) of .sup.55Mn in laser ablated samples from tumours of 55 patients with classic seminoma. Each grey square represents the median value calculated from a total area of approximately 1,800 voxels of the tumour of a single patient. FIG. 22A. The top histogram illustrates the 51 seminomas without HMRs(.sup.55Mn). FIG. 22B. The lower histogram shows tumours with HMRs(.sup.55Mn) from four patients denoted S4 through S7 which were determined at the standard T4 threshold of 2 median. Black squares are the median values of the HMRs(.sup.55Mn) found in the tumour from each of patients S4 to S7. The HMR(.sup.55Mn) values are joined via a dotted line to the bulk median .sup.55Mn value from the tumour of each of patients S4 to S7. The bins in the histogram are 200 CC/S units. Patients denoted S1, S2 and S3 are outliers with high median values, even though their tumours do not contain HMRs(.sup.55Mn) under the standard threshold.

(23) FIG. 23A-FIG. 23B: A histogram of median .sup.55Mn contents expressed as Calibrated Counts per Second (CC/S) of .sup.55Mn in laser ablated samples from tumours of 10 patients with mesothelioma. Each grey square represents the median value calculated from a total area of approximately 1,800 voxels of the tumour of a single patient. FIG. 23A. The top histogram illustrates 9 tumours without HMRs(.sup.55Mn). FIG. 23B. The lower histogram shows a tumour with a single HMRs(.sup.55Mn) determined at the standard T4 threshold of 2 median. The black square is the median value of the HMR (.sup.55Mn) found in the tumour of one mesothelioma patient. The HMR(.sup.55Mn) value is joined via a dotted line to the bulk median .sup.55Mn value of that tumour. The bins in the histogram are 200 CC/S units.

(24) FIG. 24A-FIG. 24B: Histograms showing median .sup.55Mn contents expressed as Calibrated Counts per Second (CC/S) of .sup.55Mn in laser ablated samples from tumours of 25 patients with either glioblastoma multiforme or astrocytomas of the brain. Each grey square represents the median value calculated from a total area of approximately 1800 voxels of the tumour of a single patient. FIG. 24A. The top histogram illustrates 24 tumours without HMRs(.sup.55Mn). FIG. 24B. The lower histogram shows a tumour with a single HMRs(.sup.55Mn) determined at the standard T4 threshold of 2 median. The black square is the median value of the HMR (.sup.55Mn) found in that tumour. The HMR(.sup.55Mn) value is joined via a dotted line to the bulk median .sup.55Mn value of that tumour. The bins in the histogram are 200 CC/S units.

(25) FIG. 25A-FIG. 25B: Histograms showing median .sup.55Mn contents expressed as Calibrated Counts per Second (CC/S) of .sup.55Mn in laser ablated samples from tumours of 64 patients with melanoma. Each grey square represents the median value calculated from a total area of approximately 1,800 voxels of the tumour of a single patient, except in the case of the two melanoma patients each with two major lineages within the same tumour, where the sampled area is less than 900 voxels per lineage. FIG. 25A. The top histogram illustrates the tumours without HMRs(.sup.55Mn). FIG. 25B. The lower histogram shows the tumours with HMRs(.sup.55Mn) determined at the standard T4 threshold of 2 median. Black squares are the median values of the HMRs(.sup.55Mn) found in the tumours. The HMR.sup.55Mn) values are joined via a dotted line to the bulk median .sup.55Mn value of that tumour. Some tumours have multiple HMRs(.sup.55Mn) and these are shown on the same dotted line for that tumour. The bins in the histogram are 200 CC/S units. Note that the scale has been compressed since many CC/S values exceed values of 10,000.

(26) FIG. 26: Schematic diagram of the complex polymeric structure of melanin which binds the metals used for analysis in this application.

(27) FIG. 27A and FIG. 27B are photographic images of a standard 5 micron H&E stained tissue section from a formalin-fixed paraffin-embedded block from a malignant melanoma of the chest wall of a 45 year old female illustrating the tissue and cellular morphology. FIG. 27C. .sup.66Zn levels in individual voxels of this section with contiguous voxels of high .sup.66Zn content shown as black areas. FIG. 27D. .sup.63Cu levels in individual voxels of this section with contiguous voxels of high .sup.63Cu content shown as black areas. FIG. 27E. .sup.56Fe levels in individual voxels of this section with contiguous voxels of high .sup.56Fe content shown as black areas. FIG. 27F. .sup.55Mn levels in individual voxels of this section with contiguous voxels of high .sup.55Mn content shown as black areas.

(28) FIG. 28A-FIG. 28C: Histograms showing median .sup.55Mn contents expressed as Calibrated Counts per Second (CC/S) of .sup.55Mn in laser ablated samples from a total area of approximately 1,800 voxels of the tumours of 64 patients with melanoma. As described previously, two of the 64 patients each had two major lineages within their tumours, and each such lineage is represented separately in the histogram with a square. The melanoma samples have been grouped into those that derive from the primary site (FIG. 28A: top histogram), those that derive from lymph nodes, (FIG. 28B: middle histogram), and those that derive from a distant site (FIG. 28C: lower histogram).

(29) FIG. 29A. Photographic image of a standard 5 micron H&E stained tissue section from a formalin-fixed paraffin-embedded block from a small cell carcinoma of the lung of a 54 year old male illustrating the tissue and cellular morphology of areas of darkly staining cancerous cells and the more lightly staining stromal regions. FIG. 29B. 2D relief image of the same section with .sup.66Zn levels above a selected threshold shown in white/grey and the stromal components below this threshold shown in black.

(30) FIG. 30A. Photographic image of a standard 5 micron H&E stained tissue section from a formalin-fixed paraffin-embedded block from a malignant melanoma of the rectum of a 66 year old male illustrating the tissue and cellular morphology of areas of darkly staining cancerous cells, even darker staining melanin concentrations, and the more lightly staining stromal regions. FIG. 30B. 2D relief image of the same section with .sup.66Zn levels above a selected threshold shown in white/grey and the stromal components below this threshold shown in black.

(31) FIG. 31A-FIG. 31B: Histograms of median .sup.55Mn contents expressed as Calibrated Counts per Second (CC/S) of .sup.55Mn in laser ablated samples from tumours of 15 patients with breast cancer. Each grey square represents the median value calculated from a total area of approximately 1,800 voxels of the tumour of a single patient. FIG. 31A. The top histogram illustrates 4 tumours without HMRs(.sup.55Mn). FIG. 31B. The lower histogram shows 11 tumours with HMRs(.sup.55Mn) determined at the standard T4 threshold of 2 median. Black squares are the median values of the HMRs(.sup.55Mn) found in the tumour of each of the 11 patients. The HMR.sup.55Mn) values are joined via a dotted line to the bulk median .sup.55Mn value of that tumour. Some tumours have multiple HMRs(.sup.55Mn) and these are shown on the same dotted line for that tumour. The bins in the histogram are 200 CC/S units.

(32) FIG. 32A. Photographic image of a standard 5 micron H&E stained tissue section from a formalin-fixed paraffin-embedded block from well differentiated carcinoma of the breast taken from a 39 old female illustrating the tissue and cellular morphology. FIG. 32B. 2D relief image of .sup.55Mn levels in individual voxels of this section after LA-ICP-MS. Multiple clusters of contiguous voxels of high .sup.55Mn content are shown as black areas.

(33) FIG. 33A-FIG. 33H: Two-dimensional representation of distribution and emergence of HMRs(.sup.55Mn), but no emergence of HMRs(.sup.66Zn) in voxel matrices generated via LA-ICP-MS using different threshold criteria from the 39 year old female in FIG. 32A-FIG. 32B. The four left hand side panels (FIG. 33A:T1, FIG. 33B:T2, FIG. 33C:T3 and FIG. 33D:T4) are .sup.55Mn data from laser ablation of a 3131 voxel area from carcinoma of the breast at the different standard thresholds, (T1, 0.5 median; T2, 1 median; T3, 1.5 median and T4, 2 median). The four right hand side panels (FIG. 33E:T1, FIG. 33F:T2, FIG. 33G:13 and FIG. 33H:T4) are .sup.66Zn data from laser ablation of the identical 3131 voxel area and the same threshold criteria. Dark voxels in each panel are those in which .sup.55Mn or .sup.66Zn values exceeded the threshold for that panel.

(34) FIG. 34A-FIG. 34B: Analysis of the variation in the distribution of .sup.55Mn and .sup.66Zn voxel values in the breast cancer sample from the 39 year old female whose 2D data were presented in FIG. 33A-FIG. 33H. FIG. 34A shows the right hand side Skew of the .sup.55Mn voxels, with the frequency of voxels in a particular bin on the Y axis, and voxel values on the X axis. FIG. 34B illustrates the near symmetrical distribution of .sup.66Zn voxel values from the same sample.

(35) FIG. 35A. Photographic image of a standard 5 micron H&E stained tissue section from a formalin-fixed paraffin-embedded block from an invasive ductal carcinoma of the breast from a 48 year old female illustrating the heterogeneous tissue and cellular morphology. FIG. 35B. Description of the major morphological features of the above image, showing cancerous cells in lymphatic ducts, (C1 through C5), normal ducts (N) surrounded by associated adipocytes, and a concentration of immune cells, (immune), together with other stromal components.

(36) FIG. 36A and FIG. 36B are photographic images of a standard 5 micron H&E stained tissue section from a formalin-fixed paraffin-embedded block from an invasive ductal carcinoma of the breast from the 48 year old female of FIG. 35A-FIG. 35B, illustrating the heterogeneous tissue and cellular morphology. FIG. 36C to FIG. 36F. 2D relief images of the differential distribution of the four metals .sup.55Mn (C), .sup.66Zn (D), .sup.56Fe (F) and .sup.63Cu (E) in the areas containing adipocytes and in some parts of the stromal regions.

(37) FIG. 37A. Illustration of the clinical data from patient X with adenocarcinoma of the prostate and a PSA of 8.1 showing the abnormalities in different regions of the prostate with 7 out of 12 regions showing changes of little significance, and five regions with cancer as defined by Gleason scores no higher than 7 (FIG. 37A). FIG. 37B. Diagnostic summary (Part A through Part L) of the extent of involvement of the twelve core needle biopsies from patient X.

(38) FIG. 38: Histograms of median .sup.55Mn contents expressed as Calibrated Counts per Second (CC/S) of .sup.55Mn in laser ablated samples from tumours of 10 patients with cancer of the prostate. Each grey square represents the median value calculated from a total area of approximately 1800 voxels of the tumour of a single patient. No HMRs(.sup.55Mn) were present under the standard threshold conditions of 2 median applied herein.

(39) FIG. 39 A-FIG. 39E Magnetic Resonance Images from the brain of patient Y, a 70 year old male. FIG. 39A. Presence of a contrast enhancing lesion in the paramedial left parietal lobe of the brain (arrowed) prior to drug treatment and radiation. FIG. 39B. MRI of the same brain region after radiation and immunotherapeutic drug treatment. FIG. 39C. Photographic image of a standard 5 micron H&E stained tissue section from a formalin-fixed paraffin-embedded block from a primary melanoma of the same patient resected years prior to the brain scans. FIG. 39D and FIG. 39E. Higher power illustrations of the regions of interest analyzed by LA-ICP-MS.

(40) FIG. 40: The calibrated signal tracks (arrowed .fwdarw.) of .sup.55Mn from five separate areas of cancerous cells of the primary melanoma of patient Y. There were six contiguous ablation tracks of standard 35 micron35 micron5 micron voxels in each chosen area, with track lengths of 19, 19, 13, 19 and 19 voxels respectively. The raw numerical values of calibrated counts per second for .sup.55Mn are shown for each of the adjacent voxels in each of six ablation rows.

(41) FIG. 41A: Patient Z with a squamous carcinoma of the oral region prior to radiation treatment. FIG. 41B: Patient Z six months after radiation treatment.

(42) FIG. 42: Flow diagram for melanoma patients whose solid tumour(s) are analyzed by LA-ICP-MS. The optional pathways that provide evidence for a physician as to whether to advise radiation treatment for the patient, or to refrain from the use of such a modality, are based on the homogeneity or heterogeneity of the ATI. The key decision points concern whether the .sup.55Mn voxel values fall above or below certain designated thresholds, and the degree of melanization of the tumour, in terms of metal binding.

(43) FIG. 43: Flow diagram for the non-melanoma patients whose solid tumour(s) are analyzed by LA-ICP-MS. The optional pathways that provide evidence for a physician as to whether to advise radiation treatment for the patient, or to refrain from the use of such a modality, are based on the homogeneity or heterogeneity of the ATI. The key decision points concern whether the .sup.55Mn voxel values fall above or below certain designated thresholds.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

(44) In work leading up to the present invention the inventor noted that the 2D spatial characteristics of many tumours, in particular, melanoma, varies greatly, e.g., as shown in FIG. 1 for comparative purposes. FIG. 1 shows a standard 5 micron, H&E stained tissue section from a formalin-fixed paraffin-embedded (FFPE) block from a stage Ib malignant melanoma of the neck of a 48 year old male, with the TNM staging criteria of T2aN0M0, where T2 represents invasion of the muscularis propria; N0 represents no lymph node metastasis, and M0 represents no distant metastasis. Note that the top right-hand side of this section consists of pale amelanotic tumour cells, while the darker regions on the left-hand side and bottom of the Figure are composed of cells containing darkly staining melanin, and with even larger patches of melanin, both intra- and extra-cellular, as black spots in the lower part of the image. As described above herein, the pathological diagnosis of such tumours, which is utilized to determine treatment modalities, and upon which the decision is made as to whether or not radiation will be used, is the subjective analysis of such H&E stained slides. Whilst the pathological diagnosis is also aided by probing such tissue sections with a panel of antibodies that vary from one type of tumour type to the next, such results are also subjective and far from quantitative. Until the present invention, there was no quantitative test available, or a test that predicts sensitivity or resistance to radiation, that aids the oncologist in making a decision on whether or not to utilize radiation as part of a treatment regimen.

(45) In particular, the Lancet Oncology Commission recently presented new evidence on the issues involved in expanding global access to radiotherapy (Atun et al., Lancet Oncology 2015, 16, 1153-1186). It highlighted radiotherapy as a fundamental component of effective cancer treatment and control and that it is generally used sub-optimally. The Lancet editors pointed out that radiotherapy is more scalable than other treatment modalities and is uniquely placed to deliver effective curative and palliative care (Coburn & Collingridge, Lancet Oncology, 2015, 16, 1143). However, there are two continuing roadblocks to tailoring radiotherapy to the needs of the individual patient. First, and not until the present invention, there has been no quantitative test that measures the extent to which a given tumour, in a given patient, will respond to a given regimen of radiotherapy. Second, there has been no quantitative test that measures the extent to which any tumour is likely to reoccur following radiation therapy.

(46) Until the present invention, clinical decisions were based on medical art, not quantitative measures to determine radiation responsiveness.

(47) Without being bound to any particular theory, the inventor advances the following explanation for the unexpected finding that ATI could be used to predict radiation responsiveness: ionizing radiation, such as gamma-rays and X-rays lead to the radiolysis of water which leads to the formation of the same type of chemical entities in all cells, be those cells bacterial, algal, fungal, invertebrate, or the hundreds of cell types in a human body, or abnormal cell types which arise as a result of perturbations leading to cancerous cells.

(48) Radiolysis universally generates the same types of reactive molecules, the three key ones being; the hydroxyl radical OH., the superoxide radical O.sub.2. and hydrogen peroxide (H.sub.2O.sub.2), (Daly, M; Nature Reviews Microbiology, 7, 237-245. 2009).

(49) In normal healthy mammalian cells that have not been irradiated, the same three molecules are also formed as part of the normal mitochondrial respiratory processes occurring via the mitochondrial electron transport chain. If oxygen receives less than its full complement of electrons, the result is the formation of O.sub.2. and H.sub.2O.sub.2. If the H.sub.2O.sub.2 is not dealt with immediately, any stray iron Fe.sup.2+ atoms will react with it and generate the highly dangerous hydroxyl radical OH.

(50) In mammalian cells, the superoxide radical O.sub.2. is handled by mitochondrial, cytosolic and extracellular enzyme systems, namely manganese superoxide dismutase MnSOD located in the mitochondrion, copper-zinc superoxide CuZnSOD in the cytosol and extracellular superoxide dismutase ecSOD predominantly anchored to endothelial cells.

(51) The H.sub.2O.sub.2 is cleared by both catalases and glutathione peroxidases to produce water and molecular oxygen.

(52) The highly dangerous hydroxyl radical OH. is not dealt with by enzymological processes. One example of the consequences of the high level of oxidative metabolism is in the mammalian brain, where it makes brain cells very vulnerable to lipid peroxidation from OH.

(53) The inventor reasoned that the chemical elements and levels of elements contributes to radio-responsiveness, for example manganese. The inventor noted that whilst a whole body exposure of 10 Gray (Gy) is lethal to most vertebrates, some bacteria such as D. radiodurans survive doses in excess of 17,000 Gy. One contributing mechanism by which it may achieve this is that it accumulates 150 times more manganese and 3 times less iron (Fe) than radiation sensitive bacterial species (Daly, M; Nature Reviews Microbiology, 7, 237-245, 2009). Bacterial species with the highest manganese-to-iron ratios are the most radioresistant, whereas those with the lowest Mn/Fe ratios are hypersensitive. The mechanistic underpinnings of radio-responsiveness reveal that manganese accumulation shields proteins with iron-sulphur (FeS) complexes from superoxide radicals such as (O.sub.2.) formed during irradiation. This shielding by manganese prevents the release of ferrous ions (Fe.sup.2+) from iron-sulphur containing proteins, thus preventing the highly damaging interactions of Fe.sup.2+ with hydrogen peroxide. If Fe.sup.2+ manages to react with H.sub.2O.sub.2, the result is an hydroxyl radical OH. which is dangerous and will oxidize almost every type of biological molecule.

(54) In contrast to hydrogen peroxide, O.sub.2. does not easily cross membranes and hence builds up in cellular compartments. Thus any cellular system that can effectively shield FeS containing proteins from exposure to the O.sub.2. as well as minimizing the amount of Fe.sup.2+ available for the Fenton reaction will minimize damage following irradiation and enhance radio-resistance. The inventor reasoned that the bacterial data indicate that the manganese ion is a protective metal and even at high concentrations is largely innocuous to a bacterial cell, and likely well tolerated by many cell types of multi-cellular organisms. Conversely, any bacterial system that is low in manganese and high in free iron is likely less able to protect proteins and lipids from damage and the Fenton reaction will ensure that the damaging OH. will increase its level leading to protein and lipid damage and death of the cell, hence cellular radio-sensitivity. A similar situation will pertain to eukaryotes.

(55) The intracellular availability of free iron is known to play a key role in irreversible protein damage via protein carbonylation. In yeast, carbonylation levels are increased when yeast lack a particular iron storage protein, the homolog of which is the human mitochondrially located frataxin protein. Introduction of the human ferritin into such a defective yeast strain, partially restores the iron storage capacity of such yeast, decreases free iron levels and counteracts the elevated carbonylation levels. Thus, iron storage proteins are likely to be important players in preventing cellular damage following ionizing radiation.

(56) It is noted that although the superoxide radical O.sub.2. is highly charged, it does not react with DNA. Rather, it reacts with selected targets, these being any exposed iron-sulphur (FeS) groups of certain proteins.

(57) In summary, OH. is extremely damaging to all cellular components, but the collateral damage it causes is restricted to a few Angstroms from its site of formation owing to its short lifetime. H.sub.2O.sub.2 by contrast, can diffuse throughout the cell and reacts with Fe.sup.2+, yielding one of the most powerful oxidizing reactions known. This reaction produces more OH. The bacterial data reveal that responsiveness to ionizing radiation is a continuous biological variable, which has multiple inputs: from cellular manganese concentration and from the associated enzymology that scavenges radicals, sequesters free iron, reduces hydrogen peroxide and minimizes and remanufactures proteins.

(58) In vitro systems have demonstrated that the human manganese superoxide dismutase converts the superoxide radical O.sub.2. to H.sub.2O.sub.2 and O.sub.2. In a number of human in vitro cellular systems, human MnSOD protein levels and activity have been correlated with an increased resistance to ionizing radiation. Similarly, lowering the level of MnSOD protein and activity in cellular systems results in decreased radio-resistance.

(59) Thus, in carefully controlled experimental circumstances, where the genetic background of the experimental and control cells is kept constant, and the only variable is either the introduced gene, its antisense product, or an empty vector, the mitochondrially-located MnSOD protein helps to protect cells from the effects of ionizing radiation.

(60) There are only three superoxide dismutases in human cells that handle the superoxide radical O.sub.2. and these superoxide dismutases all have different cellular locations. Manganese superoxide dismutase MnSOD is located in the mitochondrion, copper-zinc superoxide CuZnSOD is found in the cytosol and the nucleus, while extracellular superoxide dismutase ecSOD is predominantly anchored to endothelial cells. The three superoxide dismutases deal with O.sub.2. in these three different locations.

(61) Whilst much attention has been focussed on superoxide dismutases, the basic chemistry and biochemistry of chemical elements, such as Mn and Cu, their location and usage within a cell is far more widespread than the specificity of superoxide dismutases e.g., Mn exists in complex formation (Daly et al., PLoS ONE, 5, e12570, 2010; Slade and Radman, Microbiology and Molecular Biology Reviews, 75, 133-191, 2011).

(62) FIG. 2 illustrates three different manganese-binding entities. Manganese forms complexes with orthophosphate and scavenges O.sub.2.. Manganese also forms complexes with free amino acids or peptides to scavenge and decompose hydrogen peroxide, hydroxyl radicals (OH.) and O.sub.2.. Nucleosides, free amino acids and sundry organic metabolites scavenge hydroxyl radicals (Slade and Radman, Microbiology and Molecular Biology Reviews, 75, 133-191, 2011).

(63) Inorganic metals, such as, iron, copper, zinc and manganese are most commonly thought of as catalytic cofactors for proteins.

(64) Notwithstanding the foregoing suggestion that metals play a role in radio-resistance, until the present invention, manganese has not been quantified in a 2D cancerous context of a tumour compared to normal, whereby the radio-sensitivity/radio-resistance is determined and subsequently utilized to firstly make decisions regarding radiation treatment, and secondly, to predict the probability of tumour reoccurrence after radiation treatment.

(65) All the hardware and software for carrying out the invention are currently available both commercially and clinically. It will be apparent to a person skilled in the art that technological variations may be useful. Preparation of the tissue sections, cells or populations of cells for analysis according to the invention may be performed according to any known method in the art. For example, unstained tissue sections, unstained frozen sections, or cells deposited as a monolayer via a SurePath system are prepared according to the art and for example, as described herein. Any system available in the art may be used, e.g., the SurePath system yields a monolayer of cells on a slide in a compact circular area, and material prepared in this manner would be suitable for LA-ICP-MS methodology. Any method known in the art may be used to prepare the test sample and/or control sample of the invention provided the 2D integrity has not been lost. A number of ablation tracks can be made, e.g. via laser ablation-Inductively coupled plasma-mass spectrometry (LA-ICP-MS), laser ablation-time-of-flight-mass spectrometry (LA-TOF-MS), inductively coupled plasma-optical emission spectroscopy (ICP-OES), microwave plasma-atomic emission spectroscopy (MP-AES), laser induced break down spectroscopy (LIBS), secondary ion mass spectrometry (SIMS), X-ray absorption near edge structure (XANES), atomic absorption spectroscopy (AA) or X-ray fluorescence (XRF).

(66) If the values within a single or multiple ablation tracks all fall either above, or below, a designated threshold, there is no need for an examination by a pathologist. Conversely, for any tracks that are flagged by the computer software, then the companion stained slide is examined by a pathologist.

(67) Alternately, a slide is first examined by a pathologist, who would direct where the preferred ablation tracks should be done.

(68) It should be noted that a single ablation track (of width less than 110 microns) would ablate a 1 cm track of tumour tissue in about 70 seconds. In this respect, multiple tracks across a tumour sample, or multiple dispersed areas within the tumour, located and designated by a pathologist for subsequent ablation analysis, may be run in minutes with current technology.

(69) Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry (LA-ICP-MS), was originally introduced by Gray, Analyst, 110, 551-556. 1985, employing a ruby laser. This was subsequently superseded by solid state Nd:YAG and excimer-based lasers. The use of the former is described in Hare, D, et al., Analyst 134, 450-453. 2009 (Table 1 of Hare et al. for operational parameters). In regard to excimer lasers, the beam is generated by a gas mixture which is a combination of the noble gases (argon, krypton or xenon) with a reactive gas, such as chlorine or fluorine. Under high pressure and electrical stimulation, a pseudo molecule, (XeCl, KrF or ArF), termed exciplex, is created which gives rise to laser light in the ultraviolet.

(70) A popular excimer-based system in the context of the LA-ICP-MS, is the Agilent 7700 ICP-MS coupled to a New Wave excimer generating a 193 nanometer wavelength, the laser being first used by Gunther et al., J. Anal. At. Spectrom. 12, 939-944, 1977. The elemental analysis of free ions is performed on a LA-ICP-MS, e.g., using ICP-MS instrument Agilent Technologies 7700 series which is interfaced with a New Wave Research Excimer 193 laser ablation unit. In addition, the ICP-MS Instrument is fitted with an octopole collision/reaction cell. Other ICP-MS instruments are also available including Thermofisher's iCAP Q ICP-MS, Perkin Elmer's Nexion Series, Shimadzu ICP-MS 2030, and Tofwerk's ICP-TOF-MS. Instead of an ICP-MS system, the ATI may also be determined with atomic emission spectroscopy techniques such as inductively coupled plasma-optical emission spectroscopy (ICP-OES), microwave plasma-atomic emission spectroscopy, or laser induced break down spectroscopy (LIBS), or secondary ion mass spectrometry (SIMS), or X-ray absorption near edge structure (XANES), or atomic absorption spectroscopy (AA), or X-ray fluorescence (XRF)

(71) High purity liquid argon (Ar) is used as the carrier gas and plasma source, while ultra-high purity (99.999%) hydrogen (H2) is used as the reaction gas. The LA-ICP-MS system is tuned on a daily basis and for both standard mode and reaction mode using NIST 612 Trace Elements in Glass for maximum sensitivity and to ensure low oxide formation. Low oxide production is assured by measuring a mass-to-charge ratio (m/z) of 248/232 (representing .sup.232Th.sup.16O.sup.+/.sup.232Th.sup.+) and is consistently less than 0.3%. The instrument is fine-tuned for tissue analysis using matrix-matched tissue standards.

(72) Typical operational parameters for the LA-ICP-MS system are given below, and for clarity, they are shown separately for the 7700 ICP-MS and for the laser.

(73) ICP-MS Parameters.

(74) The radio frequency power is 1250.0 Watts; the cooling gas flow rate is 15.0 liters/minute; the carrier gas flow rate is 1.2 liters per minute, the sample depth is 4 millimeters; the quadrupole bias is 3.0 Volts; the octopole bias is 6.0 Volts; the dwell time is 62 milliseconds per m/z; extraction lens 1 is 5.0 Volts; extraction lens 2 is 100.0 Volts and the hydrogen collision gas is 3.1 milliliters per minute.

(75) Parameters of the New Wave 193 Excimer Laser.

(76) Wavelength 193 nanometers; repetition frequency 40 Hertz; laser energy density 0.3 to 0.5 Joules per square centimeter; spot diameter 35 micrometers; line spacing 35 micrometers; monitored mass/charge ratio (m/z), 55 (Mn), 56 (Fe), 63 (Cu) and 66 (Zn). The raster speed is 140 micrometers per second (four times 35 micrometers).

(77) In brief, a glass microscope slide with an unstained section from a particular tissue is placed in an ablation chamber and a high energy laser beam (of varying diameter) is focussed onto the section and some of the biological material is vaporized as a result of energy transfer from the laser beam. The resulting particulate matter is moved by a carrier gas, which is usually Argon, Helium, or Argon plus Helium (in the system used for one embodiment of this invention, it is Argon), into the Inductively Coupled Plasma which, at a temperature exceeding 7,000 degrees Celsius, but below 10,000 degrees Celsius, atomizes and ionizes the particulate matter to its constituent elements. The use of collision/reaction cells (Tanner et al., Spectrochim. Acta. Part B, 57, 1361-1452, 2002) minimizes spectral interferences, and removes polyatomic ions such as oxygen:argon species (.sup.16O.sup.40Ar.sup.+) that would otherwise appear as a 56 signal and be incorrectly attributed to .sup.56Fe. In the dynamic reaction cell, interfering polyatomic ions are converted to a different species at a higher m/z and no longer interfere with the target ion. Following emergence from the collision cell, the ions are focussed into a quadrupole mass filter where they are separated by their m/z ratios, and detected and quantified.

(78) Instead of a quadrupole system, ICP-MS systems can incorporate a double focussing sector field mass spectrometer, or a Time-of-Flight (TOF) analyser. An ICP-TOF-MS has a very high throughput acquisition capacity of 30,000 full mass spectra per second (Resano et al., Mass Spectrom. Rev. 29-55-78, 2010).

(79) To generate a 2D elemental map of the material on a microscope slide, the laser is rastered across a sample from side to side, one track at a time from top to bottom, where the track has a chosen width. The resulting image is visualized as adjacent pixels, but since the ablated material has depth, each pixel represents a volume of tissue, a voxel.

(80) It will be apparent to the skilled artisan that the present invention provides the following advantages:

(81) a) minimal sample preparation to avoid introduced artefacts;

(82) b) the ability to simultaneously and rapidly extract multi-elemental data;

(83) c) the deconvolution of 2D information about the structure of the tumour in terms of its tumour cells, stromal cells, abnormal vasculature, transiting immune cells, and non-cellular material such as collagen bundles, all of which form the interacting milieu that constitutes the tumour and which impinges on its position between the lower boundary of radiation sensitivity and the upper boundary of radiation resistance;
d) it directly translates to the use of, or avoidance of, a therapeutic modality, namely ionising radiation for any tissue or organ type. Where radiation therapy is used, the invention determines the probability of tumour reoccurrence.
e) it is not specific for a particular type of tumour. It is applicable to any tumour, localized or metastatic, benign or malignant. As such, it is pan-diagnostic. For example, a Prostate-Specific Antigen test measures a single circulating entity in the bloodstream which may be indicative of the presence of a benign or malignant tumour, or simply benign prostatic hyperplasia, or strenuous exercise. As such it does not suggest the type of therapeutic intervention that is required. It is also specific to males. In contrast the present invention is not restricted by specificity of gender or tumour type and hence there is no need to develop specific antibodies or drugs to a particular tumour type, such as Tarceva for non-small cell lung cancer.
f) radiation, which can be external beam or implanted seeds, is delivered to a localized anatomical region, unlike small molecular weight drugs, chemotherapeutics or antibody based biologicals which are delivered via the vascular system, which spread to all tissues, and invariably have off-target effects in normal tissues and cause various levels of toxicity, such as vemurafenib in melanoma, which induces keratoacanthomas.
g) the present invention is direct. It has no intermediate steps as regards multiple preparation steps for a sample. The assay is not confounded by potential biases inherent in methods that rely for signal amplification on processes such as PCR, where the enzymes commonly used in such procedures can introduce systematic bias through differential rates of amplification of different sequences. There is no hybridization of antibodies to tissue sections with its varying specificities of hybridization and measurement of amplification signals. The present invention measures what is in a sample without the distortions that occur as a result of multi-step processing.
h) there is a seven order linear dynamic range over which measurements can be made. Both range and the linearity are crucial since they allow a true measurement of elemental abundance in a cell population without introducing potential errors by the prior art methods that require conversion or amplification of entities.

(84) The present invention is particularly suitable for detection of disease states, differentiation states of stem cells and derivative cell populations, detection or measurement of effects of medication on cell state, and any other situation where an accurate indication of cellular state is useful.

(85) The present invention will now be described in more detail with reference to specific but non-limiting examples describing specific compositions and methods of use. It is to be understood, however, that the detailed description of specific procedures, compositions and methods is included solely for the purpose of exemplifying the present invention. It should not be understood in any way as a restriction on the broad description of the inventive concept as set out above.

Example 1

(86) Preparation of Tissue Standards

(87) Normalisation and calibration experiments were run using 30 m thick sections of matrix-matched tissue standards. These standards were prepared from chicken breast tissue removed of any fat or connective material, and were partially homogenized using an OmniTech TH tissue homogenizer fitted with a polycarbonate probe (Kelly Scientific, North Sydney, New South Wales, Australia), and subsequently spiked with standard Ca, Mn, Fe, Co, Cu, and Zn solutions. Solutions were prepared using high purity (min 99.995%) soluble chloride, sulfate, or nitrate metal salts (Sigma-Aldrich, Castle Hill, New South Wales, Australia) dissolved in 1% HNO.sub.3 (Choice Analytical, Thornleigh, New South Wales, Australia) and diluted to concentrations of ca. 100,000 g mL-1 and 10,000 g mL-1. Aliquots of the chicken breast were then spiked with varying concentrations of each of the elements and homogenized at low speed for 5 min. Six ca. 250 mg aliquots of each homogenized tissue standard were digested in 5:1 Seastar Baseline grade HNO.sub.3/H.sub.2O.sub.2 (Choice Analytical) in a Milestone MLS 1200 closed vessel microwave digester (Kelly Scientific) and analysed using solution ICP-MS to confirm the concentration and homogeneity of each element in the tissue standards. The spiked tissue was frozen and sectioned into 30 m sections and placed onto glass microscope slides for analysis. This methodology can be simply applied to sections cut at 5 microns, a small part of which is added to each tumour slide.

(88) Calibration and Background Analysis

(89) The prepared matrix-matched tissue standards were used to construct calibration curves for sensitivity normalisation of analytes under conditions described later for running the LA-ICP-MS. Data for a 10 s period prior to ablation of the tissue were collected to obtain a background signal for each m/z from the gas blank. This methodology was used to allow for a comparison between different runs of different tumours. For example, a 2K calibrated counts/second was achieved.

(90) Sensitivity Threshold(s)

(91) An alternate way of setting thresholds to that described earlier as standard deviations relative to means and medians of a radiation sensitive control, or as empirically determined values from any tumour type, is to set a sensitivity threshold by apportioning, for example, the testicular samples in the top histogram of FIG. 11A-FIG. 11F, such that the threshold is defined as a percentage of that histogram. For example, the seminomas of FIG. 11A-FIG. 11F derive from 55 patients and 41/55 of these patients constitute 75% of the histogram who fall below this 75% threshold, and all these patients are radiation sensitive. Next, 47/55 patients constitute 85% of the histogram and will be radiation sensitive. Finally, 52/55 patients constitute nearly 95 of the histogram, but some of these outlier patients are likely to have some degree of radiation resistance. These percentage figures representing a proportion of the histogram are another way of expressing the variable threshold that can be chosen for radio-responsiveness.

Example 2

(92) In one example, a section of normal brain tissue on a microscope slide as described herein, was laser ablated and its metallomic contents were analyzed by the LA-ICP-MS system.

(93) FIG. 3A shows a representative H&E stained 5 micron tissue section from a block of a normal human cortex from a 50 year old human female showing the characteristics of normal cortical tissue. This Figure illustrates the position of a three-track ablation (taken from the larger multi-track tissue ablation) that was carried out on an adjacent unstained section from the same block, but on a different slide. The H&E stained area shown in FIG. 3A corresponds to the equivalent unstained area analysed by LA-ICP-MS on the different slide. The unstained 5 micron tissue section was cut from the same block, laser ablated and analysed using the Agilent technologies 7700 LA-ICP-MS system, with detailed operational parameters described previously. Unstained tissue sections are generally used in order to avoid the introduction of any chemical elements that occur in the stains, staining solutions or which are leached from the containers during processing. Stained sections can also be used.

(94) To generate an elemental map of the unstained material on the separate microscope slide, the laser was rastered across the slide laterally, one ablation track at a time from top to bottom. Generally, this produces up to hundreds of horizontal tracks depending on the size of the tissue sample, with each track being 35 microns in width. Lateral resolution is a function of the integration time of the quadrupole mass spectrometer and was chosen in these data sets to give a resolution of 35 microns. A complete image was generated, which consists of a basic unit of a 35 micron by 35 micron pixel. Since the tissue section is 5 microns thick, each pixel represents a 35355 volume of tissue, namely a voxel of 6125 cubic microns. An ablation run of 50 horizontal tracks, in which each ablation track consists of, say, 70 voxels, generated a 2D image consisting of 3500 voxels.

(95) The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn, and .sup.56Fe was determined. The results are shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 70 voxels per track (FIG. 3B). The background-subtracted median values per voxel for these metals, expressed as calibrated counts per second, are 3,409, 6,651 and 441,317 respectively. Measurements of .sup.63Cu were unavailable for technical reasons.

(96) These raw data are presented in graphical form in FIG. 3C-FIG. 3E, where each track consisted of 70 voxels. FIG. 3C-FIG. 3E shows graphs of raw numerical values of calibrated counts per second CC/S for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows. The unstained 5 micron tissue section was cut from the same block, laser ablated and analysed using the Agilent technologies 7700 LA-ICP-MS system, with detailed operational parameters described below.

(97) As the laser rasters across the unstained tissue section on the slide, the values for each of the metals change from background levels to those characteristic of the tissue sample and then return to background values again. The variation in values generally follows the morphological changes and variations in tumour architectures seen in the stained tissue sections.

(98) In this Example, the raw free ion manganese values in FIG. 3C, reveal that there is a spike in voxel 55 in ablation row two, while none of the eight voxels contiguous to it show this large deviation from the median (as seen from the data of FIG. 3B). These large single spikes for an individual voxel are due to machine jitter and such spikes are almost always more than 3 standard deviations from the median value. While shown here for the illustrative purposes of raw data presentation, they will be removed prior to statistical analyses as they represent machine noise. They are sufficiently infrequent that even if they remained part of the data analysis, they have a miniscule effect on important outcomes such as median values.

Example 3

(99) Neoplastic Samples with Contrasting Radiation Resistances and Sensitivities

(100) As discussed above, the Lancet Oncology Commission highlighted radiotherapy as a fundamental component of effective cancer treatment and control and that it is generally used sub-optimally (Atun et al., Lancet Oncology 2015, 16, 1153-1186). Until the present invention, the clinically accepted norm is that some tumour types such as glioblastoma, mesothelioma and melanoma are largely radiation resistant, while most lymphomas, small cell tumours of the lung, and tumours of the testis, such as classic seminomas, are radiation sensitive. Other major tumour types such as cancers of the breast and prostate are generally thought to be of intermediate radio-responsiveness, but intermediate is context dependent and difficult to determine. The clinical reality relates to what percentage of primary cancers can be eradicated with radiotherapy alone and the dose that is required, as well as the percentage susceptibility of various metastatic lessons to ablative radiotherapy. Thus large seminoma and lymphoma masses are easily eradicated with radiotherapy while sarcomas and pancreatic carcinomas are rarely eradicated, despite the dose. In addition, radiotherapy is considered curative for some low/intermediate risk prostate cancers and some prostate cancer metastases to the bone, and so such prostate cancers could certainly be classified as very radiosensitive.

(101) Given the above clinical variation, we have chosen six tumour types for evaluating the ATI, as the first three provide the bookends for radiation resistance, and the second three provide the bookends for radiation sensitivity. It is clinically acknowledged, however, that even within each of these six tumour types, there are a minority which do not respond to radiation treatment in the expected manner. This is due to problems with methods used prior to the present invention e.g., low resolving power of H&E stained pathological material and the considerable disagreement among pathologists of the taxonomy within any tumour type. Even with the addition of antibody staining of formalin-fixed paraffin-embedded and Hematoxylin and Eosin (H&E) stained tissue sections, the interpretation of the variation within a tumour type remains subjective and the concordance among pathologists is variable (Elmore et al., JAMA, 313; 1122-1132. 2015).

(102) Melanoma illustrates the challenges of personalizing radiotherapies. Historically it has been considered to be intrinsically radio-resistant, a perception originating from early cell culture studies and analyses of survival curves. This belief has recently been evaluated on the basis of data from 4 decades of the clinical use of radiation therapy in melanoma patients (Mahadevan et al., Oncology, 29, (10): 743-751, 2015). The newer interpretation is that the radio-responsiveness of melanoma is diverse. It is accurately summarized by Burmeister; I've been working with melanoma for over 25 years and it still amazes me how in some patients the disease just melts away and in others it just laughs at you and kills the patient within a few weeks or months . . . there is an incredible variation in the behaviour of this disease in individual patients.

(103) As the atomic data presented herein reveal, there is now, for the first time, (i) a quantitative underpinning of which melanoma patients, (as well as other tumour types) are suitable candidates for radiotherapy, (ii) a quantitative basis for determining in which patients a cancer is likely to re-occur after radiotherapy, and (iii) a measurable basis for tumour variation. It should also be noted that such heterogeneity is not confined to melanoma. Breast and prostate cancer patients and their tumours are also heterogeneous at multiple levels, one being their response to radiation treatment.

(104) 1. Radiation Resistant Tumours

(105) In another Example, the metallomic values of neoplasms that are considered to be at the radiation resistant end of the clinical spectrum were analyzed. For example, a deadly brain neoplasm, glioblastoma multiforme, which is considered to be at the radiation resistant end of the clinical spectrum, was analysed as described in Example 2. This example illustrates how the metallomic values of a brain neoplasm differ from those of normal brain tissue, when both the normal and neoplastic samples are present on the same slide and are evaluated on the same LA-ICP-MS system under the same experimental run conditions.

(106) FIG. 4A shows a representative H&E stained 5 micron tissue section from a block of a 19 year old human female with a brain neoplasm classified as glioblastoma multiforme. The H&E stained area shown in the Figure corresponds to the equivalent unstained area analysed by LA-ICP-MS on a different slide.

(107) The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe was determined and is shown for each voxel of three contiguous ablation tracks from a tissue section that ablated 68 voxels per track (FIG. 4B). The background-subtracted median values per voxel for these metals, expressed as calibrated counts per second, were found to be 3,198, 5,066 and 219,905 respectively. A graph of raw numerical values of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows from the human female with glioblastoma multiforme was prepared and is shown in FIG. 4C-FIG. 4E.

(108) To obtain a measure of the variation between different sections taken from the same block of the same human subject with glioblastoma multiforme, two unstained areas on the same slide were analysed by LA-ICP-MS in the same machine run. The results are shown in FIG. 5A-FIG. 5E.

(109) FIG. 5A shows a representative H&E stained 5 micron tissue section from a different region of the same block of the 19 year old human female with glioblastoma multiforme shown in FIG. 4A-FIG. 4C.

(110) The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe was determined for each voxel of three contiguous ablation tracks from a tissue section that ablated 69 voxels per track (FIG. 5B). The background-subtracted median values per voxel for these metals, expressed as calibrated counts per second, were found to be 2,980, 4,155, and 190,127 respectively. A graph of raw numerical values of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows from the human female with glioblastoma multiforme was prepared and is shown in FIG. 5C-FIG. 5E.

(111) These results demonstrate that variation in background-subtracted median calibrated counts per second in the same LA-ICP-MS machine run, for different tissue sections taken from the same block of the same individual is excellent. For .sup.55Mn the CC/S values were 3,198 and 2,980; for .sup.66Zn the CC/S values were 5,066 and 4,155, and for .sup.56Fe the CC/S values were 219,905 and 190,127 respectively.

(112) The next example is of a different neoplasm, which is also considered to be at the radiation resistant end of the clinical spectrum. This example is a neoplasm of mesothelial origin and comes from a 60 year old male with malignant mesothelioma of the abdominal cavity, and its characteristics are illustrated in FIG. 6A-FIG. 6F.

(113) FIG. 6A shows a representative H&E stained 5 micron tissue section from a 60 year old human male with malignant mesothelioma.

(114) The calibrated signal of each of four metals .sup.55Mn, .sup.66Zn, .sup.56Fe and .sup.63Cu was determined for each voxel of three contiguous ablation tracks from a tissue section that ablated 67 voxels per track (FIG. 6B). The background-subtracted median values per voxel for these metals, expressed as calibrated counts per second are 4,522, 8,805, 112,772 and 1,097 respectively. A graphical representation of the raw numerical values of calibrated counts per second for the four metals, .sup.55Mn, .sup.66Zn, .sup.56Fe and .sup.63Cu in adjacent voxels in each of three ablation rows from the 60 year old human male with malignant mesothelioma was determined and is shown in FIG. 6C.

(115) A further example of a different neoplasm, which is also considered to be at the radiation resistant end of the clinical spectrum, comes from a 50 year old male with malignant melanoma of the esophagus (stage IIa, T3N0M0). The metallomic characteristics are illustrated in FIG. 7A-FIG. 7E.

(116) FIG. 7A shows a representative H&E stained 5 micron tissue section from a 50 year old male with malignant melanoma of the esophagus.

(117) The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe was determined for each voxel of three contiguous ablation tracks from a tissue section that ablated 68 voxels per track (FIG. 7B). The background-subtracted median values per voxel for these metals, expressed as calibrated counts per second are 10,565, 6,961, and 121,495 respectively. A graphical representation of the raw numerical values of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe in adjacent voxels in each of three ablation rows from the 50 year old human male with malignant melanoma of the esophagus was determined as shown in FIG. 7C-FIG. 7E.

(118) It should be noted that there is much more variation in metallomic content between voxels in some melanomas than in normal tissues and in other neoplasms. One of the major contributors to this increase in variation is the presence of intracellular and extracellular inhomogeneities in the distribution of melanin, an entity that has storage capacity for different chemical elements, particularly metals. This morphological inhomogeneity is clearly visible in the first figure presented in this application, which is part of an H&E stained section from a stage Ib malignant melanoma of the neck of a 48 year old male (FIG. 1). Within the same tumour, there was an amelanotic lineage and a highly melanised different area. The metallomic characteristics of this tumour are further described herein (see Example 5).

(119) 2. Radiation Sensitive Tumours

(120) The characteristics of three different types of neoplasms that are considered to be at the radiation-sensitive end of the clinical spectrum will now be described. They are diffuse B-cell lymphomas, small cell cancers of the lung, and seminomas of the testis.

(121) Lymphomas: the first example of a neoplasm considered to be at the radiation sensitive end of the clinical spectrum, comes from a 57 year old male with diffuse B-cell lymphoma in the testis. Its metallomic characteristics are illustrated in FIG. 8A-FIG. 8E.

(122) FIG. 8A shows a representative H&E stained 5 micron tissue section from a 57 year old male with diffuse B-cell lymphoma.

(123) The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe was determined for each voxel of three contiguous ablation tracks from a tissue section that ablated 79 voxels per track (FIG. 8B). The background-subtracted median values per voxel for these metals, expressed as calibrated counts per second are 776, 18,892, and 247,923 respectively. A graphical representation of the raw numerical values of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe, in adjacent voxels in each of three ablation rows from the 57 year old male with diffuse B-cell lymphoma in the testis was prepared and is shown in FIG. 8C-FIG. 8E. The Y axis shows calibrated counts per second (CC/S) of that metal ion, while the X axis shows the voxel number of the three ablation tracks.

(124) Malignant carcinoma: The second example of a neoplasm considered to be at the radiation sensitive end of the clinical spectrum, comes from a 38 year old male with a small cell undifferentiated malignant carcinoma of the lung (stage IIIb, T4N1M0). Its metallomic characteristics are illustrated in FIG. 9A-FIG. 9E.

(125) FIG. 9A shows a representative H&E stained 5 micron tissue section from a 38 year old male with a small cell undifferentiated malignant carcinoma of the lung.

(126) The calibrated signal of each of three metals .sup.55Mn, .sup.66Zn and .sup.56Fe was determined for each voxel of three contiguous ablation tracks from a tissue section that ablated 65 voxels per track (FIG. 9B). The background-subtracted median values per voxel for these metals, expressed as calibrated counts per second are 1,749, 4,836, and 88,317 respectively. A graphical representation of the raw numerical values of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe, in adjacent voxels in each of three ablation rows from the 38 year old male with a small cell undifferentiated malignant carcinoma of the lung was prepared and is shown in FIG. 9C-FIG. 9E.

(127) Seminoma: the third example of a neoplasm considered to be at the radiation sensitive end of the clinical spectrum, comes from a 52 year old male with seminoma. Its metallomic characteristics are illustrated in FIG. 10A-FIG. 10E.

(128) FIG. 10A shows a representative H&E stained 5 micron tissue section from a 52 year old male with seminoma.

(129) The calibrated signal of each of three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe, was determined for each voxel of three contiguous ablation tracks from a tissue section that ablated 79 voxels per track (FIG. 10B). The background-subtracted median values per voxel for these metals, expressed as calibrated counts per second are 1,565, 13,528, and 52,115 respectively. A graphical representation of the raw numerical values of calibrated counts per second for the three metals, .sup.55Mn, .sup.66Zn and .sup.56Fe, in adjacent voxels in each of three ablation rows from the 52 year old male with seminoma was prepared and is shown in FIG. 10C.

(130) In toto, the above examples demonstrate the metallomic characteristics of three types of neoplasm that are at the radiation resistant end of the clinical spectrum, (glioblastoma, mesothelioma and melanoma), and three types of neoplasms that are at the radiation sensitive end of the clinical spectrum (B-cell lymphomas, small cell cancers of the lung and seminomas of the testis).

Example 4

(131) Application of the LA-ICP-MS Method to Patient Cohorts

(132) While the above examples are illustrative of the metallomic characteristics of single individuals in each category of cancer patients, the in-depth analysis of 55 individuals with seminoma, 10 with lymphoma, 20 with small cell lung cancer, 64 with melanoma, 25 with glioblastoma multiforme or astrocytoma, and 10 with mesothelioma, revealed a dichotomy in regards to their total manganese content and the known clinical outcomes of these cancer types to radiotherapy, was determined and is shown in FIG. 11A-FIG. 11H. The total median manganese free ion contents (expressed as calibrated counts per second, CC/S of .sup.55Mn), in tumours from patients with cancers of the testis (seminoma), lymphoma, small cell carcinoma of the lung, mesothelioma, brain (glioblastoma multiforme and astrocytoma) and melanoma, were determined and are shown. Each square represents a single patient, except in the case of two melanomas where two of the 64 patients each had two major lineages within their tumours, and each lineage was represented separately in the histogram with a square.

(133) The accepted clinical situation is that approximately 85% of seminomas, lymphomas and small cell carcinomas of the lung are sensitive to radiation, resulting in a great reduction, and sometimes complete elimination, of tumours and increasing life expectancy. Consistent with these values and as determined herein, 89% of the individuals with seminoma, 80% of those with small cell carcinoma of the lung and 90% of those with lymphoma, were found to have total tumour manganese contents that fall below 2,000 CC/S. In this respect, this is the first demonstration of the correlation of such radio-sensitive tumours having values that fall below a threshold of 2,000 CC/S.

(134) In contrast to the above, the three tumour types that are considered to be largely resistant to radiation (mesotheliomas, glioblastoma multiforme, astrocytomas and melanomas), and which have a proportion that are differentially sensitive to it, present a different outcome. Not only is the variance within each of these three tumour types greatly increased, but 90% of the mesotheliomas, 85% of the glioblastoma multiforme, astrocytomas, and 59% of the melanomas, fall above the 2,000 CC/S manganese value.

(135) Given the above dichotomy, the metallomic content e.g., the manganese content of tumours becomes a critical indicator for radiation therapy that has simply not been previously recognized in terms of practical clinical care until the present invention. In these examples using a 35 micron pixel, individuals having tumours with a low total manganese content (e.g., below 2,000 CC/S) are highly likely to benefit from radiation therapy, whereas those with a manganese content (e.g., significantly above 2,000 CC/S) are likely to benefit by avoiding radiation therapy. As pixel size is increased, the CC/S will increase, and hence a higher threshold is contemplated. As the range of values remains well within linearity, there are no statistical issues that are problematic.

(136) Given that the majority of cancer patients receive radiation therapy based on medical art, rather than quantitative data relevant to tumour characteristics, such data support the use of quantitative data using e.g., LA-ICP-MS to clinical practicality. Medical art relates to what percentage of primary cancers can be eradicated with radiotherapy alone and the dose that is required to do so, as well as the percentage susceptibility of various metastatic lesions to ablative radiotherapy. As a population-based approach it does not apply to the tumour(s) of a specific patient and is not personalized medicine.

(137) Statistical Analysis

(138) The two sample nonparametric Kolmogorov-Smirnov (K-S) test compares the cumulative distributions of two data sets. The null hypothesis is that both data sets were sampled from populations with the same distribution.

(139) From the data above, a comparison of the distribution of values for testicular cancer (seminoma) versus lymphoma yielded a D value of 0.2182, which has an associated non-significant P value of 0.762; seminoma and lymphoma were not significantly different and were sampled from populations with the same distribution.

(140) A comparison of the distribution of values for testicular cancer (seminoma) versus small cell lung yielded a D value of 0.3182, which has an associated non-significant P value of 0.081; seminoma and small cell lung were not significantly different and were sampled from populations with the same distribution.

(141) A comparison of the distribution of values for small cell lung versus lymphoma yielded a D value of 0.3000, which has an associated non-significant P value of 0.507; small cell lung and lymphoma were not significantly different and are sampled from populations with the same distribution.

(142) Conclusion: the data from seminoma, lymphoma and small cell lung represented a statistically valid unitary grouping.

(143) A comparison of the distribution of values for brain cancers (glioblastomas and astrocytomas) versus mesotheliomas yielded a D value of 0.3250, which has an associated non-significant P value of 0.371; brain cancers and mesotheliomas were not significantly different and were sampled from populations with the same distribution.

(144) A comparison of the distribution of values for brain cancers (glioblastomas and astrocytomas) versus melanomas yielded a D value of 0.2226, which has an associated non-significant P value of 0.277; brain cancers and melanomas were not significantly different and were sampled from populations with the same distribution.

(145) A comparison of the distribution of values for melanomas versus mesotheliomas yielded a D value of 0.4303, which has an associated non-significant P value of 0.056; melanomas and mesotheliomas were not significantly different and were sampled from populations with the same distribution.

(146) Conclusion: brain cancers, mesotheliomas and melanomas are a statistically valid unitary grouping.

(147) Comparison Between Groups

(148) A comparison of the distribution of values for testicular cancer (seminoma) versus brain cancers yielded a D value of 0.7712, which has an associated significant P value of 0.000; seminoma and brain cancer were significantly different and were sampled from populations with different distributions.

(149) A comparison of the distribution of values for testicular cancer (seminoma) versus mesothelioma yielded a D value of 0.8455, which has an associated significant P value of 0.000; seminoma and mesothelioma were significantly different and were sampled from populations with different distributions.

(150) A comparison of the distribution of values for testicular cancer (seminoma) versus melanoma yields a D value of 0.6879, which has an associated significant P value of 0.000; seminoma and melanoma are significantly different and are sampled from populations with different distributions.

(151) A comparison of the distribution of values for lymphoma versus brain cancers yields a D value of 0.7750, which has an associated significant P value of 0.000; lymphoma and brain cancer were significantly different and were sampled from populations with different distributions.

(152) A comparison of the distribution of values for lymphoma versus mesothelioma yields a D value of 0.9000, which has an associated significant P value of 0.000; lymphoma and mesothelioma were significantly different and were sampled from populations with different distributions.

(153) A comparison of the distribution of values for lymphoma versus melanoma yielded a D value of 0.6727, which has an associated significant P value of 0.000; lymphoma and melanoma are significantly different and are sampled from populations with different distributions.

(154) A comparison of the distribution of values for small cell lung versus brain cancers yielded a D value of 0.7333 which has an associated significant P value of 0.000; lymphoma and brain cancer were significantly different and were sampled from populations with different distributions.

(155) A comparison of the distribution of values for small cell lung versus mesothelioma yielded a D value of 0.9000, which has an associated significant P value of 0.000; lymphoma and mesothelioma were significantly different and were sampled from populations with different distributions.

(156) A comparison of the distribution of values for small cell lung versus melanoma yielded a D value of 0.6273, which has an associated significant P value of 0.000; lymphoma and melanoma were significantly different and were sampled from populations with different distributions.

(157) Conclusion:

(158) seminomas, lymphomas and small cell lung cancer were found to be a unitary grouping that is statistically distinct from the unitary grouping of brain cancers, mesotheliomas and melanomas.

(159) Whilst the statistical analysis indicates lymphoma, small cell lung, brain, and mesothelioma as being drawn from a normal distribution, such is not the case for the melanomas and the seminomas, neither of which conform to a normal distribution nor to a log normal distribution.

(160) A statistical analysis using the Kolmogorov-Smirnov test indicates that the distribution of melanoma values depicted in FIG. 11A-FIG. 11H neither conforms to a normal distribution, nor a log normal distribution. As shown in later examples, the melanomas exhibit a heterogeneity that is increased by intra- and extra-cellular deposits of melanin, a heterogeneous set of polymers which bind multiple metals, and by the presence of High Metallomic Regions (HMRs), distributed within many melanomas. These HMRs are a source of hidden atomic variation that has been uncovered for the first time owing to the technology described in this application. Such regions are undetectable by conventional pathological examination.

(161) It is likely that the seminomas in FIG. 11A-FIG. 11H consist of two populations, one with a small variance that makes up the bulk of the individuals and one that is more diverse and extends from 2K to 4K CC/S. It is probable that the bulk of the seminomas in FIG. 11A-FIG. 11H correspond to the clinical group that has an excellent response to radiation, whereas the remaining subgroup is relatively more radioresistant. This finding is consistent with clinical data (below). It should also be noted that the seminomas depicted in FIG. 11A-FIG. 11H are a unitary group, namely classic seminomas in a pathological context. Despite this outwardly appearing morphological homogeneity, the metallomic analysis has revealed an underlying heterogeneity that is hidden from conventional analyses.

(162) 1. Cancers of the Testis

(163) For tumours of the testis, the major division is relatively straightforward with testicular Germ Cell Tumours (GCT) falling into two large categories, seminoma and non-seminoma. For patients presenting with testicular cancer, approximately 50% are diagnosed with seminoma. Of these, approximately 85% present with stage I disease with the remainder being clinical stage IIA and IIB. (The distinction between stage IIA and stage IIB, is that lymph nodes are 2 cms in the former and 2 to 5 cms in the latter).

(164) Postoperative radiation treatment for testicular seminoma has been the mainstay of adjuvant therapy for more than half a century. Seminomas are one of the most sensitive tumour types to radiation, where a clinical trial of stage I testicular seminoma reveals that treatment doses of 20 Gray, given as 10 fractions over 2 weeks, is sufficient to lead to almost 98% cure rates at 5 years (Jones et al., Journal of clinical Oncology, 23, 1200-1208, 2005; Medical Research Council Trial TE18/European Organisation for the Research and Treatment of Cancer Trial 30942 (ISRCTN18525328)). This radiation dose compares to 60 Gy (and up to 90 Gy), for high grade gliomas, such as glioblastoma multiforme, where overall survival is increased only by months.

(165) The extreme radio-sensitivity of early stage seminoma is well described, with dose reductions being taken as low as 13 Gy for testicular intraepithelial neoplasia, which is a precursor to a more invasive form of cancer (Sedlmayer et al., Int J Radiat. Oncol. Biol. Phys. 50, 909-913, 2001; Classen et al., Br J Can. 88, 828-831, 2003). The European Germ Cell Consensus Group summarized its position in 2008, with 20 Gy in single doses of 2 Gy each being the recommended radiotherapy (Krege et al., Eur. Urol. 53, 478-496, 2008). The authors point out that for stage II testicular seminoma, a total dose of 30 Gy for stage IIA and a total dose of 36 Gy for stage IIB seems reasonable. The statement that these two dosages seem reasonable highlights the inexactness of the prior art (Krege et al., Eur. Urol. 53, 497-513, 2008), It should also be noted that although 36 Gy is the recommended dose, it has been pointed out earlier for stage II by Classen et al., J Clin Oncol 21, 1101-1106, 2003), that there is a potential for dose reduction. It should also be noted that relapse-free survival for stage IIA is 95%, and for stage IIB it is 89%. Overall survival is close to 100%.

(166) The American recommendations for stage I and stage II seminomas are in line with the European ones described above. Up to 2015, Mead et al., evaluated the evidence from a number of clinical trials Mead et al., Journal of the National Cancer Institute, 103, 241-249, 2011) and the recommendations for stage I seminoma are as above. For stage IIA seminomas, low dose radiotherapy to the paraaortic lymph nodes and superior ipsilateral pelvis is recommended with a total dose of 25 to 35 Gy in areas of gross adenopathy.

(167) The exquisite sensitivity of germ line tumours is put into perspective, by comparing the radio-sensitivity of testicular cancers with those of brain cancers.

(168) 2. Cancers of the Brain

(169) There is considerable variation in clinical practice for the management and treatment of adult gliomas, which in Australia and New Zealand occur with the following frequencies; 4% astrocytoma grade I; 10% astrocytoma grade II; 22% astrocytoma grade III; 52% glioblastoma multiforme grade IV; and with oligodendrogliomas and oligoastrocytomas making up the remainder (Cancer Council Australia/Australian Cancer Network/Clinical Oncological Society of Australia, ISBN 978-0-9775060-8-8; 2009). Of interest for this invention is glioblastoma multiforme, the most advanced of these gliomas, as it is considered by radiologists to be radiation resistant. In addition, the comparison of glioblastoma with stage III astrocytomas is revealing in terms of their metallomics, as against the clinical comparisons (see below). The standard radiotherapeutic treatment for high-grade astrocytomas is 60 Gy in 2 Gy fractions (sometimes with a 10 Gy boost), but there are no data in which the optimal dose has been determined for grade IV gliomas such as glioblastoma multiforme. For grade III gliomas many radiologists use 59.4 Gy with a fractionation procedure of 1.8 Gy, on the expectation that the 10% reduction per fraction from 2.0 Gy to 1.8 Gy in the case of grade III gliomas may cause less tissue damage.

(170) Two phase III clinical trials bear on these data; the Radiation Therapy Oncology Group (RTOG7401) and the Eastern Cooperative Oncology group (ECOG1374) (Nelson et al., NCI Monogr. 6, 279-284, 1988 and Chang et al., Cancer, 52, 997-1007, 1983). There were no survival differences between treatment arms (radiation; radiation plus radiation boost; radiation plus BCNU, and radiation plus CCNU) and little difference between anaplastic astrocytoma and glioblastoma. It is salient in the context of minor differences in survival outcome between anaplastic astrocytoma and glioblastoma, that the metallomic data for these categories are heterogeneous, their distributions overlap, and they are not significantly different from each other when tested using the Kolmogorov-Smirnov test (P=0.581).

(171) The central issue with radiation therapy to the whole brain is the occurrence of multiple complications such as neurocognitive problems, leukoencephalopathies and endocrinopathies. While some of these have been ameliorated with Involved Field Radiation Therapy (IFRT), the frequency of relapse has not altered, nor the percentage of patients with multilocal failures.

(172) There is clearly a need to better personalize therapies for patients, e.g., radiotherapy to those patients who are at the radiation sensitive end of the spectrum, while sparing those who are likely to be radiation resistant. The current one-size-fits-all therapy, e.g., radiation therapy, is obsolete, given the new knowledge provided by the date of this application including the metallomic data.

(173) 3. Cancers of the Serosal Membranes; Mesotheliomas

(174) Mesotheliomas are considered to be radiation resistant, and treatment protocols are of the order of a dose of 54 Gy delivered in 1.8 Gy fractions.

(175) The clinical situation with mesotheliomas is little different to that confronting physicians with other tumour types. In first examining surgically resectable tumours, there is the classic trimodality: a definitive surgical procedure, radiation therapy and systemic chemotherapy. In the case of mesotheliomas, the role of definitive surgery is controversial and it is unknown whether resection of a tumour yields an improvement in survival for that particular patient and no prospective clinical trials bear on this matter. There are also no adequately powered randomized phase III clinical trials that bear on the integration of radiation therapy and chemotherapy before and or after surgery, the closest being a phase II trial using a hemithoracic radiation therapy of 54 Gy following extrapleural pneumonectomy, (Rusch et al., J Thorac Cardiovasc Surg 122, 788-795, 2001).

(176) 4. Small Cell Lung Cancer

(177) Radiologists and physicians consider small cell lung cancer to be a radio-sensitive tumour with good response rates to radiotherapy. The European Society of Medical Oncology (ESMO) has documented its modified Tumour Node Metastasis classification and stage grouping, and released its Guidelines for diagnosis, treatment and follow up (Fruh et al., Annals of Oncology, 24, (Supplement 6), vi105, 2013; Table 3). The best overall survival outcome with curative intent is from a total dose of 45 Gy, with daily 1.5 Gy fractions, together with chemotherapy (Turrisi et al., N Engl J Med, 340, 265-271, 1999).

(178) 5. Lymphomas

(179) The Cancer Council Australia and the Australian Cancer Network, have set out clinical practice guidelines for the diagnosis and management of lymphoma, (ISBN: 0-9775060-0-2; 2005). For clinical stage I-Ill low grade follicular lymphoma it recommends involved-field radiation therapy of 30-36 Gy. For adult non-Hodgkin's lymphoma, the US National Cancer Institute recommends doses between 25 Gy and 50 Gy. For the diffuse B-cell lymphomas described herein, the National Comprehensive Cancer Network (NCCN.org; 2015) recommends 30-36 Gy after chemotherapy, 45-55 Gy as primary treatment for refractory to chemotherapy, or non-candidates for chemotherapy, and 30-40 Gy for salvage pre- or post-stem cell transplantation. In the varied spectrum that comprises non-Hodgkin's lymphoma, some patients have tumours that remain indolent for long periods of time, others evolve rapidly and require immediate treatment. As can be seen from the above in terms of radiation therapy, it is largely a one-size-fits-all situation.

(180) 6. Melanomas

(181) Melanomas have generally been considered to be a radioresistant tumour, but the data are conflicting, with much of the evidence deriving from cell lines that have demonstrated a wide spectrum of radio-responsiveness. The clinical practice guidelines for the management of melanoma in Australia and New Zealand have been set out in detail by the Cancer Council Australia/Australian Cancer Network/Ministry of Health New Zealand. For Australia see ISBN 978-0-9775060-7-1 and for New Zealand see ISBN (electronic) 978-1-877509-05-6. The clinical data have been reviewed by Wazer et al., UptoDate, 2015 and they too are conflicting. The clinical trials reviewed by Wazer et al., led them to conclude that melanoma is a radio-responsive tumour, but the optimal dose and fraction remain uncertain. In terms of metastatic disease, there are substantial differences in radio-responsiveness between cutaneous, lymph node, visceral metastases or metastases to bone or brain. (It should be noted that the conflicting clinical trial data make better scientific sense in terms of the metallomic data, where there is a wide range of manganese values between patients, and as revealed in FIG. 1, large differences within a single tumour). Depending on the site of melanoma, doses have varied from 32 Gy up to 100 Gy. Finally, while radiation therapy is useful in palliative care in the settings of bone, brain and visceral metastases, it is quite unclear whether large dose fractions improve palliation.

Example 5

(182) Melanin and Morphological Inhomogeneities and Radiation Sensitivity/Insensitivity

(183) As described above, the distribution of melanin, which has storage capacity for different chemical elements, particularly metals, can be a major contributor to variations seen with melanomas. FIG. 1 is an image of an H&E stained melanoma which shows this morphological non-homogeneity, which is clearly visible as two distinct lineages within it; one lineage is light coloured and largely amelanotic and the other is darkly staining and contains both intracellular and extracellular melanin. In addition, the cells of the amelanotic region are more uniform in terms of their nuclear sizes and morphologies, than cells in the heavily melanized regions, a finding that impinges on the ATI. The lower the variance of a population of cells in any of multiple characteristics, the more restricted will be its response to any perturbogen. The metallomic content of the top part of the same tissue section as depicted in FIG. 1 was measured. A two-dimensional representation of the ablation tracks taken through the melanoma is shown in FIG. 12. These data illustrate the variation in manganese level in different parts of the same tumour and emphasize the importance of characterizing a 2D landscape in order to make informed choices on whether a patient should be radiation treated, or not. The background-corrected median values for manganese in these visually-selected areas are 31,939 calibrated counts per second and 3,045 calibrated counts per second.

(184) The 2D landscape provides an estimate, within a tumour, of the proportion of a sampled area that exists above a designated threshold. In the above sample, this can be defined, for example, as to what proportion of the landscape falls below the 10.sup.th percentile, the 20.sup.th percentile, the 30.sup.th percentile, the 40.sup.th percentile the 50.sup.th percentile etc. and accordingly, quantitation is achieved. Thus, for example, if 90 percent of the area sampled is very low in manganese, then the tumour as a whole is likely to be susceptible to radiation. On the other hand, there is little point in irradiating a tumour in which 90% of the area is high in manganese, as radiation will leave most of the tumour intact. Note that these percentiles cannot be derived from a sample of a tumour that has been ground up for molecular work.

(185) The same tumour of FIG. 1 was analyzed further and the content of all 4 metals, .sup.55Mn, .sup.66Zn, .sup.56Fe and .sup.63Cu, was determined and shown in FIG. 13A-FIG. 13D. The dichotomy, as seen morphologically, is visible from these results. As the ablation travels from the left hand side with very high values, to the right hand side with far lower values, the inflexion point occurs around voxel 34.

(186) On the other hand, melanomas with uniform morphology, may still display very different metallomic content. For example, FIG. 14A-FIG. 14B shows an H&E stained tissue section of a melanoma from the left forefinger of a 54 year old male: stage IV, T4N0M1. This particular melanoma has a uniform morphology and the pathology would appear unremarkable to an experienced pathologist (FIG. 14A). The metallomic content was measured in this melanoma, and is shown in FIG. 14B. The underlying 2D metallomic content is very different. Indeed, there are 2 large concentrations of cells with very high levels of manganese.

(187) The median .sup.55Mn value for the bulk of this tumour sample was determined to be 817 background corrected calibrated counts per second, which places this tumor in the very sensitive end of the radio-responsiveness spectrum. On the other hand, the larger of the two whitish flat areas, namely the HMR, was found to have a value exceeding 6,500 calibrated counts per second, placing it at the radio-resistant end of the spectrum.

(188) These two regions are apparent from visual inspection of a voxel matrix. However a rigorous analysis of thresholds is necessary for an in-depth analysis of these two HMR regions and this analysis is illustrated in FIG. 15A-FIG. 15D, and FIG. 16A-FIG. 16C. As used herein, when referring to a specific metal in a High Metallomic Region (HMR), the HMR is designated as e.g., HMR(.sup.55Mn), HMR(.sup.66Zn), HMR(.sup.56Fe) and HMR(.sup.63Cu) respectively, with HMR(.sup.AM) referring to the generic case of Any Metal.

(189) FIG. 15A shows the uniform morphological nature of a tumour area of 2260 voxels (sampled from an approximately 1800 voxel area) as visualized by standard H&E pathology. FIG. 15B shows the two HMR(.sup.55Mn) regions as found by visual inspection of the numbers in each voxel. FIG. 15C shows the HMRs that emerge from this 2260 voxel 2D landscape when one uses as a threshold the median value per voxel of the whole area plus the standard deviation (St.Dev.) of the voxels of the whole area (median+St.Dev.). A different threshold method also robustly uncovers the same two regions in this 2260 voxel 2D landscape (FIG. 15D), this second method utilizes two times the median of the voxel values of the whole area (2 median).

(190) As seen by the distribution of clustered voxels that appear above threshold in FIG. 15C and FIG. 15D (black squares in a background of grey squares), the outcome of using (median plus St.Dev.) versus (2 median), is near identical in terms of finding HMRs(.sup.55Mn). Analysis of the six tumour types reveals that the 2 median method may be more consistent than the median plus St.Dev. method in detecting HMRs(.sup.55Mn), HMRs(.sup.66Zn), HMRs(.sup.56Fe) and HMRs(.sup.63Cu). Whilst it will be apparent to the skilled artisan that either method may be used according to the invention, by way of non-limiting example, the 2 median methodology has been used in subsequent analyses shown herein.

(191) A further important finding relating to the distribution of HMR(.sup.55Mn) and HMR(.sup.66Zn) in the same atomic landscape, is illustrated in FIG. 16A-FIG. 16C. FIG. 16A shows the identical H&E tumour landscape of FIG. 15A-FIG. 15D, FIG. 16B shows the two detected HMR(.sup.55Mn) areas, and FIG. 16C reveals the distribution of zinc in the identical voxels to those of the voxel matrix in FIG. 16B. In contrast to the two HMRs(.sup.55Mn), only a small region of HMR(.sup.66Zn) emerges above the 2 median of the zinc threshold. Zinc, iron and copper values generally remain at those of the bulk tumour area in regions of HMR(.sup.55Mn). This generalization does not hold in melanin-rich regions of melanoma tumours, where all four metals are found together in high concentrations in concert with melanin granules.

(192) It is acknowledged by the skilled addressee that there are a number of mathematical and statistical methods of arriving at thresholds that are useful in revealing regions of metallomic interest in tumours of most use for clinical decision making for radiotherapy.

(193) FIG. 17A-FIG. 17H provides an example of the extent to which HMRs(.sup.55Mn) occur in different tumour types, such as radiation sensitive versus radiation resistant ones. This comparison is of a 51 year old female patient with a small cell undifferentiated carcinoma of the lung, (Stage 1, T1N0M0) with the 54 year old patient who had a melanoma of the left forefinger (stage IV, T4N0M1). The region used herein for illustrative purposes corresponds to an approximately 1 mm by 1 mm sample of each of the two tumours, (31 by 31 voxels totaling 961 voxels). FIG. 17A-FIG. 17H reveal areas of .sup.55Mn concentrations using four different thresholds, (i) T1, above a threshold of 0.5 the median value, (ii) T2, above a threshold of 1.0 the median value, (iii) T3, above a threshold of 1.5 the median value, and (iv) T4, above a threshold of 2.0 the median value.

(194) In the case of the small cell carcinoma of the lung, the voxels that remain above the highest threshold of 2 the median value, T4, are generally non-contiguous ones (singletons) that are scattered throughout the sample. There are very few contiguous or adjacent concentrations of HMR(.sup.55Mn) voxels in this radiation sensitive tumour sample. In contrast, the four right hand side panels (FIG. 17E-FIG. 17H) from the melanoma patient reveal the emergence of contiguous HMR(.sup.55Mn) voxels that are readily discernible at the T3, 1.5 the median value, and these HMRs(.sup.55Mn) remain even at the more stringent T4 level of 2 the median value.

(195) The data of FIG. 18A-FIG. 18H show, by way of non-limiting example, that when the same threshold criteria are applied to the identical voxels of these two tumours, but now simultaneously assayed for HMRs(.sup.66Zn), no HMR(.sup.66Zn) areas are found at thresholds of T1, T2, T3 and T4 in the small cell carcinoma of this particular patient and only a minor concentration of contiguous Zn voxels is apparent in the tumour of the melanoma patient. The distribution of .sup.66Zn voxel values remains relatively constant in both these two tumours, while HMRs(.sup.55Mn) predominate in the melanoma, as shown in FIG. 17A-FIG. 17H.

(196) This comparison of two tumour types, one radiation sensitive and the other radiation resistant is further highlighted by way of non-limiting example, in the next seven Figures, (FIG. 19A through FIG. 25B). These illustrate the data from the 184 patients whose tumours are exemplary of the two spectral ends of radiation responsiveness: small cell lung, lymphoma and testis at the radio-sensitive end of the spectrum and mesothelioma, tumours of the brain and melanoma at the radio-resistant end of the spectrum.

(197) By way of non-limiting example, an overview of tumour regions with high metallomic concentrations and their clinical implications is provided below.

(198) FIG. 19A-FIG. 19B shows the .sup.55Mn voxel data from each of twenty patients with small cell carcinoma of the lung when a tumour area of approximately 1,800 voxels was sampled. The median .sup.55Mn value of the bulk of the tumour from each patient is represented by a square and is placed in bins of 200 consolidated counts per second over a range from zero CC/S to 7,000 CC/S. All twenty patients have tumours where the bulk .sup.55Mn values fall below 3,000 CC/S, (grey squares). Fourteen of the twenty patients, shown by the grey squares above the top line, have no HMRs(.sup.55Mn) when a T4 threshold of 2 median is used. By contrast, the tumours of six patients, t1 through t6, contain single HMRs(.sup.55Mn) (black squares connected to the grey squares by a dotted line). In a specific example, the bulk of the tumour of patient t6 has a median .sup.55Mn value that falls in the 2,400 to 2,600 bin, (grey square), but this tumour also contains an HMR(.sup.55Mn) whose value falls in the 6,400 to 6,600 bin (black square). The number of above threshold contiguous voxels of the six HMRs(.sup.55Mn) in patient samples t1 through t6 is 4.9%, 3.1%, 2.1%, 12.3%, 12.6% and 5.2% respectively of the sampled tumour areas. In three of these patients, t2, t5 and t6, the .sup.55Mn levels of their HMRs(.sup.55Mn) exceed 4,000 CC/S. Such regions are likely to contain cells that exhibit a degree of radio-resistance.

(199) By way of non-limiting example a minimum size of 88 voxels efficiently revealed regions of high metallomic content. A person skilled in the art will appreciate that below this size, and depending upon the sensitivity of the instrument used, smaller and smaller HMRs are picked up until finally they become indistinguishable from a randomly generated background of singletons, as seen in the small cell lung carcinomas of FIG. 17A-FIG. 17D. An 88 HMR-containing landscape is bounded in both the X and Y directions. It is contemplated that HMRs above X8Y8 voxels can individually increase by integers in the X and Y directions, yielding HMRs of X[8+1]Y[8]: X[8] Y[8+1]: X[8+1] Y[8+1]: X[8+2] Y[8]: X[8+2] Y[8+1]: X[8+2] Y[8+2]: X[8+1] Y[8+2]: X[8]Y[8+2] through X[8+n] Y[8+n] where n is an integer that can vary from 1 to thousands. Preferably, this integer varies from 1 to 100, more preferably from 1 to 22, when a sample of 3030 voxels (1 mm.sup.2 is sampled). A person skilled in the art will understand that there is no impediment to setting a threshold lower than a minimum of 8 voxels. The value of 8 has been used here because it is empirically-derived and serves as an example of an efficient search tool for HMRs.

(200) Sizes of HMRs

(201) An important descriptor of any HMR is its size, shape and content. Tumours of the six types from a total of 184 patients were analyzed. A person skilled in the art will appreciate that HMRs can be variable in size, shape and content and that defining a minimal area will depend upon machine jitter, and in some instances a single voxel may represent a rare, low frequency electronic spike in the data. In addition, machine stutter may occasionally produce two or three consecutive high values in the X or Y direction, but is readily recognizable by appropriate image filtering software.

(202) In view of the above, an HMR for any metal [HMR(.sup.AM), where .sup.AM=Any Metal] may contain two adjacent voxels to fulfill the criterion of voxel contiguity. FIG. 20A shows the eight different theoretical ways in which two adjacent voxels (a doublet), can be configured from a reference single voxel (starred). The size of HMR(.sup.AM) can vary from 2, 3, 4, 5, 6, 7, 8, 9, 10 etc contiguous voxels, to an integer which is the sample size chosen for analysis of that particular tumour. A person skilled in the art will appreciate that size of HMR(.sup.AM) according to the invention is in any range measurable using existing technology. For example, HMR(.sup.AM) size can range from 2 voxels to about 800,000 voxels and any value in between (this being the maximum number of 3535 micron pixels that can be analysed if a conventional slide is loaded to the edges with tumour material). A person skilled in the art will also appreciate that an HMR can apply to cancer cells in a tumour, or to a region of cellular and non-cellular material, generally referred to as associated stroma, which itself can exist in an activated state owing to its interaction with neighbouring cancer cells.

(203) For a clinical application, a single sample size is typically an area of 1 mm.sup.2, as such a size has a precedent in routine pathological analysis. Thus when pathologists examine a tumour section for mitotic rate, for example in melanoma, they have typically used the number of mitoses per high-power field, or per ten high-power fields (Burton et al., 2012, Am. J Surg. 204, 969-974). By way of non-limiting example, in respect of HMRs(.sup.AM), a 1 mm.sup.2 sample conveniently approximates 1,000 voxels, in this application. Multiple 1 mm.sup.2 samples may be laser ablated from a single tissue section on a glass slide. Multiple 1 mm.sup.2 samples may be laser ablated from multiple sections from the same tumour block. Multiple 1 mm.sup.2 samples may be laser ablated from multiple blocks representing different samples from the same tumour, or its metastatic derivatives. The cumulative HMR(.sup.AM) size determined from multiple samples can vary from 4 voxels (2 doublets) to many millions for a single tumour. It will be understood that the upper limit is constrained by the practicality of the time required for analysis and health care expense.

(204) Shapes of HMRs

(205) By way of non-limiting example, within a sample size of 1 mm.sup.2, or 1000 voxels of the 35355 micron type, there can exist a large number of HMR(.sup.AM) shapes that conform to the condition of voxel contiguity. As described herein for the first time, a common shape is that shown by example in the melanoma patient data of FIG. 14A-FIG. 14B and FIG. 15A-FIG. 15D, and represented stylistically in FIG. 20B-FIG. 20C. Patterns such as those depicted in FIG. 20D-FIG. 20H manifest themselves in different tumour types. For example, FIG. 20D is one in which cancerous epithelial cells line a lymphatic vessel or duct, such as in breast and prostate cancer. FIG. 20E and FIG. 20G would be representative of the classic Indian File movement of cells which follow a structural motif in a tumour such as a collagen bundle, or a major nerve tract in glioblastomas or astrocytomas, or the single file pattern seen in breast cancer sections. FIG. 20F illustrates the case in which cancerous cells would be contained within a blood or lymphatic vessel, such as in Ductal Carcinoma In Situ (DCIS) in breast cancer. FIG. 20H is representative of the distribution of cells and melanin deposits in some melanotic tumours.

(206) It is understood that any known mathematical and statistical methods may be used to analyze patterns and inhomogeneities in matrices and the associated software. Such methods may reveal concentrations of voxels of many shapes and sizes, which may then be mapped onto the underlying pathological landscape.

(207) Footprints of HMRs(.sup.AM) and their Value in Radio-Responsiveness

(208) In a sample of 1,000 voxels using a 2 median threshold, the HMRs(.sup.AM) (black squares in FIG. 20A-FIG. 20H), FIG. 20B, FIG. 20C, FIG. 20D, FIG. 20E, FIG. 20F and FIG. 20G represent 4%, 3.7%, 2.4%, 1.5%, 3.7%, and 0.9% of the sample.

(209) As shown in FIG. 19A-FIG. 19B, it is contemplated that if patients with tumours t2, t5 and t6 are treated with radiotherapy on either an intention-to-treat or palliative basis, the bulk of the cancer cells (with values <800, <1,800 and <2,600) would either be killed or sufficiently damaged as to be unable to undergo further cell division. By contrast, most of the cells of the HMRs(.sup.55Mn) of these tumours (with values >4,000, >4,200 and >6,400) would survive and, in time, generate a new form of the tumour, giving rise to the familiar clinical finding of tumour reoccurrence.

(210) HMR Data from Lymphomas, Tumours of the Testis, Mesotheliomas, the Brain and Melanomas

(211) FIG. 21 shows the data from the 10 patients with diffuse B cell lymphomas when a tumour area of approximately 1,800 voxels was sampled for .sup.55Mn content by the Laser Ablation technology. None of the tumours had HMRs(.sup.55Mn) when a threshold of 2 median was used, as shown by the absence of black squares. In addition, all median .sup.55Mn values fell below 3,200 CC/S. The expectation is that if the tumours of these 10 patients were treated with radiotherapy on an intention-to-treat or palliative basis, the bulk of the cells would either be killed or sufficiently damaged as to be unable to undergo further cell division. It is accepted from everyday clinical practice in this area that large lymphoma masses are easily eradicated with radiotherapy.

(212) FIG. 22A-FIG. 22B show the data from the 55 patients with tumours of the testis when a tumour area of approximately 1,800 voxels was sampled. These 55 tumours are all classic seminomas and form a unitary group on the basis of pathological criteria. Radiation oncologists and medical oncologists consider them to be very radiation sensitive, since large seminoma masses are easily eradicated with radiotherapy. Fifty one of these seminomas have no HMRs(.sup.55Mn) when a threshold of 2 median is used, as shown by the 51 grey squares above the top line. By contrast, four patients, S4, S5, S6 and S7, have single HMRs(.sup.55Mn) (black squares) relative to their bulk values (grey squares). The sizes of these four HMRs(.sup.55Mn) ranged from 1.8 to 6.3% of the sampled area. In two of these patients, S6 and S7, the CC/S value of the HMRs(.sup.55Mn) exceeds 5,000 CC/S and such regions are likely to contain cells that are radiation resistant. These data support the probability when these 2 patients are treated with radiotherapy on an intention-to-treat basis or palliative basis, the bulk of the cells are killed or sufficiently damaged as to be unable to undergo further cell division, while the cells of the HMRs(.sup.55Mn) survive and in time generate a new form of the tumour.

(213) Finally, as was found from the earlier statistical analysis of these seminomas, they are not a unitary group. It may well be that the tumours of the three seminoma patients, 51, S2 and S3, which populate the extreme right hand tail of the top distribution, and whose bulk median .sup.55Mn values are between 3,000 and 4,000, have a degree of radio-resistance. The data from these three outliers (and from patients S4, S5, S6 and S7, whose tumours do harbor HMRs(.sup.55Mn)), are consistent with the clinical data which reveal that a small proportion of seminomas exhibit signs of radiation resistance.

(214) FIG. 23A-FIG. 23B show the data from the 10 patients with mesothelioma when a tumour area of approximately 1,800 voxels was sampled (with mesothelioma considered to be a radiation resistant entity). Nine of the patients have no HMRs(.sup.55Mn) when a threshold of 2 median is used, as shown by the grey squares above the top line. The one patient with the lowest median CC/S voxel value has a single HMR(.sup.55Mn) whose CC/S exceeds 5,000 (black square) and the size of which was 3.2% of the sampled area. Importantly, most mesothelioma .sup.55Mn values exceed those of the three sensitive tumour types.

(215) FIG. 24A-FIG. 24B reveals the findings from the 25 patients with tumours of the brain (glioblastomas and astrocytomas) when a tumour area of approximately 1,800 voxels was sampled (with tumours of the brain considered to be largely radiation resistant). Twenty-four of the patients have no HMRs(.sup.55Mn) when a threshold of 2 median is used, as shown by the grey squares above the top line. One patient with a low median .sup.55Mn value has a single HMR(.sup.55Mn) whose CC/S exceeds 6,000 and the size of which was 3.6% of the sampled area. While most .sup.55Mn brain values exceed those of the three sensitive tumour types, these data indicate that some glioblastomas and astrocytomas will respond to radiation treatment, as eight of these patients have median .sup.55Mn values below 3,000 CC/S.

(216) FIG. 25A-FIG. 25B shows the data from the 64 melanoma patients, when a tumour area of approximately 1,800 voxels was sampled, with melanoma historically considered to be a radiation resistant tumour. The melanomas reveal extensive heterogeneity relative to the other five tumour types, with 29 patients having tumours without HMRs(.sup.55Mn), and 35 patients having tumours with HMRs(.sup.55Mn). Unlike the previous five tumour types where HMRs(.sup.55Mn) have existed as single entities within the tumour sample, half of the melanomas have multiple HMRs within the same tumour. The sizes of the HMRs(.sup.55Mn) vary from 1.6 to 21.8% of the sampled area. Owing to the stringency used herein for the threshold (2 median), the number of HMRs(.sup.55Mn) per sample is actually a minimal estimate. FIG. 25A-FIG. 25B also reveals that for nearly 30% of tumours, the median value of the bulk of a tumour falls below a CC/S threshold of 2K (grey squares) and for two thirds of the tumours, the median value of the majority of voxels falls below a CC/S threshold of 4K (grey squares).

(217) In hindsight, these atomic data indicate that many melanomas have areas of radio-sensitivity. In terms of metastatic disease, there are substantial differences in radio-responsiveness between cutaneous, lymph node, visceral metastases or metastases to bone or brain. The widely held opinion of melanomas being radiation resistant is based on conflicting data, particularly evident by the findings that some melanomas melt away after radiation treatment, while others can rapidly kill the patient in spite of radiation treatment. The metallomic data presented herein, show for the first time, how traditionally conflicting data can be partially resolved by the discovery of the previously unknown HMRs(.sup.55Mn) in melanoma tumours.

(218) FIG. 25A-FIG. 25B also shows that 88% of the melanoma HMRs(.sup.55Mn) exceed a CC/S value of 4,000, (black squares) and nearly half of the HMRs(.sup.55Mn) exceed values of 10,000 and some are near 100,000. These novel data indicate that many melanomas may indeed initially be radiation sensitive, but since most harbour HMRs which contain substantial levels of manganese, such cell populations would likely survive radiation treatment and the tumour would regenerate from radiation resistant remnants. The clinical outcome is that the patient returns post-radiotherapy with a reoccurrence of the cancer, which is now more resistant than the initial tumour. Second, many of the melanomas contain large quantities of extra- and intra-cellular melanin, and as our data show, again for the first time in tissue sections analyzed by laser ablation, melanins can contribute to radiation resistance by the biochemical processes outlined below.

(219) Melanins are heterogeneous polymers of uncertain 3D structure that form multilayered complexes consisting of overlapping sheets of dihydroxyindole and benxothiazine rings and sundry unidentified chemical groups (Zecca et al., Trends in Neurosciences, 26; 578-580, 2003). In the case of neuromelanin, there is also a large class of polyunsaturated lipids. Neuromelanins act as sinks for many metals including chromium, cobalt, mercury, lead, and cadmium, and significantly for this application, they also contain isotopes of Mn, Zn, Fe and Cu (FIG. 26). Melanins can be released extracellularly and, owing to their chemical characteristics, can survive in the extracellular milieu for long time periods. Melanins have radio-protective properties, attested to by melanized fungi living on the walls of the nuclear reactor at Chernobyl and in the cooling water of currently operational nuclear reactors. Melanins scatter X-rays and act as a shield for radiation. Despite the massive literature on the properties of isolated melanins, the first systematic attempt to explain its radio-protective properties was only made in 2007 (Dadachova et al., Pigment Cell Melanoma Res. 21; 192-199, 2007). Until the present invention, the physical explanation of melanin's radio-protective abilities at the level of pathological cancer material and their predictive relevance to patient treatment with radiotherapy has not been realized. Furthermore, most chemical and structural studies to date have been based on isolated melanins, not on their in situ molecular properties, and there has not been a 2D analysis of tissue sections from tumors as described herein. This is highlighted by the examination of melanin concentrations in a melanoma from a 45 year old female with a malignant melanoma of the chest wall (stage II, T4N0M0) (FIG. 27A-FIG. 27F).

(220) For clarity of presentation and for comparison of the concentrations of different entities, FIG. 27A and FIG. 27B are duplicates of the same H&E stained section. The distribution of melanin within this section is sufficiently high that staining by melanin specific antibodies has not been necessary to reveal it. The distribution of the four metals, .sup.66Zn, .sup.63Cu, .sup.56Fe and .sup.55Mn, is shown in the FIG. 27C-FIG. 27F, and can be visually compared to the distribution of melanin.

(221) FIG. 27C shows the 2D distribution of .sup.66Zn, the highest concentration of which (in black) co-localizes with the major melanin tracks. The bulk values for .sup.66Zn in the lightly stained areas are from 3,000 to 18,000 CC/S, while the .sup.66Zn levels within the melanin tracks are from 18,000 to 45,000.

(222) FIG. 27D shows the 2D distribution of .sup.63Cu (in black) which again closely tracks the melanin distribution. The bulk values for .sup.63Cu in the lightly stained areas are from 200 to 1,000 CC/S, while the .sup.63Cu levels within the melanin tracks are from 1,000 to 4,000.

(223) FIG. 27E shows the high concentration of .sup.56Fe (in black) which closely tracks the melanin distribution. The bulk values for .sup.56Fe in the lightly stained areas are from 20,000 to 80,000 CC/S, while the .sup.56Fe levels within the melanin tracks are from 80,000 to 100,000.

(224) FIG. 27F reveals the concentration of .sup.55Mn which closely follows the melanin tracks. The bulk values for .sup.55Mn in the lightly stained areas are from 1,000 to 2,500 CC/S, while the .sup.55Mn levels within the melanin tracks are from 2,500 to 10,000. As the earlier example of FIG. 12 showed, HMRs(.sup.55Mn) that consist of heavily melanized regions, can reach .sup.55Mn levels of 100,000. Melanin is thus a repository for high levels of metals that, if released, are likely to be a significant factor in radio-resistance. To our knowledge, this is the first demonstration of in situ .sup.55Mn concentrations in melanin in a 2D tissue section description of cancerous tissue, and the first quantitative predictions that follow from such co-localization of metals and melanin for patient treatment in radiotherapy. Melanomas have a dual defence to radiation. The first is the accumulation of .sup.55Mn in cells. The second is the production of large amounts of melanin, which not only act as a storage capacitor for .sup.55Mn, but also likely provide some structural shielding from ionizing radiation.

(225) The sizes and shapes of the melanin-rich voxel tracks in FIG. 27A-FIG. 27F can be compared with the sizes and shapes of some possible voxel aggregations as outlined in FIG. 20A-FIG. 20H, (particularly in FIG. 20H).

(226) The melanomas are instructive in a further clinical sense, since the data presented in FIG. 25A-FIG. 25B can be subdivided into melanoma samples that are either from primary tumours, or ones that are from metastatic sites. These data are shown in FIG. 28A-FIG. 28C. A comparison of the distribution of median CC/S values for the different patients with primary or metastatic melanomas yielded a D value of 0.1889 which has an associated non-significant P value of 0.557. Thus in terms of the median .sup.55Mn values derived from a melanoma, the primary and metastatic melanomas are consistent with being sampled from populations with the same distribution.

(227) A statistical comparison of the distribution of .sup.55Mn values from the primary tumours and specifically from the metastatic ones that have already spread to the lymph nodes, yielded a D value of 0.2222 which has an associated non-significant P value of 0.442. Thus median .sup.55Mn values from primary tumours and those that have been sampled from lymph nodes, are consistent with being sampled from populations with the same distribution.

(228) An examination of the distribution of tumours with HMRs(.sup.55Mn) and those without HMRs(.sup.55Mn) to determine if the frequency of HMRs(.sup.55Mn) differs between primary and metastatic tumours reveals that they do not, (yielding a non-significant P value of 0.457). Thus in terms of the distribution of median .sup.55Mn values, primary and metastatic tumours are indistinguishable. This finding is supported by there being no a priori biochemical set of processes that would favour differential .sup.55Mn accumulation between the primary tumour and its metastatic derivative(s) in lymph nodes.

(229) While .sup.55Mn levels and their distribution within a 2D area in tumour tissue are exemplified, another metal, zinc, may also be helpful by way of non-limiting example, as a discriminator in focussing on optimal regions for .sup.55Mn analysis. Many regions of a tumour contain cancerous cells intermixed with different cell lineages, as well as normal cellular and non-cellular components. The tumour milieu can contain fibroblasts, extracellular matrix components, collagen bundles, capillaries, lymphatics and support cells such as pericytes and smooth muscle cells, macrophages, osteoblasts, osteoclasts and components such as hydroxyapatite in bone niches. If in the first instance it is cancer cell populations that are chosen to be quantified for an ATI, then zinc can be used to differentiate between cells that are cancerous and cells that are not. Hence, zinc can provide a filter for ensuring that only the most relevant cancer cell-containing voxels are used for determining .sup.55Mn values and their distributions. (Note that zinc does not differentiate between radio-sensitive and radio-resistant cells.) Zinc voxel values allow an initial avoidance of regions that may mislead in identifying HMRs(.sup.55Mn). This is illustrated in FIG. 29A-FIG. 29B and FIG. 30A-FIG. 30B.

(230) FIG. 29A illustrates an H&E stained section from a 54 year old male with a small cell undifferentiated carcinoma of the lung (Stage IIIa; T2N2M0). Lightly stained regions of largely non-cancerous and non-cellular material are apparent from the H&E view. This lightly-staining region has low .sup.66Zn voxel values relative to the cancerous cells that are darkly staining. When an appropriate threshold is applied to the .sup.66Zn data, the resulting image in FIG. 29B is an excellent approximation to the differing regions seen in the H&E image above it. This means that the analysis of the .sup.55Mn values in this tumour sample is more appropriately based only on those .sup.66Zn voxels shown in FIG. 29B, without a contribution from voxels of the non-cancerous regions. This is not meant to downplay the clinical significance of stromal components, it serves simply to compartmentalize data analysis. This pre-screening of a tissue section landscape using .sup.66Zn, will be particularly useful in tumours such as those of the breast and prostate, where non-cancerous material is much more intimately intermingled with cancer cell-containing voxels. The reciprocal of the above is that zinc can also be useful in determining which parts of a tumour can best yield the ATI of a stromal population.

(231) A similar example is evident in FIG. 30A-FIG. 30B, in which FIG. 30A illustrates an H&E stained section from a 66 year old male with a malignant melanoma of the rectum (Stage IIB; T4N0M0). Non-cancerous cell regions are again evident, particularly in the lower portion of FIG. 30A at 6 o'clock. When an appropriate threshold is applied to the .sup.66Zn data, the resulting image in FIG. 30B mirrors the differing cellular components seen in the H&E image above it. This pre-screening with .sup.66Zn again shows that subsequent analysis of the .sup.55Mn values can be more accurately determined using only those .sup.66Zn voxels shown in FIG. 30B, without contamination from voxels of the non-cancerous/stromal regions.

(232) It will be appreciated by those skilled in the art that other atomic elements and their associated isotopes, besides .sup.66Zn, may provide the same useful function of differentiating between cancerous, normal, activated and non-cellular components of the stroma, particularly in different tissues.

Example 6

(233) Cancers of the Breast

(234) The data presented herein demonstrate for the first time that there is a correlation between .sup.55Mn levels and the two spectral ends of tumours that relate to radiation sensitivity and radiation resistance. The data presented herein also provide an insight into the previously unknown existence of HMRs that are hidden from conventional pathological examination and which play a major role in radiation resistance and an indication that a tumour may reoccur after radiation treatment. In addition, the data also provide a basis for metallomic contributions that derive from the interactions between stromal components and cancer cell populations.

(235) By way of non-limiting example, tumours that are loosely classified as being of intermediate radiation sensitivity, e.g., breast and prostate, were also analysed according to the invention.

(236) The same quantitative approach of Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry has been applied to 15 tumours of the breast, as was used for the six tumour types described herein. An overview of these data is presented in FIG. 31A-FIG. 31B. Examination of the bulk .sup.55Mn values of these breast tumours (grey squares), indicates values between 1,000 and 4,400 CC/S. At this level of inspection, they are indeed intermediate between the sensitive values found in the seminomas, small cell lung and lymphomas, and the higher .sup.55Mn values of the glioblastomas, mesotheliomas and melanomas. However, over 70% of these breast cancers have HMRs(.sup.55Mn) and their values extend from 4,000 to 16,000 CC/S (black squares in FIG. 31B). In addition, some tumours have double and triple HMRs(.sup.55Mn) using the stringent 88 detection threshold. Until the present invention, it was not realized that the atomic data are consistent with the clinical observation that breast cancers are very heterogeneous.

(237) In a personalized medicine context, the current medical art of describing breast cancers as heterogeneous is not helpful in deciding which patients will benefit from radiotherapy and which patients should avoid the harmful effects of radiation treatment. The clinical reality is that in the absence of a useful predictive metric, and in the presence of uniform H&E pathology, most breast cancer patients are irradiated after surgical resection of the tumour, e.g., in the USA, so that all possible treatment modalities have been seen to be applied. The resection margins for tumours of the breast can be large, so that the chance of surgically removing any residual cancer cells that may be at the periphery of a tumour is increased. While radiation treatment further increases the probability of destroying cells that have escaped resection, the harms of radiation could be avoided in those cases where the atomic data indicate very high levels of .sup.55Mn. In such high .sup.55Mn cases, radiation treatment in an intent-to-treat situation is largely futile.

(238) FIG. 32A illustrates a well differentiated invasive ductal carcinoma of the breast (Stage IIIa, T3N1M0) from a 39 year old female. The H&E section is unremarkable and uniform by pathological inspection. By contrast, the atomic data seen in FIG. 32B reveal the hidden heterogeneity of HMRs(.sup.55Mn) (large black areas) and their important clinical implications. This breast section has been examined as an approximately 1 mm.sup.2 area, to keep it in line with the standard sample size typically used in measuring mitotic rate.

(239) FIG. 33A-FIG. 33H shows the analysis of an approximately 1 mm.sup.2 area of 3131 voxels which was simultaneously analysed for the two atomic elements .sup.55Mn and .sup.66Zn using the standard thresholds applied previously (T1, 0.5 median; T2, 1 median; T3, 1.5 median and T4, 2 median). Two HMRs(.sup.55Mn) become evident in the Mn analysis using the T4 threshold, while the identical voxels for .sup.66Zn remain at their bulk tumour levels. It is .sup.55Mn levels that are the critical characteristic of these HMRs. A potential HMR(.sup.55Mn) that has a size below the 88 threshold, is visible above the other two in panels T3 and T4 (FIG. 33C and FIG. 33D).

(240) Further analysis of this approximately 1 mm.sup.2 area demonstrates the usefulness in analysing HMRs compared with bulk analyses of seemingly homogeneous pathological samples which pool data. Analysis of all 961 voxels in this sample in histogram form is shown in FIG. 34A-FIG. 34B for .sup.55Mn and .sup.66Zn values. In FIG. 34A, the heterogeneity in high .sup.55Mn levels in voxels is seen as the peaks and the smear of values that are in excess of 5000 CC/S values and which reach a value around 15,000. This heterogeneity can be examined in a rigorous statistical manner e.g., by an analysis of voxel contents using Bowley's Non Parametric Skew Statistic (NPSS) (Bowley, 1901, Elements of Statistics, P S King and Son, publishers, Westminster, London, UK). By way of non-limiting example, this was applied to the .sup.55Mn data and the NPSS value, [mean minus median]/[St.Dev] and found to be 0.25, whereas for the much more symmetrical .sup.66Zn data it was found to be 0.02, representing a large difference. It will be clear to a person skilled in the art that there are a multitude of statistical approaches to measuring heterogeneity, which may be used with the Skewness shown herein.

(241) The further area analysis revealed a distribution of the high .sup.55Mn values that consists of voxels spread uniformly throughout the sample, or their existence as aggregates. Finally, compared to the heterogeneity seen in .sup.55Mn levels, FIG. 34B shows that the identical Zn voxels were more uniform in their values. This is correlated by the lack of peaks or a smear to the right of the major peak in FIG. 34B.

(242) Cancerous Cells in Lymphatic Vessels/ducts in the Process of Metastasis

(243) In contrast to the example of the morphological homogeneity of the pathology seen in the invasive carcinoma described above, FIG. 35A illustrates the morphological heterogeneity of an invasive ductal carcinoma of the breast (Stage IIa, T2N0M0) from a 48 year old female as visualized by H&E. It is an informative sample from the perspective of the ATI, as this section contains, (i) an area of normal breast lobules at 3 o'clock (denoted N), with associated adipocytes, (ii) five areas denoted C1 through C5 in which cancer cells (denoted C), are visible inside lymphatic vessels and hence in the process of metastasizing, and (iii) an area of concentrated immune cells (denoted immune), together with a stromal component, including adipocytes spread throughout the middle and right hand side of the sample.

(244) The .sup.55Mn values of the normal lymphatic vessels, designated N, in the normal portion of the breast, yielded CC/S values of 3,204, 3,104 and 3,155. The lymphatic vessels that contain cancerous cells are informative. Lymphatic vessel C1 yields a value of 2,904, C2 and C3 yield 2,804 and C4 yields 2,771. Thus the .sup.55Mn contents of these transiting metastatic cells are little different from their progenitor cells constituting a normal duct of the breast. In contrast, the transiting cells in lymphatic vessel 5 have a median CC/S of 8612 and are predicted to be radio-resistant. This example illustrates the atomic microheterogeneity that occurs in a small single sample of the breast. This microheterogeneity cannot be extracted from the pathology of the cancer cells in C1 through C5 which are in the process of metastasis, since at the microscopic level they appear indistinguishable.

(245) The metallomic data reveal that all four metals, .sup.55Mn, .sup.66Zn, .sup.63Cu and .sup.56Fe readily resolve the areas that consist of adipocytes as against the cellular and acellular regions seen in the H&E stained sections (FIG. 36A-FIG. 36F). The metallomic contents of adipocytes are diluted by the large concentration of lipids in such cells.

(246) The heterogeneity issue, seen so clearly in this breast cancer section at both the morphological and metallomic levels, is also pervasive in other cancer types, even in different large regions within the same organ.

(247) There are a number of important issues that are taken into account by medical professionals regarding the treatment of a patient with cancer, including, but not limited to, the age of the patient; the current health condition of the patient; comorbidities; the locations of the tumour(s) (primary or metastatic); whether surgery, chemotherapy, drug treatment, immunotherapy or radiation is an option and whether the treatment is made as intent-to-cure, or palliative. This list exemplifies the clinical decision network that must be navigated to yield the best options. The central decision-node after blood tests and scans have been completed, is the pathology of the tumours. All other decisions follow from the information obtained at this node, since an inference is made as to whether the tumour is likely to be benign or likely to progress. Except in the case of measuring mitotic rate, until the present invention, the inference is currently not based on quantitative data, but on the general experience with a particular tumour type. For tumours that are relatively uncommon, there are usually few clinical trials that provide guidance on which therapeutic step is the best option. Even for common tumour types, such as those of the breast, prostate, brain, skin and ovary, the shortcomings in the multitude of clinical trials that do exist have lead to continuing controversies. The major problem is that the data gathered at this pivotal pathological-decision node of a particular tumour type are not of requisite quantitative quality, and until this application, are not generalizable to all tumour types.

(248) The clinical value of the crucial pathological information (namely a pointer to whether a major therapeutic option such as radiation should be implemented), rests largely on subjective interpretation of staining methodologies, combined with antibody information, or various forms of in situ hybridizations, which are further complemented by newer genomic and proteomic technologies. As reinforced by further examples below, the atomic data provide a new measure of quantitation that has hitherto been lacking at this key pathology/radiation-treatment, decision node. The Atomic Therapeutic Indicator of any tumour type provides for the first time, a quantitative underpinning of which patients are suitable candidates for radiotherapy, and in which patients a tumour is likely to reoccur after radiotherapy.

Example 7

(249) Cancer of the Prostate

(250) FIG. 37A-FIG. 37B shows a clinical example of the heterogeneity in pathological status in different parts of a tissue or organ, in this case the prostate gland. It is illustrated by a Final Prostate Biopsy Report from the USA (with the patients' full permission of disclosure). The patient (hereinafter denoted patient X), has adenocarcinoma of the prostrate with a PSA level of 8.1 ng/ml. The biopsy report documents the abnormalities in different regions of the prostate with 7 out of 12 regions showing changes of little significance, and five regions with cancer as defined by Gleason scores no higher than 7 (FIG. 37A). The diagnostic summary of the extent of involvement of each of the twelve core biopsies is also shown (FIG. 37B; Part A through Part L). Before the present invention, this patient would be assessed for radiation treatment without any quantitative measure that demonstrates such treatment would be warranted. This patient underwent radiation treatment on the basis of personal preference after discussions with his primary care physician, the other option offered to him being radical prostatectomy. In accordance with the present invention, the chemical distribution of elements within each of these regions is measured, e.g., the .sup.55Mn levels, 2D distributions and HMRs(.sup.AM), is determined by laser ablation. The results are analysed using available statistical methods and statistical theory, including calculating a measure of central tendency, where the common measures of central tendency are the median, arithmetic mean and mode. Any other measures of central tendency known in the art may also be used, including but not limited to geometric mean, medimean, winsorized k-times mean, K-times trimmed mean, and weighted mean, and the data may also be transformed prior to calculating a central tendency. For example, the data are also tested using non-parametric statistics, including the Kolmogorov-Smirnov test, which involves making no assumptions about the distribution of data. Other statistical methods include a combination of Bayesian and more standard statistics, clearly set out by Lee and co-authors in Demystify Statistical Significance-Time to Move on from the P value to Bayesian Analysis, (Lee, J Journal of the National Cancer Institute, 103, 2-3 2010) and Berry, Carlin, Lee & Muller. Bayesian Adaptive Methods for Clinical Trials. Chapman and Hall/CRC Biostatistics series. ISBN 9781439825488. 2010; Nuzzo, Nature, 506, 150-152. 2014).

(251) Other statistical methodologies include those set out in: Talfryn et al., British Medical Journal, 316, 989-991, 1998; Sterne & Smith, British Medical Journal, 322, 226-231, 2001; Bland & Altman, British Medical Journal, 328, 1073, 2004. In a clinical setting, other methods include: Rubinstein et al., Journal of the National Cancer Institute, 99, 1422-1423, 2007; Krzywinski & Altman, Nature Methods, 10, 1041-1042, 2013)

(252) Upon completion of the statistical analysis, a decision is made on the radio-sensitivity or radio-resistance of the different regions of the tumour and the most appropriate treatment. For example, a tumour that is deemed to be radio-sensitive at the bulk level and in its cancer cell-laden HMRs(.sup.55Mn) (if any), is irradiated. For one that is radio-resistant, or has large HMRs(.sup.55Mn), radiation treatment is not recommended. Selective radiation treatment is also recommended for specific areas of a tumour that is radio-sensitive. For example, any radiation treatment that selectively targets a specific area of a tumour is recommended. Such selective radiation treatments that are available and known in the art are contemplated. A person skilled in the art is able to readily determine which selective treatment is warranted.

(253) The same quantitative approach of LA-ICP-MS has been applied to 10 tumours of the prostate as used for the previous seven tumour types described so far. An overview of these data is presented in FIG. 38.

(254) Conclusion: No tumours had high HMRs(.sup.55Mn) and their voxel values are generally at the low end of the ATI spectrum (i.e. low range CC/S) and hence, from this sample, the tumours are expected to be mostly radiation sensitive.

Example 8

(255) Primary Melanoma of the Skin Metastatic to the Brain

(256) FIG. 39A-FIG. 39E and FIG. 40 reveal the detailed clinical data and their comparison to quantitative atomic data from a 70 year old patient (hereinafter denoted patient Y), who first presented with a primary melanoma and was subsequently diagnosed with a brain tumour. The patient was treated with whole brain radiation, sundry drugs, immunotherapy (prembrolizumab) and finally underwent a stereotactic craniotomy. This example illustrates the level of clinical detail that is required to match the atomic data, as well as the clinical baseline involved in placing an ATI into routine medical practice. The patient has provided full permission of disclosure.

(257) Patient Y was initially diagnosed with a primary melanoma of the upper back, with clear sentinel lymph nodes, as well as a number of basal cell carcinomas and squamous cell carcinomas. A number of years later, following a fit, MRI scans revealed a cerebral neoplasm of the left parietal lobe (FIG. 39A, arrowed). Following radiation treatment, MRI scans revealed the presence of residual material at the tumour site, with its associated ambiguities of interpretation (FIG. 39B).

(258) The detailed pathology report stated that the sections from the skin of the primary melanoma showed an ulcerated nodular melanoma with the tumour cells being positive for MelanA and negative for 34Be12, which is consistent with melanoma (FIG. 39C). The Breslow thickness was 3.8 millimeters; Clark level 4; ulceration, 3 millimeters; percent of dermal invasive tumour width, approximately 50%; dermal mitoses 9 per mm.sup.2 (which is considered to be a high mitotic rate; dominant cell types, naevoid and epithelioid; intravascular and lymphatic invasion was not seen and actinic/solar elastosis was mild.

(259) The melanotic area was widely excised from the upper right back and excisions of four sentinel nodes were carried out. Sections were treated with S100, HMB45 and MelanA to confirm the lesion as being a primary melanoma.

(260) While the primary ellipse revealed melanoma, no further tumour was evident in the wider excision. The adjacent epidermis did reveal reactive changes. Sentinel lymph nodes 1, 2, 3 and 4 revealed no evidence of further malignancy based on staining with H&E and immunoperoxidases, while the status of node 3 was not reported. Following the excision, the patient was declared clear of cancer and after 6 months showed no signs of recurrent melanoma.

(261) Three years later the patient collapsed but recovered, yet exhibited signs of incoordination of the left leg when walking. Similar episodes of involuntary movement of the leg recurred up to 20 times per day and initially they were accompanied by a strange sensation on the left side of the head. A cerebral MRI revealed a 14 mm contrast enhancing lesion in the paramedian left parietal lobe around the precentral gyrus (FIG. 39A-FIG. 39E). The radiologist gave a differential diagnosis of a low grade glioma, but it was possible that since the patient had a melanoma 3 years previously, that the cranial lesion could instead be a recurrent melanoma. This ambiguity is reflected by the limitations of MRI. The radiologist concluded that the nature of this lesion is uncertain. It is likely to represent a neoplasm of the brain which is low grade particularly as there is no mass effect or oedema associated with the lesion. The findings on the diffusion weighted scan of the low signal on the ADC (Apparent Diffusion Coefficient) map make it likely to be a low grade lesion such as a ganglioglioma, but also consider a DNET (Dysembryoplastic Neuroepithelial Tumor) or pleomorphic xanthoastrocytoma without any cystic component. A small oligodendroglioma remains a possibility.

(262) It was recommended that neurosurgery not be undertaken at the time, with stereotactic radiosurgery being the appropriate option. Given that the patient had been diagnosed with a primary melanoma, it was possible that the brain lesion was a metastatic melanoma to the brain. The patient underwent radiotherapy for assumed metastatic melanoma with the total delivered radiation dose being 25 Gy and also began an immunotherapeutic regimen course of pembrolizumab 2 months after radiation therapy. A further month later an MRI revealed that the midline frontal metastasis had diminished in diameter from 16 mm to 11 mm. At face value, the tumour had sensitive and resistant components. The tumour became better defined with a thinner enhancing margin and with a more discretely hypointense centre. Most of this reduction in tumour size is due to radiation and not the immunotherapeutic drug pembrolizumab, as tumour regression after one month of this drug treatment averages only 6%, (Hamid et al., 2013, New England Journal of Medicine, 369, 134-144). This patient's tumour reduced from diameters of 16 mm:16 mm:16 mm to 11 mm:11 mm:11 mm. This is a reduction of (5+5+5)/(16+16+16), 31%, of which 6% can be attributed to the drug, and 25% to radiation. This means that 80% (25/31) of the tumour's initial regression was due to radiation. A further MRI scan 3 months later revealed a blush of peripheral contrast enhancement in the white matter adjacent to the tumour nidus, suggesting tumour progress since the last examination. The tumour was still at 11 mm longest diameter 2 months later. An even later MRI revealed marked progression around the tumour which now measured 23 mm in the longest diameter (FIG. 39B), however it was not possible to tell if this was radio-necrosis or progressive tumour. The patient also began treatment with Avastin, and it was decided to proceed to craniotomy. Prior to craniotomy, a hyperintense centrally necrotic tumour was noted, approximately 1833 mm transaxial and 33 mm superior oblique, features in keeping with a high grade tumour. The patient underwent a stereotactic craniotomy to remove this left frontal lesion.

(263) The histopathology of the resected brain tumour revealed 2 pieces of tan and brown ragged friable soft tissue measuring 15 mm10 mm5 mm and 9 mm6 mm5 mm. There were very few and scattered devitalized melanoma cells with smudgy nuclei and no unequivocal evidence of residual viable malignancy in this material. The material was negative for the melanoma markers melanA and HMB45 and no definite pigment was seen. In the absence of marker confirmation, there is insufficient evidence to unequivocally state that this was a metastatic melanoma or an independent brain lesion, although the pathologist leaned towards a melanoma.

(264) Atomic Analysis of the Primary Tumour Biopsy

(265) As there were no biopsies of the brain lesion prior to radiation, the only available evidence for tumour identity and radio-responsiveness stems from the primary tumour (FIG. 39C-FIG. 39E). A standard unstained 5 micron tissue section was obtained (with all ethical, legal and patient consent issues fulfilled), and the entire area of FIG. 39C was laser ablated and then examined for the presence of nests of cancerous cells amongst the normal stromal heterogeneity that is visible in most of FIG. 39C.

(266) The numerical data obtained from the laser ablation are shown in FIG. 40 where five different regions of the primary tumour were analysed. The median .sup.55Mn value from all five tracks after background subtraction was 3,747 CC/S (ATI) placing the primary tumour in an intermediate range of radio-responsiveness. The clinical data are compatible with this value since the tumour was ablated with radiation.

(267) As noted above, the patient was also treated with an immunotherapeutic regimen of pembrolizumab which targets PD-1, so it is a combination of all these factors that has contributed to the final outcome. However, as pointed out above, 80% of the reduction in initial tumour size can be attributed to the radiation treatment.

Example 9

(268) Tumour Status after Irradiation of a Visible Tumour

(269) There is one source of tumours that are favourable for a more quantitative approach in terms of ATI and these are externally visible tumours whose status, progress, and condition after radiation treatment can be more readily measured than internal tumours. Patient Z is an example of such a case (FIG. 41A-FIG. 41B). For privacy reasons, this case serves only as an illustration of the directly measurable and unambiguous outcome of radiation of an externally visible tumour, and highlights the difference between the ambiguities experienced with the internal tumour issues experienced with the brain lesion of patient Y.

(270) Patient Z was radiation treated for a squamous cell carcinoma, (FIG. 41A), which after radiation treatment resolved almost completely within six months (FIG. 41B). Had a biopsy been available for atomic analysis, a comparison of the clinical response and the ATI would have been useful.

(271) Unlike other diagnostics which provide information on whether a patient has a particular tumour type, but do not provide the next therapeutic step tailored specifically for that patient, the ATI is applicable to all tumour types. ATI is pan-diagnostic. It is not restricted in the manner of a PSA test for example, where having obtained a value above 4 ng/ml, the next question is; what is the therapeutic intervention? Is it radical prostatectomy, radiation (external beam, brachytherapy, or proton beam), watchful waiting, cryotherapy, or androgen deprivation therapy? Unlike the ATI, the PSA test itself does not provide the therapeutic pointer.

(272) In the case of cancers of the breast, even if complete resection of a tumour with wide margins is carried out, followed by chemotherapy, radiation treatment, hormone therapy and drug treatment with Herceptin and/or Avastin and/or immune checkpoint inhibitors and/or immune agonists and/or vaccines, what is the probability that the tumour will reoccur if there is no information of the primary source as regards its potential radio-responsiveness? If the breast tumour biopsy had a sub-threshold ATI for example, then the probability of its reoccurrence after radiation would be lower than if the ATI of the biopsy was above threshold and if the tumour had one or multiple HMRs(.sup.55Mn), from which cells may have already migrated. If tumour cells have already spread to the nearby lymph nodes from the breast, then measurements of the ATI (high or low) in these nodes will provide a clinician with an indication of whether a more vigilant monitoring of the patient is required.

(273) Similarly in the BRAF.sup.V600E mutation in melanoma, and in many genomic tests, the presence of a targetable driver mutation is inferred from a cell sample or from circulating nucleic acids in the vasculature. This does not have the high value of a 2D visualization of the tumour landscape, since the former are a pooled group of entities. This is a critical differentiator between the use of an ATI and a pooled sample, where in the latter, it is impossible to tell whether a high reading derives from the output of a small group of cancer cells or activated stromal cells, or whether most cells in the sample contribute to the reading. The clinical implications for treatment are very different. A melanoma patient treated with vemurafenib, for example, who has a small number of cells in the tumour producing the altered protein, will hardly benefit from treatment, whereas the drug will be far more efficacious in a melanoma patient where a large number of cells in the tumour are producing a defective protein product. This distinction is difficult to make unless a 2D landscape is available.

Example 10

(274) Measurements Using Radiation Sensitizers/Synergizers

(275) .sup.10Boron

(276) The radio-responsiveness of a tumour is determined by measuring .sup.55Mn and its calibrated signals according to the invention, and the radio-responsiveness may also be influenced by the addition of a sensitizer. In such a case, the success of Boron Neutron Capture Therapy will depend both upon the total .sup.55Mn calibrated signal and that of the sensitizer. Adding a sensitizer such as p-boronophenylalanine to a tumour cell population that is high in manganese, may not be as useful as adding it to a cell population that is low in manganese. In this example, the ATI for a tumour is determined using LA-ICP-MS, before using Boron Neutron Capture Therapy.

(277) A tumour sample is taken from a patient who has been previously infused intravenously with an FDA approved sensitizer, e.g., a .sup.10Boron derivative, such as p-boronophenylalanine, or the intravenous infusion of liposomes containing boron derivatives as previously described (Heber et al., Proc. Natl. Acad, Sci. USA, 111, 16077-16081. 2014), or boron nanoparticles as previously described (Petersen et al., Anticancer Research, 28, 571-576, 2008). The tumour sample is then examined for the 2D distribution of .sup.10Boron to determine whether its levels and distribution will be beneficial in terms of radiation. Simultaneously, or separately, the distribution and level of Mn is determined. The relative amounts of .sup.55Mn and .sup.10B determines the suitable tumours of patients for radiation.

(278) Boron Neutron Capture Therapy (BNCT) is briefly described. A number of external entities are known to make tumours more sensitive to radiation e.g., Boron, .sup.10B. A thermal neutron is captured by the nucleus of .sup.10B and the ensuing fission reaction yields .sup.7Lithium as a recoil, an alpha particle, a weak gamma-ray (0.5 MeV gamma photon), and 2.4 MeV of kinetic energy. The .sup.7Li ion and the alpha particle are classified as high linear energy transfer radiation and are highly destructive.

(279) In studies in mice, subcutaneous injection of cells pre-incubated with boron nanoparticles into mice, which were then irradiated with neutron radiation, lead to longer survival, since the growth of tumours was delayed (presumably because the tumours with boron were made more sensitive to neutron irradiation) (Petersen et al., Anticancer Research, 28, 571-576, 2008). A clinical trial for the treatment of head and neck tumours has been initiated by the Boneca Corporation (ClinicalTrials.gov identifier; NCT00114790). In addition, a phase I/II clinical trial on Argentinian patients (with multiple subcutaneous metastases of melanoma), who have been treated with p-boronophenylalanine and neutron radiation yield an almost 70% response rate (Menendez P Appl. Rad. Isot. 67, (7-8 Suppl.)S50-S53). 2009). Boron Neutron Capture Therapy has also been used for non-small cell lung cancer (Farias et al., Phys. Med. 30,888-897. 2014).

(280) 2 Deoxy-D-Glucose, 2-DG

(281) Other useful radiosensitizers include, for example, those as summarized in Shenoy & Singh, Cancer Investigation 10, 533-551, 1992. These include 2 deoxy-D-glucose, 2-DG, which is a close analog of glucose but without the hydroxyl group in position 2.

(282) 2-DG is taken up avidly by those tumour cells that preferentially use glucose as available fuel, but upon phosphorylation by hexokinase, 2-DG is not further metabolized. Thus by competing with glucose uptake and subsequent steps, 2-DG causes metabolic stress and renders cells more sensitive to radiation. In cell lines exposed to ionizing radiation and simultaneously treated with 2-DG, radiation damage was increased in some, and the usual heterogeneity between cell lines was observed (Dawrkanath et al., Int. J Radiat. Oncol. Biol. Phys. 50, 1051-1061, 2001). It is thought that the radiosensitization of some tumour cell populations occurs via disturbances in thiol metabolism (Lin et al., Cancer Research, 63, 3413-3417, 2003).

(283) In early phase I/II clinical trials on patients with advanced brain tumours, the toxicity and feasibility of using 2-DG in combination with large fraction, 5 Gy, radiotherapy, (2-DG plus RT) was found to be well tolerated (Mohanti, B, Int. J. Radiat. Oncol, Biol. Phys, 35, 103-11, 1996).

(284) Treatment of glioblastoma multiforme patients with increasing oral doses of 2-DG and radiation revealed that up to 250 milligrams/kilogram of body weight was well tolerated with no significant damage to the normal brain. In addition, some of 60 patients in this cohort revealed median survivals that exceeded those of patients that only received radiotherapy (Singh et al., Strahentherapie and Onkologie, 8, 507-514, 2005). A summary of patient treatment and outcomes is found in Dwarakanath, J. Cancer Research and Therapeutics, 5, 21-26, 2009).

(285) Finally in nude mice with heterotropic pancreatic tumours, treatment with 2-DG plus radiation resulted in inhibition of tumour growth and increased survival, compared to controls (Coleman et al., Free Radical Biology and Medicine, 44, 322-331, 2008).

(286) Testing a tumour for .sup.55Mn (and any other relevant metallomic data), prior to radiation, is therefore another practical application of the technology of the invention to radiosensitizers, not just 2-DG.

(287) Immunotherapies

(288) Radiation, and the antitumour immune responses that follow, form an interacting system with an increased presentation of antigens on the surfaces of cancer cells and the release of a host of proteins and peptides (and metals bound to proteins and peptides), that influence the responses of antigen-presenting cells. Thus primary tumours that have been irradiated (or distant metastatic growths within the same individual that have not been irradiated), can become sensitized to attack by various immune cells. The mechanistic bases for this increased sensitivity are actively debated but remain unresolved, (Sharabi et al., 2015, Oncology [Williston Park] 2015, 29(5), pii:211304; Formenti, J Natl Cancer Inst. 105, 256-265, 2013). Melanomas that are highly resistant to radiation have high levels of melanin (which store a host of metals). In the context of the present invention, efficacy of immunotherapies is also addressed. Without being bound by any particular theory, the extensive and very different metabolic properties of radiation resistant cancer cells, versus sensitive ones, will not likely be treated equally by the immune system, either before or after radiation treatment. Thus radiation concomitant with immunotherapy; radiation preceding immunotherapy, or immunotherapy preceding radiation, will yield very different populations of cells within a tumour owing to differential selection. In this example, a tumour is first characterized by its ATI according to the method of any aspect, embodiment or example herein and a useful immunotherapy is then applied. Efficacy of the immunotherapy and/or when to administer radiation treatment, e.g., before, during or after immunotherapy is contemplated by performing the method of the invention according to any aspect, embodiment or example herein. Different types of immunotherapy are thought to interact differently with the same type of radiation. For example, Sipuleucel-T for castration-resistant prostate cancer (a dendritic cell vaccine designed to induce immunity against prostatic acid phosphatase), ipilimumab (anti-CTLA4) for unresectable metastatic melanoma, pembrolizumab and nivolumab (anti-PD-1) for melanoma, nivolumab for melanoma and advanced squamous non-small-cell lung cancer, and tremelimumab and lirilumab, are likely to produce different responses to radiation than immunotherapies involving chimeric antigen receptor T cells (CAR-T based immunotherapies). Checkpoint Blockade Immunotherapy in combination with stereotactic radiation delivery is underway for the treatment of glioblastoma, but glioblastoma is being treated as a unitary entity. Accordingly, any glioblastoma is characterized according to the method of any aspect, embodiment or example herein and an evaluation is provided for patients that best respond to immunotherapies.

(289) Rose Bengal and Melanoma

(290) Rose Bengal (4,5,6,7-tetrachloro-2, 4, 5, 7-tetraiodo-fluorescein) is an industrial chemical patented in 1882 that turns yarn and food red. When applied intra-lesionally to cutaneous melanomas, there can be significant shrinkage of some tumour(s) (Thompson et al., Melanoma Research, 18, 405-411, 2008; Thompson et al., Ann. Surg. Oncol. 22, 2135-2142, 2015). While treated skin lesions decreased in volume after Rose Bengal intra-lesional treatment, some of the distant untreated tumours in the same patient also shrank, indicating that an immune response was likely involved. The addition of radiotherapy (RT) to Rose Bengal (RB) treatment can further enhance tumour ablation, as shown for 3 patients who underwent both therapeutic modalities, RB plus RT (Foote et al., Melanoma Research, 2010, 20, 48-51, 2010). Note however, that radiation treatment of these patients was not based on any a priori knowledge of the radio-sensitivity or radio-resistance of their multiple tumours, as radioresponsiveness was not measurable prior to this application. In fact, it was noted that there is still no consensus on the optimal dose and fractionation in melanoma (Foote et al., Melanoma Research, 20, 48-51, 2010).

(291) Rose Bengal may be considered as an agent that has multiple modes of action: a sensitizer of cells to radiation, a sensitizer via augmentation of the immune system, an additive cell kill agent, or a synergizer. It is not possible to differentiate between these as the molecular mechanism(s) of these interactants is unknown, and the spatial arrangement of cancerous and stromal cells that have taken up RB remains unknown. The clinician is thus left with the difficult task of the management of melanoma patients, particularly those with regional metastatic regions, such as local, satellite and in-transit recurrence. The current treatment guidelines predominantly include surgical excision, local ablation, intra-lesional chemotherapy and targeted drugs such as vemurafenib. All of these are challenging due to disease heterogeneity and frequent and persistent proliferation of lesions (Thompson et al., 2015, Ann. Surg. Oncol. 22, 2135-2142, 2015).

(292) The relevance of the current application to Rose Bengal and therapeutic treatment options, is that the Rose Bengal molecule contains 4 iodine atoms that are easily measured in tissue sections by LA-ICP-MS. Thus a section of any tumour that has been intra-lesionally injected by Rose Bengal can be simultaneously analysed in each voxel for Iodine, .sup.55Mn or any other atom, prior to, and after radiation treatment, to determine which cell populations are susceptible to radiation. In this manner one can more precisely target susceptible lesions. There are no data as yet to determine the relative clinical efficacy of RB followed by radiation, or radiation followed by RB. What is clear is that any tumour that is accessible to intra-lesional injection of RB, can be analysed by the methods of this application to provide quantitative data on radiotherapeutic options.

(293) By way of non-limiting examples, injection of RB into the prostate via a multiple core needle approach, and simultaneous Laser Ablation analysis of Iodine and .sup.55Mn from tissue sections, will provide information on which cell types of normal, cancerous and stromal populations preferentially retain RB. These spatial distributions will enhance decision making as regards radiotherapy.

(294) Similarly, intra-lesional deposition of RB into breast tumour regions will allow simultaneous Laser Ablation analysis of Iodine and .sup.55Mn from tissue sections, and provide information on which cell types of normal, cancerous and stromal populations preferentially retain RB (as depicted for .sup.55Mn, .sup.66Zn, .sup.56Fe and .sup.63Cu in FIG. 35A-FIG. 35B and FIG. 36A-FIG. 36F in this application). These spatial distributions will enhance decision making as regards radiotherapy.

Example 11

(295) Clinical Implementation of the Atomic Therapeutic Indicator

(296) Analysis of the data from eight different tumour types has revealed a number of findings that place the technological and clinical aspects of ATI into perspective. Not unexpectedly, the melanomas are different to all other epithelial tumours, since no other tumour types synthesize melanin (unless it is a fortuitous activation of all pathways that culminate in melanin production in an unrelated cell type, or the result of cell fusion between immune cells and melanoma cells). In addition, melanomas derive from the initial embryological derivatives of neural crest cells, which are migratory cells that populate and set up different embryonic structures. Except for the initial migratory nature of germ cells, neural crest cells are the only other transitory cell type that migrates over long distances. As described herein for various melanotic tumours, melanins colocalize with the high concentrations of .sup.55Mn, .sup.66Zn, .sup.56Fe and .sup.63Cu. Of these, .sup.55Mn is most likely to provide radiation protection by its ability to bind O.sub.2., H.sub.2O.sub.2 and the highly dangerous hydroxyl radical OH. when .sup.55Mn is bound to various chemical entities (FIG. 2). In addition, high melanin concentrations throughout a tumour may provide a degree of physical shielding from radiation that is not available to any other tumour type. The only melanomas that can reasonably be placed with all other tumour types, are those melanomas that through the inactivation of steps leading to melanin production, are completely amelanotic. It is salient in this regard that the melanomas that have been examined herein, and which have unambiguously low concentrations of melanin pigment granules by visual microscopical examination (which is not quantitative), have low .sup.55Mn CC/S values (median voxel values of 1,939, 1,239, 817, 1,439, 1,278, 939, 1,617 and 1,678 CC/S). These all fall below the 2K ATI threshold illustrated in FIG. 11A-FIG. 11F. If this finding is subsequently confirmed quantitatively, then the amelanotic condition, in the absence of HMRs, may be indicative of a low .sup.55Mn ATI, and radiation sensitivity.

(297) In a clinical context, one modus operandi for the application of ATI to tumour biopsies are the flow diagrams shown in FIG. 42 and FIG. 43. In the former, easily accessed melanomas would be removed by surgery with wide excision as exemplified by patient Y. Analysis of voxel contents via Bowley's Non Parametric Skew Statistic (NPSS) (Bowley, 1901, Elements of Statistics, P S King and Son, publishers, Westminster, London, UK), would provide the first indication of homogeneity or lack thereof. Second, the conditions for homogeneity would be that, (i) the NPSS values for .sup.55Mn, .sup.66Zn, .sup.56Fe and .sup.63Cu would all be low, (ii) that the sample had no HMRs(.sup.55Mn) in either the cancer cell or stromal compartments and (iii) no overt melanisation. If the sample was found to be homogeneous, its ATI would be above or below a chosen threshold of clinical importance. As the primary tumour has been resected, such data on homogeneity and above or below thresholds, would provide information on the probable characteristics of any cells that had metastasized earlier, such as the cranial metastasis in patient Y. Analysis of resected lymph nodes from a patient who had undergone removal of the primary melanoma, would be analysed in the same manner as above. An ATI would provide guidance on whether to initiate radiation should a metastatic derivative of the primary tumour arise at a later time. If a biopsy from a metastatic site is available, then the flow diagram of FIG. 42 starts anew.

(298) The right hand panels of FIG. 42 illustrate the situation when a biopsy returns a sample that is heterogeneous in terms of voxel characteristics and when the tumour is resected. The heterogeneity will derive from a number of contributing factors, (i) heterogeneity within the main landscape, (ii) the presence, size and content of HMRs(.sup.55Mn) in both the cancer cell and stromal compartments, and (iii) the presence of melanin granules and the extent of intracellular melanin concentrations. The evaluation of all HMRs(.sup.55Mn), will indicate the probability of reoccurrence of a distant tumour (if any cells had metastasized early prior to the resection of the primary tumour).

(299) The flow diagram of FIG. 42 also applies to any biopsies that are taken from sites that are not amenable to excision by surgery, and where it needs to be ascertained by determining an ATI, if radiation is a useful option as regards intent-to-treat by radiation.

(300) For all other tumour types besides melanomas, the flow diagram of FIG. 43 will lead to a clinical decision point on whether to use radiation, or to spare it. The biopsy will reveal either homogeneity or heterogeneity (the latter being an unavoidable grey area in a clinical sense). The tumour will either be resectable or not, and radiation will be used if the ATI derived from the absence of HMRs(.sup.55Mn) is below threshold. If the tumour is heterogeneous and resectable, the ATI will be an indicator of what to expect if distant metastases arise at a later time in the patient's life. If the tumour is not resectable, then the decision to irradiate or not, is made on the basis of threshold values.

(301) It will be understood by the person skilled in the art that a number of other factors unrelated to the ATI will be considered by both the attending physician and the patient. These will include, but are not limited to, the age of the patient, the current health condition of the patient, comorbidities, the locations of the tumour(s) (primary or metastatic) and any hereditary conditions that make a patient radiation-sensitive. In the case of brain lesions, some tumours will be more radiation resistant than other tumour types, because the cancer cells use a novel mechanism of interconnection via microtubes, which allows damaged cancer cells to be repaired by others within the tumour (Osswald et al., Nature, 528, 93-98, 2015).

Example 12

(302) Clinical Implementation of the Atomic Therapeutic Indicator in Combination with Other Entities

(303) The key foundation of this application is that focused pulses of high irradiance laser energy applied across a tissue section, and the analysis of the vaporized material via mass spectrometry, provide a 2D spatial atomic map which is of immediate therapeutic importance as regards radiation treatment of cancer patients via an ATI. This ATI/H&E map is the foundation onto which other different maps can be superimposed. A person skilled in the art will recognize that a judicious choice of biological entities providing multilayered/superimposed information, will further increase the clinical impact of an ATI, a situation that was not available prior to this application. We demonstrate below how integrating additional maps employing metal-labelled antibodies using elemental analysis, namely laser ablation-Inductively coupled plasma-mass spectrometry (LA-ICP-MS) or laser ablation-time-of-flight-mass spectrometry (LA-TOF-MS) or inductively coupled plasma-optical emission spectroscopy (ICP-OES), or microwave plasma-atomic emission spectroscopy (MP-AES), or laser induced break down spectroscopy (LIBS), or secondary ion mass spectrometry (SIMS), or X-ray absorption near edge structure (XANES), atomic absorption spectroscopy (AA) or X-ray fluorescence (XRF), can increase the power of the clinical decision making process.

(304) Additional Maps

(305) Gene expression can be measured in tissue sections via spatial transcriptomics using arrayed reverse-transcription oligo (dT) primers and fluorescently labelled nucleotides (Stahl et al., Science, 353, 78-82, 2016). This provides a spatial map of gene expression relative to the H&E map of the tissue section, but it comes at the cost of being labour intensive, involving library construction, amplification steps, intensity loss of fluorescence with time, staining artefacts and auto-fluorescence as well as deconvolution of large data sets of Entities of Unknown Clinical Significance. To our knowledge, no previous spatial gene expression maps of tissue sections have reported on the clinical question of whether radiation is a preferred treatment option for a patient.

(306) What has not been available, until the present application, is a clinically useful 2D map generated from a simultaneous measurement of an ATI together with specially selected biological parameters at the protein or cellular level. These parameters need to have a presumed involvement in radiotherapeutics and need to be immediately implementable with current pathological and molecular technologies. A number of instantiations of such maps are provided below.

(307) Many tumours are claimed to contain cancer stem cells (CSCs) (Clevers, Nature Medicine, 17, 313-319, 2011), that are virtually resistant to radiation (Ogawa et al., 2013, Anticancer Research, 33, 747-754). This radio-resistance of CSCs is thought to be due to a number of factors including their superior DNA repair capabilities and their heightened defence to Reactive Oxygen Species. Such CSCs are thought to self-renew, divide slowly and are capable of reconstituting a tumour. If such is indeed the case, then it would be clinically advantageous to construct and to superimpose a properties of cancer stem cells map, onto the ATI/H&E map. This can be done using metal-labelled antibodies.

(308) Current technology on formalin-fixed paraffin-embedded tissue sections generally employs antibodies to the antigens of interest, but to multiplex several protein tumour markers, say 4 or 5, which may co-localize on the same tissue section, is near impossible using current immunohistochemistry. Use of a primary antibody which is antigen specific, is followed by an amplification step which involves an enzyme labelled secondary antibody. The time factor of staining 5 sequential tissue sections, processed at different times and different staining conditions, is neither conducive to rapid and accurate throughput, nor to interpretation. However, application of antibody labelling using metals (especially lanthanides and their easily distinguished isotopes), means that different antibodies, each tagged with a different isotope, can be applied to the same tissue section which is then directly examined via LA-ICP-MS, (Giesen et al., 2011, Anal. Chem. 83, 8177-8183). Here there is no ambiguity with co-localized staining procedures, fluorescence issues or quantification. This methodology has been applied in a diagnostic context to the labelling of primary antibodies, anti-Her2, anti-CK-7 and anti-MUC 1 using the lanthanides holmium, thulium and terbium and their subsequent location in breast cancer tissue sections examined via LA-ICP-MS, (Giesen et al., 2011, Anal. Chem. 83, 8177-8183). It has also been applied to directly label anti-tyrosine hydroxylase (TH) with Ytterbium-173, Paul et al., 2015, Chemical Science, DOI:10.1039/c5sc02231b, 2015).

(309) The above data demonstrate that multiple lanthanide-labelled antibodies analysed via elemental analysis can report on the spatial distribution of antigens in the same tissue section and provide clinical information of use for drug-based patient treatment. In the context of radiotherapeutic information, however, the requirement is different. It is to measure an ATI and entities of radiotherapeutic significance in the same tissue section, or in sequential serial sections (for example in the breast cancer section exemplified in FIG. 35A-FIG. 35B, where there are cancerous cells in lymphatic vessels, and where those metastasizing cells have different .sup.55Mn levels). This can be implemented as follows.

(310) It is known that the 15 lanthanides; Lanthanum, Cerium, Praseodymium, Neodymium, Promethium, Samarium, Europium, Gadolinium, Terbium, Dysprosium, Holmium, Erbium, Thulium, Ytterbium and Lutetium, together with Scandium and Yttrium, have multiple isotopes that are readily distinguished by LA-ICP-MS. Hence instead of 17 elements, one has a minimal palette of around 32 labels from which to choose to label antibodies of choice. In the context of this application, the next step is to select those antigens that are markers for stemness, DNA repair, ROS, cell division and methylases and demethylases involved in silencing of tumour suppressors in cancer cells, and all of which impinge on radiotherapeutic relevance.

(311) In a non-limiting context, the first steps involve:

(312) (i) selecting proteins involved in sternness networks that are relevant to tumour aggressiveness/metastatic potential, such as CD44, (breast, liver and pancreas), CD133 (brain, colorectal, lung, liver), EpCAM, (colorectal and pancreatic) as set out by Clevers (Nature Medicine, 17, 313-319, 2011), as well as proteins that are targeted by drugs aimed at cancer stem cells such as NOTCH, DLL4, FAK, STAT3, and NANOG (Kaiser, Science, 347, 226-229, 2015).
(ii) selecting proteins involved in DNA repair, such as BRCA1, BRCA2 and ATM, as reviewed by Wood et al. (2001, Science 291, 1284-1289, and subsequent updates).
(iii) selecting proteins involved in metabolic networks of Reactive Oxygen Species which have a major role in the tumour niche, such as Hypoxia Inducible Factors, HIF (Simon and Keith, Nature Reviews Molecular Cell Biology, 9, 285-296, 2008), carbonic anhydrase IX and catalase.
(iv) selecting proteins involved in cell division networks, such as Ki-67 (Inwald et al., Breast Cancer Res Treat. 139, 539-552, 2013).
(v) selecting proteins such as the DNA demethylase TET1 and TET 3 (Forloni et al., Cell Reports, 16, 1-15, 2016), and DNA methylases such as the DNMT1, DNMT3a and DNMT3b that silence or activate genomic regions and influence the oncogenic potential of cells within tumours.

(313) The second step involves using primary antibodies to these protein products and then labelling either primary or secondary antibodies with a suitable lanthanide and hybridizing these antibodies to tissue sections.

(314) The third step involves the application of elemental analysis to such an antibody-rich section. This provides a simultaneous readout of the endogenous metals that have been used in this Application, Mn, Zn, Fe and Cu, to generate an ATI and to locate HMRs, plus a spatial readout of the various lanthanides that highlight the positions in the ATI/H&E map of the relevant proteins to which the lanthanides are attached. This is multiplexed rapid cartography of clinical importance. When applied to a tissue section such as that of FIG. 35A-FIG. 35B, it will allow for a more complete understanding of the characteristics of the cell populations in lymphatic vessels and capillaries of the breast (and any other tumour), the heterogeneity within primary and secondary tumours, and in the case of core needle biopsies of the prostate, an expansive view that far exceeds that of current pathological insights.

(315) This logical extension of superimposed multilayered information onto an ATI/H&E map, provides a new pathological taxonomy, not previously seen, that produces rapid, quantitative, clinically relevant information from the same or serial tissue sections which can be evaluated in a radiotherapeutic context.

(316) Advantages of the Present Invention

(317) The 2D cartographic nature of the present invention has a number of significant attributes that are relevant to the clinical radiotherapeutic treatment that can be undertaken compared to the non-cartographic alternative of homogenizing a tumour and measuring the amount of manganese per unit mass of tissue. First, all tumours are heterogeneous in terms of the amount and type of stromal material, so the relative amount of tumour cells to stromal material, which is only visible and measurable with reference to a 2D map, will influence the median ATI. Second, even when a 2D region consists of nearly all cancer cells, they can be very different in terms of their Mn levels. FIG. 1 illustrates this perfectly, where a third or so of the section consists of amelanotic tumour cells with median CC/S values around 3,000, and the rest of the section consists of areas where the Mn levels vary between 15,000-45,000 CC/S and then HMRs(.sup.55Mn) that vary between 45,000-150,000 CC/S. If one had homogenized this section, the median ATI would far exceed 10,000 CC/S. What would have been lost in a clinical sense, is that radiation treatment of this patient would have killed the third of the tumour that has a low ATI, and thus the tumour burden would have been reduced, with clinical benefit to the patient. Even more importantly, homogenizing a section, or a tumour sample, loses one of the important aspects of the invention: HMRs(.sup.55Mn) that are predictive of tumour reoccurrence. Most HMRs represent less than 10% of a section, so homogenization loses that information of the presence of an HMR(.sup.55Mn), as the median ATI is barely moved.

(318) In the example of breast cancer, FIG. 35, with its mixture of normal breast ducts, an area of immune cells, adipocytes and stromal cells, the clinically relevant cells that are most dangerous are the ones that are already in only one lymphatic vessel C5. They are in the process of metastasis and they have a median of 8,612 CC/S. If the entire section were homogenized, this clinically relevant information is completely lost, as the median of the homogenized section is now approximately 1,669 CC/S, a low value that is influenced by the proportion of the adipocyte population, the stroma and the relatively low values in the normal ducts.