PROTEIN SIGNATURE FOR SCREENING GENERAL POPULATION FOR COLORECTAL CANCER AND/OR PRE-CANCEROUS STAGE THEREOF

Abstract

The present invention refers to an in vitro method for the diagnosis of colorectal cancer and/or pre-cancerous stage thereof.

Claims

1. An in vitro method for screening for colorectal cancer or a precancerous stage thereof comprising: a) measuring the concentration level of at least protein TFF3 (Trefoil factor 3), in a plasma sample obtained from a subject, b) determining if the concentration level of TFF3 is statistically higher than a reference concentration level of TFF3 measured in healthy control subjects, and c) identifying said subject as one having colorectal cancer or a precancerous stage thereof when the measured concentration level of TFF3 is statistically higher than said measured reference concentration level.

2-13. (canceled)

14. The in vitro method according to claim 1, comprising: a) measuring the concentration levels of at least proteins TFF3 and Flt3L (Fms-related tyrosine kinase 3 ligand), in plasma sample obtained from the subject, b) determining if the concentration levels of TFF3 and Flt3L are statistically higher than the concentration levels of TFF3 and Flt3L measured in healthy control subjects, and c) identifying said subject as one having colorectal cancer or a precancerous stage thereof when the measured concentration levels of TFF3 and Flt3L are statistically higher than said measured reference concentration levels.

15. The in vitro method, according to claim 1, comprising: a) measuring the concentration level of the proteins from at least one of the following groups of proteins: TFF3, Flt3L and HGFR; TFF3, Flt3L and IGFBP2; TFF3, Flt3L and CD147; TFF3, Flt3L and CD163; TFF3, Flt3L and CYFRA21-1; TFF3, Flt3L and ADAMDEC1; TFF3, Flt3L and FASL; TFF3, Flt3L, HGFR and IGFBP2; TFF3, Flt3L, HGFR and CD147; TFF3, Flt3L, IGFBP2 and CD147; TFF3, Flt3L, CD163 and IGFBP2; TFF3, Flt3L, CD163 and HGFR; TFF3, Flt3L, CD163 and CD147; TFF3, Flt3L, CYFRA21-1 and CD147; TFF3, Flt3L, CYFRA21-1 and IGFBP2; TFF3, Flt3L, CD163, and CYFRA21-1; TFF3, Flt3L, HGFR, and CYFRA21-1; or TFF3, Flt3L, CYFRA21-1, and CEA in a plasma sample obtained from the subject; b) determining if the concentration levels of the proteins in any one or more of the groups of proteins in a) is statistically higher than the concentration levels of the same groups of proteins measured in healthy control subjects; and c) identifying said subject as one having colorectal cancer or a precancerous stage thereof when the concentration levels of the proteins measured in any one or more of the groups of proteins in a) are statistically higher than said measured reference concentration levels.

16. The in vitro method according to claim 3, further comprising processing the measured concentration values of the proteins to obtain a risk score and identifying said subject as one having colorectal cancer or a precancerous stage thereof when the concentration level of the proteins measured are statistically higher than said measured reference concentration levels.

17. The in vitro method according to claim 1, wherein the pre-cancerous stage of colorectal cancer is advanced colorectal adenoma.

18. The in vitro method according to claim 1, further comprising confirming the identification of said subject as one having colorectal cancer or a precancerous stage thereof by an imaging technique.

19. A kit for screening for colorectal cancer or a pre-cancerous stage thereof comprising: a) reagents or tools suitable for obtaining a plasma sample from a subject, and b) reagents or tools suitable for determining the concentration level of TFF3 (Trefoil factor 3) in said plasma sample.

20. The kit according to claim 19, further comprising reagents or tools suitable for determining the concentration level of Flt3L (Fms-related tyrosine kinase 3 ligand) in said plasma sample.

21. The kit according to claim 19, comprising reagents or tools for determining the concentration levels of the proteins from at least one of the following groups of proteins: TFF3, Flt3L and HGFR; TFF3, Flt3L and IGFBP2; TFF3, Flt3L and CD147; TFF3, Flt3L and CD163; TFF3, Flt3L and CYFRA21-1; TFF3, Flt3L and ADAMDEC1; TFF3, Flt3L and FASL; TFF3, Flt3L, HGFR and IGFBP2; TFF3, Flt3L, HGFR and CD147; TFF3, Flt3L, IGFBP2 and CD147; TFF3, Flt3L, CD163 and IGFBP2; TFF3, Flt3L, CD163 and HGFR; TFF3, Flt3L, CD163 and CD147; TFF3, Flt3L, CYFRA21-1 and CD147; TFF3, Flt3L, CYFRA21-1 and IGFBP2; TFF3, Flt3L, CD163, and CYFRA21-1; TFF3, Flt3L, HGFR, and CYFRA21-1; or TFF3, Flt3L, CYFRA21-1, and CEA in a plasma sample obtained from the subject.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0037] FIG. 1. Receiver-operating-characteristic (ROC) curve in colorectal cancer for: A) [TFF3 and Flt3L]. Area Under Curve (AUC)=0.8672. B) [TFF3 and Flt3L and HGFR]. Area Under Curve (AUC)=0.8672. X axis represents Specificity. Y axis represents Sensitivity.

[0038] FIG. 2. Receiver-operating-characteristic (ROC) curve in advanced colorectal adenoma for [TFF3/Flt3L/CD163/ADAMDEC1]. Area Under Curve (AUC)=0.6815.

[0039] X axis represents Specificity. Y axis represents Sensitivity.

DETAILED DESCRIPTION OF THE INVENTION

Example 1. Material and Methods

Example 1.1. Population of Study

[0040] The population of study was Spanish screening population. This means asymptomatic average risk subjects between age 50-75 referred to colonoscopy by a population screening program. Due to the low incidence of CRC in screening population, some CRC cases are patients already diagnosed with CRC that were scheduled for surgery (blood samples were obtained before colonoscopy or surgery resection). Subjects who have developed another type of cancer in the 5 years prior to their participation in the study or patients who have previously received chemotherapy or radiotherapy, or patients diagnosed with Non-Advanced Adenomas, Familiar Adenomatous Polyposis or Lynch Syndrome, Inflammatory Bowel Disease, or inadequate intestinal preparation for colonoscopy or subjects who have undergone colonoscopy/polypectomy in the previous 5 years were excluded from the study.

[0041] At the end, a total of 180 subjects from six Spanish hospitals (Hospital de Burgos, Hospital de Vigo, Hospital de Alicante, Hospital de Ourense, Instituto Valenciano de Oncologia, and Hospital de Bellvitge) were prospectively included in this study: 120 patients newly diagnosed with sporadic colorectal neoplasia (60 with CRC and 60 with AA) and 60 healthy individuals.

[0042] The results were validated in a cohort of 92 subjects from four Spanish hospitals (Hospital de Burgos, Hospital de Vigo, Hospital de Ourense, Instituto Valenciano de Oncologia): 59 patients diagnosed with sporadic colorectal neoplasia (32 with CRC and 27 with AA) and 33 healthy individuals.

[0043] All subjects without personal history of any cancer and with a recent colonoscopy confirming the lack of colorectal neoplastic lesions. Patients with AA were those with adenomas having a size of at least 10 mm or histologically having high grade dysplasia or >20% villous component. The characteristics of participants are shown in Table 1 and Table 2. Blood samples were collected prior to endoscopy or surgery in all individuals.

TABLE-US-00002 TABLE 1 Clinical-pathological characteristics of the study cohort 180 Cases Control (CTL) AA CRC TOTAL Mean age (SD)  58 (34-74)  62 (36-81) 64.7 (38-86) 61.6 (34-86) GENDER Male 30 (50%) 36 (60%) 31 (51.6%) 97 (53.8%) Female 30 (50%) 24 (40%) 29 (48.4%) 83 (46.2%) COLORECTAL FEATURES TNM stage I 13 II 13 III 18 IV 12 Unknown 4 Location Ascending colon and cecum 17 Descending colon and sigma 16 Transverse 1 Rectum 21 Hepatic flexure 5 ADVANCED COLORECTAL ADENOMA FEATURES Size => 10 mm 49 Small AA (<=15 mm) 29 Big AA (>15 mm) 23 Unknown 8 Mean size (mm) (SD) 17 No. AAs Mean (SD) 1.57 High-grade dysplasia Yes 21 No 30 Unknown 9 Villous component Yes 17 No 33 Unknown 10

TABLE-US-00003 TABLE 2 Clinical-pathological characteristics of the study cohort 92 Cases Control (CTL) AA CRC TOTAL Mean age (SD) 64 (44-36)  65.3 (39-88)   72 (54-85) 67 (39-88)  GENDER Male 16 (48.5%) 16 (59.3%) 17 (53%) 49 (53.2%) Female 17 (51.5%) 11 (40.7%) 15 (47%) 43 (46.8%) COLORECTAL FEATURES TNM stage I 3 II 9 III 9 IV 7 Unknown 4 Location Ascending colon and cecum 11 Descending colon and sigma 11 Transverse 3 Rectum 6 Unknown 1 ADVANCED COLORECTAL ADENOMA FEATURES Size => 10 mm 22 Small AA (<=15 mm) 15 Big AA (>15 mm) 11 Mean size (mm) (SD) 18.5 No. AAs Mean (SD) 2.2 Unknown 1 High-grade dysplasia Yes 9 No 16 Unknown 2 Villous component Yes 13 No 12 Unknown 2

[0044] The study was approved by the Institutional Ethics Committee of each Hospital, and written informed consent was obtained from all participants in accordance with the Declaration of Helsinki.

Example 1.2. Sample Preparation

[0045] Ten mL of whole blood from each participant were collected in EDTA K2 containing tubes. Blood samples were placed at 4° C. until plasma separation. Samples were centrifuged at 1,600×g for 10 min at 4° C. to spin down blood cells, and plasma was transferred into new tubes, followed by further centrifugation at 16,000×g for 10 minutes at 4° C. to completely remove cellular components.

[0046] Blood samples were obtained before colonoscopy or surgery resection.

Example 1.3. Molecular Analysis

[0047] The concentration of the biomarkers in plasma samples was established using commercial ELISA (Enzyme-linked immunosorbent assay) and CLIA (Chemiluminescence immunoassay) test and following their corresponding instruction manual. HGFR, ErbB4, CEA, CD163, DKK3, IGFBP2, and TFF3 was analyzed with ELISA kit from Cloud clone Corp. Level of CD147, Flt3L, FasL and Casp4, was measured using ELISA Kit form Elabscience. In the case of the DCSINGR, ELISA kit from RayBio was used. PKM2 were analyzed with ELISA kit from Aviva and ADAMDEC1 with ELISA kit from Cusabio. Related to CLIA test, CYFRA21-1 y AREG was analyzed with CLIA test from Cloud Clone Corp.

Example 1.4. Data Quantification

[0048] For the protein quantification step the samples were processed with the corresponding kit (ELISA/CLIA) and distributed in experimental plates. Each plate contained also control data used to construct a standard curve. Fluorescence data obtained from each run (expressed as integer numbers) have been background corrected for each sample and quantified using a standard curve generated using a 2-degree polynomial regression model.

Example 1.5. Statistical Analysis

[0049] Three groups of individuals were considered in the analysis. CRC (Individuals diagnosed with colorectal cancer), AA (Individuals diagnosed with advanced adenoma) and CTL (Individuals with no disease).

[0050] Raw quantification data have been transformed by applying square root function, and then centering and scaling so that, after the transformation, each protein measure has mean 0 and standard deviation 1.

[0051] To deal with non-normality issues, Wilcoxon rank-sum test was used to compare either CRC cases or AA cases against CTL individuals.

[0052] Diagnostic performance for the individual proteins and some of their combinations has been assessed by their receiver operating characteristic (ROC) curves, and the area under the ROC curves (AUC). Moreover, sensitivities, specificities, positive predictive and negative predictive values (PPV and NPV) for the different tests were calculated at the optimal cutoff point defined by the best Youden's Index (or equivalently, the point of the ROC that maximizes the sum of sensitivity and specificity).

[0053] Scores used for deriving the ROC-AUCs and the rest of performance values were obtained using univariate logistic regression model for the individual proteins and multivariate logistic regression models for the different combination of proteins considered. 95% CI for the AUCs was obtained with the DeLong methodology both in individual markers and combination of them.

[0054] All the individual performance metrics obtained from the previous analysis are summarized in Table 3 and Table 4.

[0055] Moreover, multivariate analysis has been carried out to explore if combinations of proteins improve the performance of individual markers. We have used multivariate logistic regression to fit models with all possible combinations of two, three and four proteins.

Example 2. Results

Example 2.1. Individual Marker Results

[0056] Different metrics to evaluate the individual proteins were determined, also perming the following comparisons: CRC/AA vs CTL. Table 3 and Table 4 show these metrics for individual proteins, including p-value from Wilcoxon test (p.Wilc), area under the ROC curve (AUC), and Sensitivity (Sens.) and Specificity (Spec.) values, computed in the cut-off point of the ROC curve with the best Youden's index.

[0057] It can be seen that TFF3, CYFRA21-1, Flt3L and AREG are significantly (p<0.05) different between CRC and CTL groups and their AUC are significantly different from 0.5 (as their 95% confidence interval do not include 0.5). In the case of AA group, CASP4 and CYFRA21-1 also shows statistically differences compared to CTL group.

TABLE-US-00004 TABLE 3 CRC Marker pWilcoxon AUC Sens Spec TFF3 0 0.751 (0.663-0.839) 80 63.33 78.33 65 CYFRA21-1 0 0.746 (0.66-0.833)  68.33 68.33 56.67 80 Flt3L 0.001 0.682 (0.583-0.781) 66.67 73.33 AREG 0.025 0.619 (0.518-0.719) 81.67 43.33 CASP4 0.071 0.596 (0.493-0.698) 35 86.67 IGFBP2 0.095 0.589 (0.487-0.691) 98.33 16.67 95 20 20 95 ADAMDEC1 0.122 0.582 (0.479-0.685) 71.67 48.33 DKK3 0.224 0.564 (0.461-0.668) 93.33 25 FASL 0.378 0.547 (0.442-0.651) 56.67 56.67 55 58.33 28.33 85 DCSIGNR 0.379 0.547 (0.442-0.651) 41.67 75 ErbB4 0.419 0.543 (0.439-0.647) 58.33 55 41.67 71.67 40 73.33 CD163 0.624 0.526 (0.422-0.63)  43.33 66.67 HGFR 0.696 0.521 (0.416-0.626) 61.67 53.33 CEA 0.767 0.516 (0.411-0.621) 25 88.33 CD147 0.869 0.509 (0.404-0.613) 45 65 PKM2 0.904 0.507 (0.402-0.611) 46.67 63.33

Individual Performance Metrics in CRC

[0058]

TABLE-US-00005 TABLE 4 AA Marker pWilcoxon AUC Sens Spec CASP4 0.011 0.635 (0.536-0.734) 73.33 50 CYFRA21-1 0.019 0.625 (0.524-0.725) 81.67 43.33 80 45 TFF3 0.061 0.599 (0.497-0.702) 70 55 68.33 56.67 66.67 58.33 Flt3L 0.099 0.588 (0.485-0.69) 61.67 58.33 IGFBP2 0.187 0.57 (0.466-0.674) 75 45 73.33 46.67 71.67 48.33 DCSIGNR 0.243 0.562 (0.458-0.665) 50 65 ErbB4 0.265 0.559 (0.455-0.663) 65 55 CD163 0.356 0.549 (0.445-0.653) 26.67 90 HGFR 0.372 0.547 (0.443-0.652) 65 51.67 AREG 0.613 0.527 (0.422-0.632) 90 23.33 88.33 25 ADAMDEC1 0.654 0.524 (0.419-0.628) 38.33 73.33 DKK3 0.688 0.521 (0.417-0.626) 50 60 46.67 63.33 43.33 66.67 38.33 71.67 CD147 0.854 0.51 (0.405-0.614) 61.67 48.33 60 50 FASL 0.856 0.51 (0.405-0.614) 65 43.33 CEA 0.914 0.494 (0.389-0.599) 60 53.33 58.33 55 PKM2 0.998 0.5 (0.395-0.605) 63.33 48.33 60 51.67

Individual Performance Metrics in AA

[0059] Additionally, there are statistical significances between early stage vs control (p value=0.00098), and early stage+AA vs control (p value=0.0048) indicating that TFF3 is a good marker for screening purposes (i.e. early detection).

Example 2.2. Best Combinations of Biomarkers

[0060] With the aim of improving individual diagnostic capability, combinations of proteins have been explored. All possible combinations of two, three and four proteins have been analyzed.

[0061] We have used the cohort of 180 subjects for developing models with all possible combinations and we have used the validation cohort (n=92) to obtain validated performance metrics (AUC values, Sens, Spec, PPV and NPV).

[0062] Table 5, Table 6 and Table 7 show the AUC achieved for the combinations of two, three and four biomarkers respectively, discriminating CRC from CTL.

TABLE-US-00006 TABLE 5 Combinations of two biomarkers CRC vs CTL AUC TFF3/Flt3L 0.8672 Flt3L/CYFRA21-1 0.8135 TFF3/CYFRA21-1 0.8047 HGFR/CYFRA21-1 0.8037 CYFRA21-1/CD147 0.8027 CYFRA21-1/IGFBP2 0.7988 CD163/CYFRA21-1 0.7979 CYFRA21-1/CEA 0.7949 TFF3/CD147 0.7871 Flt3L/CEA 0.7871 TFF3/HGFR 0.7813

TABLE-US-00007 TABLE 6 Combinations of three biomarkers CRC vs CTL AUC TFF3/Flt3L/HGFR 0.8672 TFF3/Flt3L/IGFBP2 0.8652 TFF3/Flt3L/CD147 0.8574 TFF3/Flt3L/CD163 0.8535 Flt3L/CYFRA21-1/CD147 0.8135 TFF3/Flt3L/CYFRA21-l 0.8105 Flt3L/HGFR/CYFRA21-1 0.8105 Flt3L/CYFRA21-1/IGFBP2 0.8105 Flt3L/CD163/CYFRA21-1 0.8096 Flt3L/CYFRA21-1/CEA 0.8096 TFF3/CYFRA21-1/CD147 0.8076 TFF3/HGFR/CYFRA21-1 0.8047 TFF3/CYFRA21-1/IGFBP2 0.8047 TFF3/CD163/CYFRA21-1 0.8027 HGFR/CYFRA21-1/CD147 0.8027 HGFR/CYFRA21-1/IGFBP2 0.8018 CD163/CYFRA21-1/IGFBP2 0.8018 CYFRA21-1/IGFBP2/CD147 0.7998 CD163/HGFR/CYFRA21-1 0.7988 TFF3/CYFRA21-1/CEA 0.7979 CD163/CYFRA21-1/CEA 0.7979 CD163/CYFRA21-1/CD147 0.7979 HGFR/CYFRA21-1/CEA 0.7959 CYFRA21-1/CEA/CD147 0.7949 CYFRA21-1/IGFBP2/CEA 0.7939 TFF3/HGFR/CD147 0.7891 TFF3/IGFBP2/CD147 0.7881 Flt3L/CEA/CD147 0.7871 Flt3L/IGFBP2/CEA 0.7822 HGFR/Flt3L/CEA 0.7793 CD163/CYFRA21-1/FASL 0.7722 AREG/HGFR/CYFRA21-1 0.7705

TABLE-US-00008 TABLE 7 Combinations of four biomarkers CRC vs CTL AUC TFF3/Flt3L/HGFR/IGFBP2 0.8652 TFF3/Flt3L/HGFR/CD147 0.8584 TFF3/Flt3L/IGFBP2/CD147 0.8574 TFF3/Flt3L/CD163/IGFBP2 0.8418 TFF3/Flt3L/CD163/HGFR 0.8389 TFF3/Flt3L/CD163/CD147 0.8193 TFF3/Flt3L/CYFRA21-1/CD147 0.8135 TFF3/Flt3L/CYFRA21-1/IGFBP2 0.8135 TFF3/Flt3L/CD163/CYFRA21-l 0.8125 Flt3L/CYFRA21-1/IGFBP2/CD147 0.8125 Flt3L/CD163/CYFRA21-1/IGFBP2 0.8115 TFF3/Flt3L/HGFR/CYFRA21-1 0.8105 Flt3L/HGFR/CYFRA21-1/CD147 0.8105 Flt3L/HGFR/CYFRA21-1/IGFBP2 0.8105 TFF3/CYFRA21-1/FASL/CD147 0.8105 Flt3L/CYFRA21-1/CEA/CD147 0.8096 DKK3/CD163/CYFRA21-1/IGFBP2 0.8096 TFF3/CYFRA21-1/IGFBP2/CD147 0.8086 Flt3L/HGFR/CYFRA21-1/CEA 0.8086 Flt3L/CD163/CYFRA21-1/CD147 0.8076 Flt3L/CD163/HGFR/CYFRA21-1 0.8066 Flt3L/CYFRA21-1/IGFBP2/CEA 0.8066 TFF3/HGFR/CYFRA21-1/CD147 0.8057 Flt3L/CD163/CYFRA21-1/CEA 0.8057 TFF3/Flt3L/CYFRA21-1/CEA 0.8047 TFF3/HGFR/CYFRA21-1/IGFBP2 0.8047 TFF3/CD163/CYFRA21-1/CD147 0.8037 TFF3/CD163/HGFR/CYFRA21-1 0.8037 TFF3/CD163/CYFRA21-1/IGFBP2 0.8037 Flt3L/IGFBP2/CEA/CD147 0.8027 HGFR/CYFRA21-1/IGFBP2/CD147 0.8018 TFF3/CYFRA21-1/CEA/PKM2 0.8008 CD163/HGFR/CYFRA21-1/IGFBP2 0.7988 CD163/HGFR/CYFRA21-1/CD147 0.7988 CYFRA21-1/IGFBP2/FASL/CD147 0.7984 CD163/CYFRA21-1/IGFBP2/CD147 0.7979 CD163/HGFR/CYFRA21-1/CEA 0.7979 CD163/CYFRA21-1/CEA/CD147 0.7979 CD163/CYFRA21-1/IGFBP2/CEA 0.7969 HGFR/CYFRA21-1/CEA/CD147 0.7959 TFF3/CYFRA21-1/CEA/CD147 0.7949 TFF3/HGFR/CYFRA21-1/CEA 0.7949 TFF3/CYFRA21-1/IGFBP2/CEA 0.7949 TFF3/CD163/CYFRA21-1/CEA 0.7939 HGFR/CYFRA21-1/IGFBP2/CEA 0.7939 CYFRA21-1/IGFBP2/CEA/CD147 0.7939 TFF3/HGFR/IGFBP2/CD147 0.7891 Flt3L/HGFR/CEA/CD147 0.7832 CD163/CYFRA21-1/FASL/CD147 0.7792 Flt3L/HGFR/IGFBP2/CEA 0.7754 CD163/CYFRA21-1/IGFBP2/FASL 0.7571 Flt3L/HGFR/IGFBP2/CD147 0.7568 CD163/CYFRA21-1/CEA/FASL 0.7560

[0063] Since the TFF3/Flt3L pair appears as the best combination of two proteins and it is also present among the top combinations of three and four markers in the CRC. vs. CTL comparison, we have explored its performance in discriminating AA from CTL.

[0064] We obtained an AUC of 0.5361 when using the pair TFF3/Flt3L for discriminating AA from CTL. We combined this pair with the rest of proteins to obtain combinations of three and four markers. Table 8 and Table 9 shows the AUC achieved for these combinations of three and four biomarkers respectively, discriminating AA vs CTL.

TABLE-US-00009 TABLE 8 Combinations of three biomarkers AA vs CNT AUC TFF3/Flt3L/ADAMDEC1 0.6430 TFF3/Flt3L/CYFRA21-1 0.6358 TFF3/Flt3L/FASL 0.6310

TABLE-US-00010 TABLE 9 Combinations of four biomarkers AA vs CTL AUC TFF3/Flt3L/CD163/ADAMDEC1 0.6815 TFF3/Flt3L/FASL/ADAMDEC1 0.6623 TFF3/Flt3L/IGFBP2/ADAMDECl 0.6454 TFF3/Flt3L/AREG/DKK3 0.6418 TFF3/Flt3L/CYFRA21-1/CEA 0.6418 TFF3/Flt3L/HGFR/CYFRA21-1 0.6394 TFF3/Flt3L/CYFRA21-1/CD147 0.6370 TFF3/Flt3L/HGFR/ADAMDEC1 0.6370 TFF3/Flt3L/CEA/ADAMDEC1 0.6358 TFF3/Flt3L/ADAMDEC1/CD147 0.6346 TFF3/Flt3L/CYFRA21-1/IGFBP2 0.6334 TFF3/Flt3L/FASL/CD147 0.6334 TFF3/Flt3L/CEA/FASL 0.6322 TFF3/Flt3L/HGFR/FASL 0.6322 TFF3/Flt3L/IGFBP2/FASL 0.6298 TFF3/Flt3L/CD163/FASL 0.6178 TFF3/Flt3L/AREG/DCSIGNR 0.6130 TFF3/Flt3L/CD163/IGFBP2 0.6010

[0065] Based on their respective AUCs, the best models have been selected. Table 10 shows the best results for CRC and Table 11 shows the best results for AA. The metrics for the best combinations of proteins are included, comprising area under the ROC curve (AUC), Sensitivity (Sens.), Specificity (Spec.), and positive (PPV) and negative (NPV) predictive values computed in the cut-off point of the ROC curve with the best Youden's index.

TABLE-US-00011 TABLE 10 Best combinations for CRC AUC Sens Spec PPV NPV Best combination of two biomarkers TFF3/Flt3L 0.8672 0.8438 0.7813 0.7941 0.8333 Best combination of three biomarkers TFF3/Flt3L/HGFR 0.8672 0.8438 0.7813 0.7941 0.8333 TFF3/Flt3L/IGFBP2 0.8652 0.8750 0.7500 0.7778 0.8571 TFF3/Flt3L/CD147 0.8574 0.9688 0.5625 0.6889 0.9474 TFF3/Flt3L/CD163 0.8535 0.8125 0.7813 0.7879 0.8065 TFF3/Flt3L/CYFRA21-1 0.8105 0.9063 0.6563 0.7250 0.8750 Best combination of four biomarkers TFF3/Flt3L/HGFR/IGFBP2 0.8652 0.8750 0.7500 0.7778 0.8571 TFF3/Flt3L/HGFR/CD147 0.8584 0.9688 0.5625 0.6889 0.9474 TFF3/Flt3L/IGFBP2/CD147 0.8574 0.9688 0.5625 0.6889 0.9474 TFF3/Flt3L/CD163/IGFBP2 0.8418 0.7500 0.8125 0.8000 0.7647 TFF3/Flt3L/CD163/HGFR 0.8389 0.8125 0.7500 0.7647 0.8000 TFF3/Flt3L/CD163/CD147 0.8193 0.7813 0.7500 0.7576 0.7742 TFF3/Flt3L/CYFRA21-1/CD147 0.8135 0.9063 0.6563 0.7250 0.8750 TFF3/Flt3L/CYFRA21-1/IGFBP2 0.8135 0.9063 0.6563 0.7250 0.8750 TFF3/Flt3L/CD163/CYFRA21-1 0.8125 0.9063 0.6563 0.7250 0.8750 TFF3/Flt3L/HGFR/CYFRA21-1 0.8105 0.8438 0.7500 0.7714 0.8276 TFF3/Flt3L/CYFRA21-1/CEA 0.8047 0.9063 0.6563 0.7250 0.8750

TABLE-US-00012 TABLE 11 Best combinations for AA AUC Sens Spec PPV NPV Best combinations of three biomarkers TFF3/Flt3L/ADAMDEC1 0.6430 0.6538 0.5938 0.5667 0.6786 TFF3/Flt3L/CYFRA21-l 0.6358 0.5385 0.7500 0.6364 0.6667 TFF3/Flt3L/FASL 0.6310 0.7692 0.5625 0.5882 0.7500 Best combinations of four biomarkers TFF3/Flt3L/CD163/ADAMDEC1 0.6815 0.6538 0.7188 0.6538 0.7188 TFF3/Flt3L/FASL/ADAMDEC1 0.6623 0.7308 0.6875 0.6552 0.7586 TFF3/Flt3L/IGFBP2/ADAMDEC1 0.6454 0.6538 0.5938 0.5667 0.6786 TFF3/Flt3L/AREG/DKK3 0.6418 0.5000 0.8438 0.7222 0.6750 TFF3/Flt3L/CYFRA21-1/CEA 0.6418 0.5385 0.7500 0.6364 0.6667 TFF3/Flt3L/HGFR/CYFRA21-1 0.6394 0.5385 0.7500 0.6364 0.6667 TFF3/Flt3L/CYFRA21-1/CD147 0.6370 0.5385 0.7500 0.6364 0.6667 TFF3/Flt3L/HGFR/ADAMDEC1 0.6370 0.6538 0.5938 0.5667 0.6786 TFF3/Flt3L/CEA/ADAMDEC1 0.6358 0.6538 0.5938 0.5667 0.6786 TFF3/Flt3L/ADAMDEC1/CD147 0.6346 0.6538 0.5938 0.5667 0.6786 TFF3/Flt3L/CYFRA21-1/IGFBP2 0.6334 0.5385 0.7500 0.6364 0.6667 TFF3/Flt3L/FASL/CD147 0.6334 0.7308 0.5938 0.5938 0.7308 TFF3/Flt3L/CEA/FASL 0.6322 0.7308 0.5938 0.5938 0.7308 TFF3/Flt3L/HGFR/FASL 0.6322 0.7692 0.5625 0.5882 0.7500 TFF3/Flt3L/IGFBP2/FASL 0.6298 0.7308 0.5938 0.5938 0.7308 TFF3/Flt3L/CD163/FASL 0.6178 0.7308 0.5625 0.5758 0.7200 TFF3/Flt3L/AREG/DCSIGNR 0.6130 0.8846 0.4063 0.5476 0.8125 TFF3/Flt3L/CD163/IGFBP2 0.6010 0.3846 0.8125 0.6250 0.6190

[0066] Table 12 has been designed to show the overlapping of the most important signatures claimed in the present invention for CRC. It is clearly shown that most of the best signatures comprise [TFF3 and Flt3L]

TABLE-US-00013 TABLE 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 CRC X X 0.8672 X X X 0.8672 X X X 0.8652 X X X X 0.8652 1) TFF3, 2) CYFRA21-1, 3) Flt3L, 4) AREG, 5) CASP4, 6) IGFBP2, 7) ADAMDEC1, 8) DKK3, 9) FASL, 10) DCSIGNR, 11) ErbB4, 12) CD163, 13) HGFR, 14) CEA, 15) CD147 and 16) PKM2

[0067] Table 13 has been designed to show the overlapping of the most important signatures claimed in the present invention for AA. It is clearly shown that [TFF3 and CYFRA21-1] is also included in combination of biomarkers showing a good performance for AA.

TABLE-US-00014 TABLE 13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 CRC X X X X 0.6815 X X X X 0.6623 X X X X 0.6454 X X X 0.6430 1) TFF3, 2) CYFRA21-1, 3) Flt3L, 4) AREG, 5) CASP4, 6) IGFBP2, 7) ADAMDEC1, 8) DKK3, 9) FASL, 10) DCSIGNR, 11) ErbB4, 12) CD163, 13) HGFR, 14) CEA, 15) CD147 and 16) PKM2