METHOD OF DETECTING BACTERIA

Abstract

A method to detect bacteria in a biological sample in order to determine if the biological sample contains tuberculous mycobacteria or non-tuberculous mycobacteria, said tuberculous mycobacteria being detected by analyzing a set of measurements consisting of a plurality of light scattering measurements by means of algorithms making use of neural networks.

Claims

1. A method to detect bacteria in a biological sample in order to determine if said biological sample comprises tuberculous mycobacteria (MTC) or non-tuberculous mycobacteria (MOTT), the method comprising: subjecting said biological sample to a plurality of light scattering measurements in order to acquire a plurality of set of measurements as a function of time; processing said set of measurements acquired, so as to determine at least the starting time of the exponential phase of bacterial growth, the maximum value of bacterial growth of said set of measurements acquired and the temporal development of at least one statistical parameter; calculating one or more characteristic values from said temporal development of said statistical parameter; and analyzing, by means of one or more algorithms making use of a neural network, at least said starting time of the exponential phase of bacterial growth, said maximum value of bacterial growth of said set of measurements acquired and said calculated characteristic values, in order to determine if said biological sample comprises tuberculous mycobacteria or non-tuberculous mycobacteria.

2. The method as in claim 1, wherein said statistical parameter is chosen from a group comprising mean, median, standard deviation, variance or other statistical parameters which can be derived from the function of auto-correlation of said light scattering measurements that make up each of said set of measurements.

3. The method as in claim 1 wherein said starting time of the exponential phase of bacterial growth is determined when the average measurement of said light scattering measurements of a group consisting of the last set of measurements acquired and of a number of sets of measurements prior to the last set of measurements acquired comprised between 1 and 12, is greater than a predefined threshold value.

4. The method as in claim 1, wherein said characteristic values are defined by the values of the progressive average of said temporal development of the standard deviation calculated starting from said starting time of the exponential phase of bacterial growth.

5. The method as in claim 1, wherein, if the exponential bacterial growth occurs before a limit time comprised between 2 and 4 days, it provides to signal that the biological sample contains contaminants.

6. The method as in claim 1, providing to subject said biological sample to said plurality of light scattering measurements with a sampling frequency comprised between 0.5 kHz and 10 kHz and for a time comprised between 0.5 seconds and 5 seconds.

7. The method as in claim 1, providing to acquire said plurality of set of measurements at intervals of time defined for a temporal period defined at least up to about 11 days following the beginning of said exponential phase of bacterial growth.

8. The method as in claim 1, comprising at least a verification step in which it is determined if said tuberculous mycobacteria (MTC) in said biological sample are able to resist one or more antibiotic drugs, determining if said mycobacteria are the multidrug-resistant type, or determining the single or crossed resistance at the same time as the bacterial growth.

9. The method as in claim 1, wherein said biological sample does not come into contact with the outside environment for the entire execution of the culture step and the bacterial detection procedure.

10. The method as in claim 1, wherein said plurality of light scattering measurements is carried out by means of an automatic scan turbidimeter, and the processing, calculation and analysis are carried out by a processing unit able to determine said statistical parameters, to calculate said characteristic values from said at least one set of measurements and to determine if said biological sample comprises tuberculous mycobacteria or non-tuberculous mycobacteria.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0052] These and other characteristics of the present invention will become apparent from the following description of some embodiments, given as a non-restrictive example with reference to the attached drawings wherein:

[0053] FIG. 1 is a diagram in which the areas of the world with the highest incidence of cases of tuberculosis are highlighted;

[0054] FIG. 2 is an enlargement under the microscope of a tubercular mycobacterium;

[0055] FIG. 3 is a bar diagram that indicates the distribution of the days of growth of tuberculous mycobacteria;

[0056] FIG. 4 is a graph that shows the radial distribution of the mycobacteria, either tubercular or not, in two clusters (A and B) using a K-means algorithm;

[0057] FIG. 5 schematically shows three sets of measurements acquired around the initial bacterial growth phase of the bacterial growth curve;

[0058] FIGS. 6 and 7 show the temporal developments of the progressive average of the deviation standard of two biological samples respectively containing MTC and MOTT.

DETAILED DESCRIPTION OF SOME EMBODIMENTS

[0059] We will now describe possible embodiments of a bacterial detection method in a biological sample, in which the method provides to determine whether the biological sample comprises tuberculous mycobacteria MTC or non-tuberculous mycobacteria MOTT.

[0060] The biological sample can be selected from one or more body fluids, such as respiratory fluids, pleural fluids, ascites, liquor, blood, bile, urine or suchlike, or it can be a different biological sample, for example feces, sputum, or other.

[0061] According to one aspect of the present invention, the detection method provides at least:

[0062] to subject the biological sample to a plurality of light scattering measurements to acquire a plurality of sets of measurements as a function of time;

[0063] to process the acquired set of measurements to determine at least the starting time of the exponential phase of bacterial growth, the value of the maximum bacterial growth of the acquired sets of measurements and the temporal development of at least one statistical parameter;

[0064] to calculate one or more characteristic values of the temporal development of the statistical parameter;

[0065] to analyze, by means of one or more algorithms that use a neural network, at least the starting time of the exponential phase of bacterial growth, the value of the maximum bacterial growth of the acquired sets of measurements and the characteristic values calculated;

[0066] to determine, on the basis of this analysis performed by the neural network, whether the biological sample comprises tuberculous mycobacteria MTC or non-tuberculous mycobacteria MOTT.

[0067] According to possible solutions, the biological sample is put in a culture broth inside a container.

[0068] The culture broth can comprise a liquid phase Middlebrook broth, with the addition of PANTA additives (Polymyxin B, Amphotericin B, Nalidixic acid, Trimethoprim, Azlocillin), to which an OADC (Oleic Albumin Dextrose Catalase) culture supplement can be added.

[0069] In particular, it has been noted that with this solution the results regarding the screening culture phase are available from 24 to 48 hours before the MGIT system (Mycobacter Growth Indicator Tube), which is a market-leader instrument and known in the state of the art.

[0070] Mycobacteria in a liquid medium are oriented particles which, when subjected to a light generated by a source, disperse the light.

[0071] Depending on the response time and frequency of the refracted light, the method also allows to determine the size and morphological structure of the mycobacteria in the biological sample.

[0072] The present invention allows to determine one or more statistical parameters from the statistical distribution of the light scattering measurements acquired, from which it is possible to discriminate quickly and accurately whether the biological sample comprises tuberculous mycobacteria MTC or non-tuberculous mycobacteria MOTT.

[0073] The set of measurements consisting of the light scattering measurements will be characterized by a statistical distribution from which it is possible to extrapolate specific statistical parameters such as mean, median, standard deviation, variance or other characteristic parameters such as for example the autocorrelation function of the light scattering measurements that constitute the set of measurements.

[0074] By subjecting the biological sample to a plurality of light scattering measurements at defined time intervals, it is possible to calculate the experimental points of the bacterial growth curve, each of the experimental points being defined by the median of the light scattering measurements acquired in a determinate time.

[0075] With reference to FIG. 5, a bacterial growth curve is schematically shown, obtained from the detection of the experimental points defined by the median of the light scattering measurements acquired from the biological sample in a determinate time.

[0076] Each experimental point of the bacterial growth curve is obtained starting from a corresponding set of measurements consisting of the light scattering measurements acquired.

[0077] FIG. 5 shows how the sets of measurements acquired around the initial bacterial growth phase have statistical distributions useful for the detection and recognition of tuberculous mycobacteria MTC.

[0078] If the exponential bacterial growth occurs before a defined limit time, that is, if it occurs before the minimum time in which the mycobacteria typically grow, the method provides to signal that the biological sample contains contaminants.

[0079] According to possible embodiments, the time limit defined can be comprised between 2 and 4 days from when the culture is obtained, that is, from the beginning of the analysis, advantageously the time limit defined is 3 days from when the culture is obtained.

[0080] In the event that no exponential bacterial growth is detected, the method provides to signal that the biological sample is negative.

[0081] Applicant has identified how the statistical distributions and in particular the temporal development of the standard deviation of the sets of measurements allows to recognize if in the biological sample there are tuberculous mycobacteria MTC or non-tuberculous mycobacteria MOTT.

[0082] By processing the sets of measurements, or from their statistical distributions, it is possible to obtain characteristic values correlated to the morphology of the tuberculous mycobacteria.

[0083] According to possible embodiments, the characteristic values are defined by the values of the progressive average of the temporal development of the standard deviation calculated from the starting time of the exponential phase of bacterial growth.

[0084] With reference to FIGS. 6 and 7 two temporal developments are shown of the progressive average of the standard deviations of two biological samples containing respectively MTC and MOTT.

[0085] As can be seen, the temporal development of the characteristic values of the sample with MTC are different from those relating to the MOTT sample.

[0086] From the differences in the two temporal developments it is possible to discriminate whether in a biological sample there are MTC or MOTT mycobacteria.

[0087] From the analysis of the characteristic values derived from the temporal development of the statistical parameter determined by the sets of measurements it is possible to determine, based on defined threshold values or using algorithms using a neural network, if the biological sample comprises tuberculous mycobacteria MTC or non-tuberculous mycobacteria MOTT.

[0088] Until now it has been known that in order to obtain information, or at least determine whether the biological sample comprises tuberculous mycobacteria MTC or non-tuberculous mycobacteria MOTT, it was necessary to have the bacterial growth curve in its entirety. This made known methods slow, with disadvantages not always admissible.

[0089] Applicant has identified a surprising link between the statistical distribution of the light scattering measurements and the morphology of tuberculous mycobacteria MTC which allows to quickly identify if the biological sample comprises tuberculous mycobacteria MTC by analyzing the set of measurements acquired.

[0090] In particular, a close link was identified between the function of autocorrelation of the set of measurements, as well as between the statistical parameters deriving from it, and the morphology of the tuberculous mycobacteria MTC.

[0091] In fact, due to the exclusive production of cord factor on the cell wall of such mycobacteria, their morphology and hence the statistical distribution of light scattering measurements allow to clearly distinguish these mycobacteria from MOTT.

[0092] Thanks to this, the detection method according to the present invention is able to provide a response much more quickly compared to the times necessary up to now.

[0093] Moreover, thanks to the present invention it is possible to determine, directly and without handling the potentially dangerous biological sample, whether it contains tuberculous mycobacteria MTC or non-tuberculous mycobacteria MOTT.

[0094] According to possible embodiments, the method provides that the biological sample is inserted in a hermetically sealed container provided with an irreversible closing device, that is, in a vial or other similar container able to seal the contents and not allow an operator to open it.

[0095] This guarantees a high level of safety for the operators, as the container, being provided with an irreversible closing device, cannot be opened once closed.

[0096] Unlike known techniques which use indirect methods, such as for example measurements of carbon dioxide or other gases, the present invention is highly sensitive.

[0097] Moreover, by acquiring and monitoring the light scattering measurements from the beginning, it is also possible to discriminate whether any possible contaminants are present in the sample, as they grow before the mycobacteria and with a different slope. In this case it will be possible to repeat the sampling of the biological sample and hence its analysis in a short time.

[0098] Furthermore, the method allows to quickly identify whether and how the mycobacteria grow in the expected time, without needing to acquire the entire bacterial growth curve.

[0099] According to possible embodiments, the plurality of light scattering measurements can be performed by means of an automatic scanning turbidimeter.

[0100] For example, vials with biological samples can be inserted into an automatic scanning turbidimeter able to simultaneously perform a plurality of light scattering measurements.

[0101] The turbidimeter can use laser light or other source of electromagnetic radiation. For example, the turbidimeter can be an HB&L-Uro4 system, produced and patented by Applicant and widely used in the sector.

[0102] According to possible embodiments, the processing of the set of measurements can be carried out by means of a processing unit able to determine at least the starting time of the exponential phase of bacterial growth, the maximum value of bacterial growth of the sets of measurements acquired and the temporal development of at least one statistical parameter.

[0103] The processing unit can also perform the calculation of the characteristic values from the temporal development, as well as the analysis using one or more algorithms using a neural network.

[0104] For example, the processing unit can be an electronic board, an integrated circuit or another similar or comparable microprocessor.

[0105] The processing unit can comprise, or be connected to, a calculator unit able to calculate one or more characteristic values from the temporal development of at least one statistical parameter.

[0106] The processing unit can comprise, or be connected to, an analysis unit able to process and execute one or more algorithms using a neural network to determine whether the biological sample comprises tuberculous mycobacteria MTC or non-tuberculous mycobacteria MOTT.

[0107] The analysis of the data is obtained with a neural network, in the case described here a known K-means algorithm, appropriately adapted for the purpose, by means of an analysis program.

[0108] According to possible variants, the analysis can be performed using algorithms based both on the K-Means algorithm and also on Self Organizing Maps.

[0109] According to possible embodiments, the biological sample is subjected to a plurality of light scattering measurements with a sampling frequency comprised between 0.5 kHz and 10 kHz.

[0110] According to possible embodiments, the biological sample is subjected to a plurality of light scattering measurements for a time comprised between 0.5 seconds and 5 seconds.

[0111] According to possible embodiments, the detection method requires that a plurality of sets of measurements is acquired at defined time intervals for a defined period of time at least up to about 11 days after the beginning of the exponential phase of bacterial growth.

[0112] This period of time allows to obtain a response with a high level of reliability, since by monitoring and analyzing the sets of measurements made according to the present invention it is possible to accurately identify whether the biological sample comprises tuberculous mycobacteria MTC.

[0113] According to possible embodiments, the method comprises at least a verification step in which it is determined whether the tuberculous mycobacteria MTC inside the sample are able to resist one or more antibiotic drugs, determining whether the mycobacteria are the multidrug-resistant mycobacteria type MDR, or determining single or crossed resistance simultaneously with bacterial growth.

[0114] By performing the antibiotic resistance tests simultaneously with the growth and detection step of the tuberculous mycobacteria MTC, it is possible to identify the most appropriate antibiotic therapy.

[0115] According to possible embodiments, the method provides that the biological sample in the culture does not come into contact with the external environment for the entire execution of the bacterial detection procedure.

[0116] In a possible application of the method according to the present invention, the biological sample was inoculated in two vials, each vial containing a culture broth. The vials can advantageously be provided with an irreversible closing device.

[0117] The first vial contains the biological sample, possibly pretreated in culture broth, while in the subsequent vial or vials, in addition to the biological sample, as in the first vial, one or more antibiotic drugs is also added.

[0118] Subsequent vials can comprise a single antibiotic to verify single sensitivity to a drug, or a cocktail of two or more different antibiotics to determine multi-resistance to antibiotics, for example cross-resistance to isoniazid/rifampicin.

[0119] If there is only one antibiotic per vial, using several vials containing different antibiotics, it is possible to analyze the sensitivity to the antibiotics SIRE (streptomycin, isoniazid, rifampicin, ethambutol).

[0120] The same method can possibly be extended to other antibiotics, for example pyrazinamide.

[0121] The use of antibiotics (individually tested or in cocktails) allows to better interpret which antibiotic or antibiotics are most suitable for health treatment.

[0122] By analyzing over a period of time, by means of light scattering techniques, the two or more test tubes under examination, it was possible to verify a growth phase of the mycobacteria (vial one) and, simultaneously with the sampling, if the strain had developed resistance to one or more MDR antibiotics, and if so to which (based on the corresponding vials).

[0123] To obtain the desired results, several measurements were performed during the day for several consecutive days, in order to evaluate the growth rate of the mycobacteria. In this case, the cultures were monitored at regular times, for example every 5 minutes or every 30 minutes for 6 weeks.

[0124] For the whole duration of the analysis, the vials remained sealed, completely eliminating the risk of airborne transmission of mycobacteria.

[0125] As we have seen, the bacterial detection method using analysis with a mathematical algorithm of data coming from light scattering tests is simple to use and provides clear results; moreover, the method described above is extremely safe and does not entail risks for the operator, as it does not expose the mycobacteria to the external environment, and therefore does not require the laboratory to meet BSL-3 safety standards.

[0126] In addition, the rapid identification of bacterial contamination allows the biological samples stored, if present, to be retested with minimal delay. Even more importantly, the parallel analysis of a second culture vial containing antibiotics (for example isoniazid and rifampicin), allows simultaneous phenotypic identification of MDR mycobacteria. In this way it also satisfies the requests made by the WHO, previously reported.

[0127] It is clear that modifications and/or additions of parts can be made to the bacterial detection method as described heretofore, without departing from the field and scope of the present invention.