Methods for predicting blood disorders

Abstract

The present invention relates to a method for determining the risk of developing a blood disorder, the method comprising the steps of: - exposing a biological sample to radiation to obtain a spectrum characteristic of the sample, which spectrum is processed to obtain a spectral signature; - comparing the spectral signature obtained in the previous step with one or more reference spectral signatures; and - concluding: * if the individual's spectral signature differs significantly from the spectral signature of a control, concluding that the individual has a high likelihood of developing a blood disorder; * otherwise concluding that the individual has a low likelihood of developing a blood disorder.

Description

[Technical Field]

[0001] The present invention relates to a method for predicting blood disorders. [Background technology]

[0002] The field of hematological disorders generally requires early screening, diagnosis, and tracking of disease progression toward more serious conditions, and typically involves multiple blood tests and analyses that may be performed on specific biological samples.

[0003] Myelodysplastic syndrome (MDS) is a pre-leukemic condition that is becoming more frequent with the aging of the population, and presents with progressive progression, with 30% of cases progressing to secondary acute leukemia. Diagnosis of MDS requires bone marrow collection for cytological analysis by a biologist specializing in hematology (bone marrow morphology).

[0004] Such analyses are invasive, lengthy, and costly, making it important to simplify the analysis and obtain results more quickly.

[0005] Fourier transform infrared (FTIR) spectroscopy is a well-known technique used to identify organic compounds and to examine the biochemical composition of biological samples (tissues or liquids).

[0006] Processes such as leukemia development can cause overall changes in cellular biochemistry, leading to differences in absorption spectra when analyzed by FTIR spectroscopy. Therefore, FTIR spectroscopy is commonly used to differentiate normal and abnormal tissues by analyzing changes in the absorption bands of macromolecules such as fats, proteins, carbohydrates, and nucleic acids.

[0007] Furthermore, the teachings in Patent Document 1 are known from the prior art, and these teachings describe a method and system for detecting and monitoring hematological cancers. More specifically, the inventors cited in this application have confirmed that mononuclear cell samples obtained from leukemia patients produce different FTIR spectra from those of healthy controls and non-cancer patients with leukemia-like clinical symptoms, such as subjects suffering from fever, thus enabling differential diagnosis of leukemia patients. By differentiating between leukemia patients, patients with leukemia-like clinical symptoms, and healthy control subjects, IR spectroscopy provides an effective diagnostic tool for diagnosing leukemia and / or other types of hematological malignancies.

[0008] However, while this literature demonstrates the use of FTIR spectroscopy in the context of leukemia diagnosis, it does not provide tools that enable the refinement of diagnosis based on disease subtype or the prediction of disease onset.

[0009] Thus, there is a need to provide methods that can improve the detection of leukemia, particularly subtypes of myeloid leukemia. [Prior art documents] [Patent Documents]

[0010] [Patent Document 1] International Publication No. 2011 / 121588 [Overview of the project] [Problems that the invention aims to solve]

[0011] The present invention aims to overcome these shortcomings of the prior art.

[0012] One of the objectives of the present invention is to provide a prediction method that enables the prediction of the onset of blood disorders in an individual in a simple manner. [Means for solving the problem]

[0013] The present invention thus relates to a method for determining, from a biological sample of an individual, in particular in vitro, the risk that the individual will develop a blood disorder, the method comprising: - exposing the biological sample to mid-infrared radiation (MIR) having wavelengths varying from 4000 cm -1 to 400 cm -1 to obtain a spectrum characteristic of the sample, the spectrum being processed to obtain a spectral signature composed of absorption peaks characteristic of the type and relative concentration of the various molecules present in the sample, by their position or wave number, and by their intensity or absorbance; - comparing the spectral signature obtained in the previous step with one or more reference spectral signatures, the one or more reference spectral signatures being obtained from a reference population of individuals; - a step of drawing a conclusion, * concluding that the individual has a high likelihood of developing a blood disorder if the intensity of a first group of peaks of the individual's spectral signature is significantly different from the intensity of these same peaks obtained in the reference spectral signature; The first group of peaks corresponds to the wave numbers of the first group of 1330 cm -1 , 1445 cm -1 , 1478 cm -1 , 1493 cm -1 , 1505 cm -1 , 1507 cm -1 , 1520 cm -1 , 1526 cm -1 , 1544 cm -1 , 1571 cm -1 , 1602 cm -1 , 1668 cm -1 , 1674 cm -1 , 1676 cm -1 , 1697 cm -1 , and 2852 cm -1 , and * otherwise concluding that the individual has a low likelihood of developing a blood disorder. relates to a method comprising.

[0014] This invention is based on the remarkable discovery made by the inventors that determining the infrared spectrum of a simple blood sample obtained from an individual makes it possible to obtain information related to the risk of that individual developing a blood disorder. [Modes for carrying out the invention]

[0015] Spectroscopy is a simple and rapid method that does not require reagents for implementation (other than appropriate materials) and allows us to obtain information related to the polymer structure of compounds contained in biological samples. Typically, infrared spectroscopy (FTIR) consists of multiple absorption bands, each corresponding to specific functional groups associated with cellular components such as fats, proteins, carbohydrates, and nucleic acids. All physiological changes that occur in an individual, including carcinogenesis, can lead to overall changes in metabolism, which will alter the absorption spectrum when a sample is analyzed using FTIR technology. Therefore, FTIR is commonly used to differentiate normal and abnormal tissues by analyzing changes in the absorption bands of molecules.

[0016] The infrared portion of the electromagnetic spectrum is divided into three regions, designated in relation to visible light: near-infrared, mid-infrared, and far-infrared. Far-infrared radiation is approximately 400-10 cm². -1 (The spectrum spans 1000-25 μm, practically 1000-30 μm), adjacent to the microwave region, possessing low energy, and can be used for rotational spectroscopy. Mid-infrared radiation is approximately 4000-400 cm⁻¹. -1 (Across a spectrum of 25–2.5 μm, practically 30–1.4 μm) it can be used to investigate the fundamental vibrations of the relevant vibrational structures. Near-infrared light is of higher energy, approximately 14,000–4,000 cm⁻¹. -1 Harmonic vibrations can be excited over a spectrum of 2.5 to 0.7 μm, and practically 1.4 to 0.8 μm. The designation and classification of these sub-regions are basically by convention. In this invention, the above definitions are used as a reference to spectroscopy in the mid-infrared region.

[0017] The infrared spectrum of a sample is established by passing an infrared light beam through the sample. Examining the transmitted light reveals the amount of energy absorbed at each wavelength. This can be achieved by using a monochromatic beam with a wavelength change over time, or by simultaneously measuring all absorbances by interferometry using a Fourier transform instrument. Thus, it is possible to generate absorbance or transmission spectra and analyze the absorption wavelengths. Analysis of these features reflects the molecular structure of the sample.

[0018] This technique works almost exclusively with samples containing covalent bonds. Simple spectra are obtained from highly pure samples with few active bonds in the infrared region. More complex molecular structures lead to more absorption bands and thus more complex spectra, but this method is still used for characterizing very complex mixtures.

[0019] The method described in this invention lies in the simplicity of the steps performed. -The first step is to obtain an infrared spectrum from an individual's biological sample. - The second step is to compare the spectrum obtained in the previous step with one or more reference spectra in order to conclude about the presence or risk of developing a blood disorder.

[0020] In the present invention, "blood disorder" means any medical condition that affects the components of the blood, and in particular means malignant blood disorders such as leukemia, lymphoma, myeloma, and myelodysplastic syndrome and myeloproliferative syndrome.

[0021] In this invention, in order to obtain a spectral signature characteristic of a given sample, the results of exposure to infrared light are processed, in particular, by a Fourier transform.

[0022] More specifically, 1330cm -1 , 1445cm -1 , 1478cm-1 , 1493cm -1 , 1505cm -1 , 1507cm -1 , 1520cm -1 , 1526cm -1 , 1544cm -1 , 1571cm -1 , 1602cm -1 , 1668cm -1 , 1674cm -1 , 1676cm -1 , 1697cm -1 , and 2852cm -1 The focus is on a specific wavenumber (the reciprocal of the wavelength), and the relative intensity of each peak corresponding to that wavenumber that follows the Fourier transform.

[0023] Thus, in summary, the given sample is at least 1330 cm². -1 , 1445cm -1 , 1478cm -1 , 1493cm -1 , 1505cm -1 , 1507cm -1 , 1520cm -1 , 1526cm -1 , 1544cm -1 , 1571cm -1 , 1602cm -1 , 1668cm -1 , 1674cm -1 , 1676cm -1 , 1697cm -1 and 2852cm -1 In order to obtain a spectral signature at a certain wavenumber, the system is exposed to mid-infrared radiation to obtain a spectrum that will be processed by a Fourier transform.

[0024] Once the spectral signature is obtained, it is compared to a reference spectral signature, or to multiple reference spectral signatures.

[0025] The reference spectral signature is obtained from a reference sample that has undergone the same infrared processing (and Fourier transform) as the sample being analyzed. For more effective comparison, it is essential that the reference spectral signature is obtained from the same type of biological sample (e.g., blood, serum, plasma, etc.) as the biological sample being tested. Thus, for example, if a blood sample is being tested for a certain individual, according to the method of the present invention, the reference sample will be obtained from another blood sample.

[0026] The reference spectral signature is obtained from a reference individual, which may be either a healthy individual, i.e., an individual without any disease, or an individual with a disease whose symptoms are different from those of a blood disorder as defined in this invention.

[0027] The reference individual may correspond to the individual being tested by the method of the present invention, and the reference sample is collected before the individual develops a blood disorder or before the likelihood of developing one increases.

[0028] During the comparison between an individual's spectral signature and a reference spectral signature, the intensities (absorbance) of various peaks corresponding to the aforementioned wavenumbers are compared.

[0029] In this comparison, if the intensities of all peaks corresponding to a given wavenumber differ significantly (increase or decrease) from the intensities of the same peaks in the reference spectral signature, then the individual whose sample has been examined by the method of the present invention is likely to develop a blood disorder. A given individual's signature thus consists of a pattern-type structure distributed across a set of spectral variables (16 in this embodiment). The reference signature (whether healthy or pathological) is thus formed by this "profile," or pattern. For each type of patient, a profile specific to the individual's physiological state is identified. Identifying any individual's physiological state is thus based on a comparison of this profile (or pattern) to the reference profile (or pattern). Calculating the distance between the individual's pattern and the reference pattern makes it possible to assign the individual to a specific class / category (e.g., healthy or diseased). The individual is "classified" according to the closest reference pattern in a 16-dimensional space (e.g., in this example).

[0030] Thus, it is possible to determine an individual's risk of developing malignant hematological disorders based on FTIR spectra obtained from biological samples of blood, blood by-products, or bone marrow.

[0031] Advantageously, the present invention, in the manner described above, further applies when an individual is at high risk of developing a blood disorder. * If the intensity of the second group of peaks in an individual's spectral signature differs significantly from the intensity of these same peaks obtained in the reference spectral signature, it is concluded that the individual is at high risk of developing leukemia. The second peak group was 3316 cm. -1 , 3283cm -1 , 3281cm -1 , 3256cm -1 , 3118cm -1 , 3116cm -1 , 1345cm -1 , 1343cm -1 , 1340cm -1, and 1338cm -1 Corresponding to the wavenumber of the first group, * Otherwise, conclude that the individual is at high risk of developing myelodysplastic syndrome. Regarding the method.

[0032] The inventors have demonstrated that, if a first group of peaks, i.e., wavenumbers, can determine whether an individual is likely to develop a blood disorder, then by examining a second group of peaks in the spectral signature, and according to the difference obtained, it is possible to determine whether an individual examined by the method of the present invention is likely to develop leukemia or myelodysplastic syndrome.

[0033] Myelodysplastic syndrome (MDS) is a clonal blood disorder originating from hematopoietic stem cells in the medulla, characterized by excessive proliferation of abnormally differentiated myelogenic cells (myelodysplasia). Excessive apoptosis of progenitor cells leads to insufficient production and peripheral cytopenia (ineffective hematopoiesis).

[0034] These are divided into several classes, as defined by the WHO (2016), depending on the type and number of cytopenia, signs of myelodysplasia (malformations of meningeal cells), and the presence or absence of blast excess.

[0035] In 70% of cases, the onset is delayed, relatively gradual, and accompanied by progressive worsening of cytopenia (bone marrow failure). On the other hand, in 30% of cases, the onset is rapid due to the accumulation of blast cells, and the disease is more invasive, progressing towards acute myeloid leukemia, which is why MDS is also called a "pre-leukemic state."

[0036] According to this embodiment, examining two first groups of wavenumbers in the spectral signature does not allow for differentiation between MDS that should be called "low-risk" and MDS that should be called "high-risk" in terms of progression to secondary leukemia.

[0037] In this invention, a distinction is made between two types of leukemia, particularly two types of acute myeloid leukemia, namely, primary acute myeloid leukemia and secondary myeloid leukemia.

[0038] Primary leukemia occurs spontaneously and directly in patients without being detected by them prior to myeloproliferative syndromes. Such leukemia is thought to be caused by the simultaneous accumulation of abnormalities affecting the proliferation and differentiation of myeloid progenitor cells.

[0039] On the other hand, secondary acute myeloid leukemia develops as myeloproliferative syndrome worsens, particularly through the accumulation of genetic abnormalities that inhibit the differentiation of progenitor cells.

[0040] In an advantageous embodiment, the present invention relates to the method described above, where an individual is at high risk of developing myelodysplastic syndrome. * If the intensity of the third group of peaks in an individual's spectral signature differs significantly from the intensity of these same peaks obtained in the reference spectral signature, it is concluded that the individual is at high risk of developing low-risk myelodysplastic syndrome. The third peak was 3060 cm. -1 , 3062cm -1 , 3396cm -1 , 3384cm -1 , and 3052cm -1 Corresponding to the wavenumber of the first group, * Otherwise, the method for concluding that an individual is at high risk of developing myelodysplastic syndrome.

[0041] The first, second, and third groups of peaks in the spectral signature can be used to differentiate between the occurrence of low-risk or high-risk myelodysplastic syndromes.

[0042] Advantageously, the present invention relates to the method described above, in which case an individual is at high risk of developing leukemia. * If the intensity of the fourth group of peaks in an individual's spectral signature differs significantly from the intensity of these same peaks obtained in the reference spectral signature, it is concluded that the individual is at high risk of developing secondary leukemia. The fourth peak was 3270 cm. -1 , 3268cm -1 , 3266cm -1 , 3264cm -1 , 3192cm -1 , 3190cm -1 , 2850cm -1 , 2840cm -1 , 1707cm -1 , 1705cm -1 , 1664cm -1 , 1662cm -1 , 1633cm -1 , 1631cm -1 , 1493cm -1 , 1491cm -1 , 1489cm -1 , 1458cm -1 , 1456cm -1 , and 1256cm -1 Corresponding to the wavenumber of the first group, * Otherwise, it is concluded that the individual has a high probability of developing primary leukemia. Regarding the method.

[0043] The first, second, and fourth groups of peaks in the spectral signature can be used to differentiate between the occurrence of primary leukemia and secondary leukemia.

[0044] More advantageously, the present invention relates to the method described above, wherein the biological sample is a plasma sample.

[0045] A favorable biological sample for implementing the present invention is plasma, which can be obtained during routine blood tests.

[0046] Plasma is the liquid part of blood. Plasma occupies about 55% of the blood volume and serves to transport blood cells, platelets, hormones, and other soluble components (proteins, metabolites, hormones, salts, etc.) throughout the body.

[0047] Advantageously, the present invention relates to the above-described method in which the spectral signature and the reference spectral signature are obtained via the second derivative of the respective infrared spectroscopy data.

[0048] The calculation of the second derivative of each of the spectra is advantageously carried out. The second derivative improves the resolution in the infrared region and the discrimination of the peaks obtained. The second derivative of the infrared spectrum provides a clear improvement over the use of the raw spectrum (non-derivative) for the characterization and identification of the compounds contained in the sample.

[0049] This process is generally carried out using software incorporated in the spectrometer. <​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​-1 , and 2852cm -1 Corresponding to the wavenumber, * Otherwise, conclude that the individual is unlikely to develop a blood disorder. Regarding the use of the sample.

[0051] Advantageously, the present invention is applicable in the above-described use when an individual is at high risk of developing a blood disorder. * If the intensity of the peak corresponding to the second wavenumber group of the spectral signature obtained by infrared spectroscopy for a sample differs significantly from the intensity of the same peak obtained from the spectral signatures of one or more control individuals, it is concluded that the individual has a high probability of developing leukemia. The second wavenumber group is 3316 cm. -1 , 3283cm -1 , 3281cm -1 , 3256cm -1 , 3118cm -1 , 3116cm -1 , 1345cm -1 , 1343cm -1 , 1340cm -1 , and 1338cm -1 Corresponding to the wavenumber, * Otherwise, conclude that the individual is at high risk of developing myelodysplastic syndrome. Regarding use.

[0052] Furthermore, the present invention, in the use described above, - If an individual is at high risk of developing myelodysplastic syndrome, * If the intensity of the peak corresponding to the third wavenumber group in the spectral signature obtained by infrared spectroscopy for a sample differs significantly from the intensity of the same peak obtained from the spectral signatures of one or more control individuals, it is concluded that the individual is highly likely to develop low-risk myelodysplastic syndrome. The third wavenumber group is 3060 cm. -1 , 3062cm -1 , 3396cm -1 , 3384cm -1 , and 3052cm-1 Corresponding to the wavenumber, * Otherwise, it is determined that the individual is at high risk of developing myelodysplastic syndrome. - If an individual is at high risk of developing leukemia, * If the intensity of the peak corresponding to the fourth wavenumber group in the spectral signature obtained by infrared spectroscopy for a sample differs significantly from the intensity of the same peak obtained from the spectral signatures of one or more control individuals, it is concluded that the individual has a high probability of developing secondary leukemia. The fourth wavenumber group is 3270 cm. -1 , 3268cm -1 , 3266cm -1 , 3264cm -1 , 3192cm -1 , 3190cm -1 , 2850cm -1 , 2840cm -1 , 1707cm -1 , 1705cm -1 , 1664cm -1 , 1662cm -1 , 1633cm -1 , 1631cm -1 , 1493cm -1 , 1491cm -1 , 1489cm -1 , 1458cm -1 , 1456cm -1 , and 1256cm -1 Corresponding to the wavenumber, * Otherwise, conclude that the individual is at high risk of developing primary leukemia. Regarding use.

[0053] The present invention further relates to a computer program product or software, which includes a portion, means, or program code instruction for performing the steps of the methods defined above when the program is executed on a computer.

[0054] Advantageously, the program may be contained on a computer-readable data storage medium. This type of medium is not limited to portable storage media such as CD-ROMs, but may also be part of a device including internal memory, one in a computer (e.g., RAM and / or ROM), one in an external memory device such as a hard disk or USB key, or one on a nearby or remote server.

[0055] To the advantage of the aforementioned computer program products or software, - Processing of infrared spectra to obtain spectral signatures, - Comparison of spectral signatures obtained from biological samples of the examined individual with a reference spectral signature, or - Both It is designed to make this possible.

[0056] The aforementioned computer program products or software can also be advantageously used to form the second derivative of the spectrum obtained after the Fourier transform.

[0057] The present invention will be better understood in light of the figures and the following examples. [Brief explanation of the drawing]

[0058] [Figure 1] This table shows the combined results of the group classifications G0, G1, G2, G3, and G4. G0: Healthy, G1: Low-risk myelodysplastic syndrome, G2: High-risk myelodysplastic syndrome, G3: Secondary acute myeloid leukemia, G4: Primary acute myeloid leukemia. [Examples]

[0059] (Example 1: Classification) (material and method) A - Biological sample Plasma was separated from whole blood samples (n=70) of individuals who provided consent and were unwell, as included in the MYLESYM test (ID-RCE 2011-A00271-40), by double centrifugation (700g, 10 minutes). This plasma was compared to plasma from 99 healthy donors recruited through the HEALTHOX test (ClinicalTrials.gov#NCT02789839).

[0060] The plasma sample (50 μL) is frozen at -80°C until use. After thawing at room temperature and homogenizing using a vortex-type agitator, 5 μL is spread onto a 96-position multiwell plate made of silica or zinc selenide (ZnSe), which are transparent under IR, dried in a sterilizer at 35°C for 15 minutes, and analyzed using an MIR spectrophotometer.

[0061] Alternatively, the sample (20 μL) may be placed on a microscope slide and left in the air for 24 hours to dry. B-plasma sample acquisition Sample size n=169 (99 healthy individuals / 70 individuals with health problems) G0: Health (60 women, 39 men).

[0062] G1: Low-risk myelodysplastic syndrome: LR-MDS (12 females, 26 males).

[0063] G2: High-risk myelodysplastic syndrome: HR-MDS (5 women, 4 men).

[0064] G3: Secondary acute myeloid leukemia: AML Sec (3 women, 8 men).

[0065] G4: Newly diagnosed acute myeloid leukemia: AML-Novo (9 females, 7 males). C-LUMOS microscope (Bruker) LUMOS is an autonomous IRTF microscope with a built-in spectrometer. Technological innovations, including a motor-driven crystal, allow the system to transition from transmission mode to reflection and ATR modes without operator intervention, enabling fully automated measurement of the sample or background noise, even when ATR mode is activated. This type of instrument is suitable for attenuated total internal reflection (ATR) measurements when the sample is placed on a glass slide, which is a material that is not transparent in the mid-infrared region.

[0066] A similar instrument to the Bruker IR Biotyper may also be used. This instrument is driven by OPUS software, which belongs to Bruker. This type of spectrophotometer makes it easy to acquire approximately 100 spectra per day, including plate preparation. The infrared spectra are thus connected in "transmission" mode, and the infrared beam passes through the sample and a multi-well plate made of crystallized ZnSe, a material that is transparent in the mid-infrared region.

[0067] In either case (measurement of reflection or transmission), the spectral resolution is 4 cm. -1 The scans are then averaged over 64 to 128 scans. Background noise is measured through an empty well. The "raw" absorption spectrum is then saved and exported to Jcamp format ("open" format) using a macro routine under OPUS. D - Quality of inspection To evaluate spectral quality based on multiple parameters, the quality of the spectrum is assessed based on several parameters such as water vapor content, signal / water ratio, and noise intensity, and abnormal spectra (outliers) that do not meet certain criteria are identified. To confirm the hydration state, the amide I band (1650 cm⁻¹) of the protein is used. -1 ) essentially reflects liquid water (3400cm -1 It is certain that it is 2 to 3 times larger than ). E-baseline correction Baseline variations can be caused by changes in conditions between acquisitions, or by variations related to equipment or the environment (e.g., temperature). F-normalization To minimize signal intensity differences caused by the instrument rather than the sample, the raw spectrum is normalized using the MSC (Multiplicative Scatter Correction) scattering prevention algorithm. This is a spectral correction method (Sun. D-Wet al. 2009). G-filtering This procedure involves selecting the target spectral range depending on the sample. For the target sample (plasma), 3800-940 cm⁻¹ is used. -1 The spectral range is fixed: 2800~1800 cm⁻¹ -1 In this band, the spectrum is truncated because it does not contain information relevant to the analysis being performed. This spectrum mainly includes the contribution of atmospheric CO2 and reflects environmental variability. H-second derivative The derivative allows for improved spectral resolution and thus reduces the effects of band overlap. It should be noted that the transition from the raw spectrum to the second derivative reduces the signal-to-noise ratio {Martens H et al. 2002}. The second derivative of the spectrum is calculated using 13 points in a sliding window type Savitsky-Golay smoothing.

[0068] (Analysis of spectral data) A-Statistical methods 1. Unmonitored analysis (descriptive analysis) PCA: Principal Component Analysis. This is the best analysis for understanding data structure and identifying potential spectra, called outliers, that exhibit different spectral profiles due to technical reasons such as improper acquisition, or biochemical reasons. 2. Monitored analysis (explanatory analysis) PLSR: Least Squares Regression (Partial Least Squares Regression). This is a statistical method that makes it possible to model the complex relationship between observed quantitative variables called manifest variables and latent variables (MIR spectrum). B - Selection of Variables The selection of variables using genetic algorithms or FADA methods makes it possible to identify a subset of discriminant variables to define the types of biochemical markers that are altered by disease status (Trevisan J et al. 2014). This means that • Improvement in the performance of the prediction model (Jouan-Rimbaud D et al. 1995), and • Improved model interpretation and understanding of the systems being investigated It has the following two advantages. C-FADA and GLM algorithms LDA / logistic regression analysis allows for the identification of the most discriminative spectral variables between different groups of subjects, in this case healthy and unwell. Based on these most discriminative variables, incremental selection is made to identify several variables that enable the best specificity and selectivity. D-Prediction Principle The results of discriminant analysis tests are traditionally presented in the form of a confusion matrix and are interpreted as shown below in Table 1.

[0069] [Table 1]

[0070] The results obtained from this study are shown in Table 2 and Figure 1 below.

[0071] [Table 2]

[0072] This study demonstrates that myelodysplastic syndromes and acute leukemia (primary or secondary) are accompanied by distinct metabolic changes revealed through specific IVIIR spectral signatures (specific "barcodes"). This opens up interesting possibilities for early and rapid diagnosis: • Identification of plasma molecules that reflect physiological pathological conditions and can serve as important biomarkers (interpretation of spectral signatures); • Support for the early detection of myelodysplastic syndrome, and • Support for tracking MDS patients.

[0073] (Example 2) Files imported under OPUS are then imported into a matrix and transposed using software written in the R environment.

[0074]

number

number

[0075] At the end of this program, an Excel file will be created, and its first tab will contain the transpose matrix of all samples to be processed (1 sample = 1 row).

[0076] The subsequent step involves calculating the second derivative of each spectrum, smoothing these derivatives over 11 or 13 points using the Savitzky-Golay sliding window routine, and then truncating these derivatives to retain only the frequency region relevant to the analysis. The retained spectral region is, in almost all cases, 3800–2800 cm⁻¹. -1 and 1800~700cm -1These are then normalized according to the principle of vector normalization (the area under the second derivative is normalized to 1). The matrix of the second derivative is saved in the second tab of the same Excel file. The following script performs this preprocessing.

[0077]

number

number

number

[0078] Note: These steps can be performed using any type of computational software, as the mathematical operations are standard. However, it is important to follow the order in which they are performed.

[0079] Some authors prefer working with raw spectra that have been corrected for scattering (multiple scattering correction or MSC routines). The inventors found better performance when working with second derivatives.

[0080] These truncated and normalized second derivatives are used to calibrate the predictive model.

[0081] The predictive model is based on a PLSR-type (Partial Least Squares Regression) analysis, which allows for the identification of the most discriminative spectral variables between the two groups. These variables are ordered according to the number of times they have been positively selected over a large number of iterations (typically 100). A manual check is then performed to reduce the number of variables that need to be considered in the predictive model as much as possible. Each time (for each combination of variables), a confusion matrix is ​​calculated, allowing for the identification of correctly classified and incorrectly classified samples.

[0082] Once this optimization is performed, validation is carried out by predicting samples that did not play a role in calibrating the predictive model. The following R script makes it possible to perform these tasks.

[0083]

number

number

number

number

number

[0084] These results are shown in the form of second derivative spectra, identification of markers (differentiating variables), and confusion matrices, as specified in Figure 1.

[0085] The present invention is not limited to the embodiments shown, and other embodiments will be clearly apparent to those skilled in the art.

Claims

1. A computer program product for providing information regarding the risk of developing a blood disorder from a biological sample of blood, blood by-products, or bone marrow of an individual at risk of developing a blood disorder, wherein when the computer program is executed on a computer, - A step of exposing the biological sample to mid-infrared radiation (MIR) with wavelengths varying from 4000 cm⁻¹ to 400 cm⁻¹ to obtain a spectrum characteristic of the biological sample, wherein the spectrum is processed by a Fourier transform to obtain a spectral signature consisting of characteristic absorbance peaks, which are determined by the position or wavenumber and intensity or absorbance of various molecules present in the biological sample, depending on their type and relative concentration. - A step of comparing the spectral signature obtained in the previous step with one or more reference spectral signatures, wherein the one or more reference spectral signatures are obtained from a reference group of individuals. Includes a portion, means, or program code instruction for executing, moreover, - This is the step to reach a conclusion. * If the intensity of the first group of peaks in the spectral signature of the individual differs from the intensity of the same peaks obtained in the reference spectral signature, then it is concluded that the individual is at high risk of developing a blood disorder. The first group of the peaks corresponds to the wavenumbers of 1330 cm －１ , 1445 cm －１ , 1478 cm －１ , 1493 cm －１ , 1505 cm －１ , 1507 cm －１ , 1520 cm －１ , 1526 cm －１ , 1544 cm －１ , 1571 cm －１ , 1602 cm －１ , 1668 cm －１ , 1674 cm －１ , 1676 cm －１ , 1697 cm －１ , and 2852 cm －１ , and * If not, conclude that the individual is unlikely to develop a blood disorder. A computer program product that includes parts, means, or program code instructions for performing a certain action.

2. The computer program product according to claim 1, wherein the biological sample is a plasma sample.

3. The computer program product according to claim 1 or 2, wherein the spectrum and the control spectrum are obtained by the second derivative of infrared spectroscopic data.

4. If the aforementioned individual is at high risk of developing a blood disorder, further, * If the intensity of the second group of peaks in the spectral signature of the individual differs from the intensity of the same peaks obtained in the reference spectral signature, it is concluded that the individual has a high probability of developing leukemia. The second group of the aforementioned peaks is 3316 cm. －１ , 3283cm －１ , 3281cm －１ 3256cm －１ , 3118cm －１ 3116cm －１ , 1345cm －１ , 1343cm －１ , 1340cm －１ , and 1338 cm －１ Corresponding to the wavenumber, * Otherwise, it is concluded that the individual is highly likely to develop myelodysplastic syndrome. A computer program product according to any one of claims 1 to 3.

5. If the aforementioned individual is at high risk of developing myelodysplastic syndrome, further, * If the intensity of the third group of peaks in the spectral signature of the individual differs from the intensity of these same peaks obtained in the reference spectral signature, it is concluded that the individual is likely to develop low-risk myelodysplastic syndrome. The third group of the aforementioned peaks is 3060 cm. －１ , 3062cm －１ , 3396cm －１ 3384cm －１ , and 3052cm －１ Corresponding to the wavenumber, * Otherwise, it is concluded that the individual is likely to develop a high-risk myelodysplastic syndrome. The computer program product according to claim 4.

6. If the aforementioned individual is at high risk of developing leukemia, * If the intensity of the fourth group of peaks in the spectral signature of the individual differs from the intensity of the same peaks obtained in the reference spectral signature, then it is concluded that the individual has a high probability of developing secondary leukemia. The fourth group of the aforementioned peaks is 3270 cm. －１ , 3268cm －１ 3266cm －１ , 3264cm －１ , 3192cm －１ , 3190cm －１ , 2850cm －１ , 2840cm －１ , 1707cm －１ , 1705cm －１ 1664cm －１ , 1662cm －１ , 1633cm －１ , 1631cm －１ 1493cm －１ 1491cm －１ , 1489cm －１ , 1458cm －１ 1456cm －１ , and 1256 cm －１ Corresponding to the wavenumber, * Otherwise, we conclude that the individual is highly likely to develop primary leukemia. The computer program product according to claim 4.

7. Use of a computer program product according to any one of claims 1 to 3, the use comprising the steps of concluding the following: * If, with respect to the biological sample, the intensity of the peak corresponding to the first group of wavenumbers in the spectral signature obtained by infrared spectroscopy differs from the intensity of the same peak obtained from the spectral signatures of one or more control individuals, then the individual is likely to develop a blood disorder. The first group of wavenumbers is 1330 cm －１ , 1445cm －１ , 1478cm －１ 1493cm －１ , 1505cm －１ , 1507cm －１ , 1520cm －１ 1526cm －１ 1544cm －１ 1571cm －１ , 1602cm －１ , 1668cm －１ 1674cm －１ 1676cm －１ 1697cm －１ , and 2852cm －１ Corresponding to the wavenumber, * Otherwise, the likelihood of the aforementioned individual developing a blood disorder is low.

8. Use of the computer program product according to claim 7, further performing the steps of concluding the following: - If the aforementioned individual is at high risk of developing a blood disorder, * If, for the aforementioned biological sample, the intensity of the peak corresponding to the second wavenumber group of the spectral signature obtained by infrared spectroscopy differs from the intensity of the same peak obtained from the spectral signature of one or more control individuals, then the individual is likely to develop leukemia. The second group of wavenumbers is 3316 cm⁻¹. －１ , 3283cm －１ , 3281cm －１ 3256cm －１ , 3118cm －１ 3116cm －１ , 1345cm －１ , 1343cm －１ , 1340cm －１ , and 1338 cm －１ Corresponding to the wavenumber, * Otherwise, the individual is highly likely to develop myelodysplastic syndrome.

9. Use of the computer program product according to claim 8, further comprising the step of concluding the following: - If the aforementioned individual is highly likely to develop myelodysplastic syndrome, * If the intensity of the peak corresponding to the third wavenumber group of the spectral signature obtained by infrared spectroscopy for the aforementioned biological sample differs from the intensity of the same peak obtained from the spectral signature of one or more control individuals, then the individual is likely to develop a low-risk myelodysplastic syndrome. The third group of wavenumbers is 3060 cm⁻¹. －１ , 3062cm －１ , 3396cm －１ 3384cm －１ , and 3052cm －１ Corresponding to the wavenumber, * Otherwise, the individual is likely to develop a high-risk myelodysplastic syndrome. - If the aforementioned individual is highly likely to develop leukemia, * If, for the aforementioned biological sample, the intensity of the peak corresponding to the fourth wavenumber group of the spectral signature obtained by infrared spectroscopy differs from the intensity of the same peak obtained from the spectral signature of one or more control individuals, then the individual is highly likely to develop secondary leukemia. The fourth group of the wavenumbers corresponds to wavenumbers of 3270 cm －１ , 3268 cm －１ , 3266 cm －１ , 3264 cm －１ , 3192 cm －１ , 3190 cm －１ , 2850 cm －１ , 2840 cm －１ , 1707 cm －１ , 1705 cm －１ , 1664 cm －１ , 1662 cm<000,094>, 1633 cm －１ , 1631 cm －１ , 1493 cm －１ , 1491 cm －１ , 1489 cm －１ , 1458 cm －１ , 1456 cm －１ , and 1256 cm －１ and * Otherwise, the aforementioned individual is highly likely to develop primary leukemia.

Images