Method for assessing clinical outcome of patient in clinical environment using transcriptome score and device for implementing same

Abstract

The present invention relates to a method for determining a transcriptome score (TScore) suitable for assessing the risk of an undesirable clinical outcome of a patient in a clinical environment from an assay sample or population of assay samples previously extracted from the patient, the method comprising, among other steps, determining a transcriptome score (TScore) suitable for assessing the risk of an undesirable clinical outcome of the patient, gene expression values of at least two different genes selected from a predetermined set of genes are measured for a patient sample or a population of samples previously obtained from the patient, and a transcriptome score is determined, where the genes are assigned or not assigned scores. The method may be computer implemented. The genes of the predetermined set of genes may be at least two genes selected from the group consisting of ADGRE3, ARL14EP, BPGM, C3AR1, CCNB1IP1, CD177, CD274, CD3D, CD74, CIITA, CTLA4, CX3CR1, GNLY, IFNgamma, IL10, IL1R2, IL1RN, IL7R, IP10 / CXCL10, MDC1, OAS2, S100A9, TAP2, TDRD9, TNF, and ZAP70. The method may be performed on an assay biological sample extracted from a patient who has been treated in a resuscitation department, an intensive care unit, or a continued care unit. The invention also relates to a method of classifying samples previously taken from a patient as a group reflecting the risk of an undesirable clinical outcome in a clinical environment, or a method of identifying a patient at risk of an undesirable clinical outcome in a clinical environment, and an in vitro or ex vivo method of screening for whether a drug has the ability to alleviate an undesirable clinical outcome in a patient. The invention also relates to a computer device for carrying out the invention, and to the use of a kit for carrying out the method of the invention.

Description

Technical Field

[0001] The present invention relates to the field of investigation methods for determining the potential outcomes of clinical conditions of patients admitted to a nursing facility, particularly an intensive care unit (ICU). The present invention provides a method for determining a transcriptome score (TScore), which is suitable for assessing the risk of adverse clinical outcomes for patients in a clinical setting, the method being based on statistical processing of measured gene expression values. The method can be computer-implemented. The present invention also relates to a method for classifying samples previously extracted from a patient, particularly a patient previously admitted to an intensive care unit, into groups reflecting the risk of adverse clinical outcomes in a clinical setting, and to a method for identifying patients with adverse clinical outcome risks in a clinical setting. The present invention particularly relates to intensive care unit-acquired infections (ICU-AI), particularly the results of nosocomial infections. The present invention also relates to an in vitro or ex vivo method for screening whether a drug has the ability to alleviate adverse clinical outcomes based on stratifying the patient's immune background. The present invention also relates to a computer device for implementing the method executed by a computer and the purposes of a kit comprising them. Background Art

[0002] introduce

[0003] Sepsis is a complex, life-threatening syndrome caused by a dysregulated host response to infection (1). The primary inflammatory response is often followed by a complex and persistent immunosuppressive response affecting both innate and adaptive immune components. This post-septic immunosuppression is associated with a higher risk of secondary infections, the most prominent of which is intensive care unit (ICU)-acquired pneumonia (2).

[0004] However, there is growing evidence that there is important heterogeneity in the type and magnitude of the immune response during post-sepsis immunosuppression (3). This heterogeneity is also reflected at the patient level, with sepsis of pulmonary origin being associated with a higher risk of subsequent ICU-acquired pneumonia compared with sepsis of other origins (4). Furthermore, the recent COVID-19 pandemic has highlighted that the type of pathogen also influences the risk of ICU-acquired infection, with patients infected with SARS-CoV-2 having a higher susceptibility to secondary pneumonia compared with patients with influenza or severe bacterial pneumonia (5).

[0005] Although biomarkers for determining the risk of developing care-associated infections in patients are known, for example from WO 2010 / 082004, WO 2021 / 058919, WO 2022008827 or WO2022008828, in this context it is not only important to identify suitable diagnostic biomarkers, but also to develop tools that allow stratification of patients according to their risk of ICU-acquired infections, especially in secondary infections and in the earliest cases.

[0006] While transcriptomic approaches hold promise for informative stratification strategies in this setting, this has been shown to be most effective in the late stages of sepsis. In a prospective cohort of patients with sepsis, blood gene expression at the onset of ICU-acquired infection showed reduced expression of genes involved in gluconeogenesis and glycolysis (2). In ICU patients with ventilator-associated pneumonia (VAP), the authors reported transcriptomic repression of genes associated with the immune synapse in the blood. However, almost all studies have investigated transcriptomic responses during ICU-acquired infection rather than identifying patients at risk for further infection. Multiplexed molecular platforms, such as Recent developments in the BioMérieux System (bioMérieux) allow for rapid and reliable assessment of the transcriptomic response of patients during nosocomial infections based on the expression of multiple genes involved in pro-inflammatory and anti-inflammatory responses (6).

[0007] The present invention aims to propose a stratification tool, in particular based on the evaluation of circulating mRNA of genes involved in the immune response. The inventors have currently emphasized that the development of such a tool is limited by the heterogeneity of patients with sepsis. The aim of the study reported herein on which the present invention relies was to evaluate the diagnostic performance of a transcriptome score that, as a proof of concept, has been performed in a series of patients between days 5 and 7 after admission to the ICU. This transcriptome score can identify a subgroup of critically ill patients who are likely to exhibit adverse clinical outcomes, including a higher rate of ICU-acquired infections, a longer ICU stay and a higher mortality rate. Thus, the present invention demonstrates the relevance of the so-called transcriptome score reported herein, which was initially based on a practical combination of the outputs of immune-related genes detected with a prototype multiplex PCR tool, but defined by a specific rationale that makes it an unweighted score, a strategy that proved to be decisive, valuable and relevant to other tools obtained, for example, by machine learning.

[0008] Thus, the present invention is based on the experiments described herein and proposes novel devices and tools that are intended to solve any or all of the problems set forth above or that become apparent from this specification. Summary of the Invention

[0009] The present invention relates to a method for determining a transcriptome score (TScore) from an assay sample or a population of assay samples previously extracted from a patient, said transcriptome score being suitable for assessing the risk of adverse clinical outcomes for a patient in a clinical setting, said method comprising:

[0010] a. measuring gene expression values ​​of a patient sample or a population of patient samples previously obtained from a patient, or if the method is computer-implemented, retrieving such gene expression values ​​for at least two different genes selected from a predetermined set of genes, and

[0011] b. For each gene in the predetermined gene set, and for each sample, the

[0012] comparing the measurement value obtained or retrieved in step a. with a predetermined threshold value, said predetermined threshold value identifying the measurement sample as being associated with a patient exhibiting an infection or not exhibiting an infection (particularly an ICU-acquired infection), and / or determining a receiver operating characteristic (ROC) curve based on the measurement value obtained or retrieved in step a., said ROC curve being intended to represent the diagnostic ability of the gene expression value measured for said gene to identify the measurement sample as being associated with a patient exhibiting an infection or not exhibiting an infection (particularly an ICU-acquired infection), and

[0013] c. For each patient sample in the assay sample or assay sample population, calculate a transcriptome score, where:

[0014] i. for each gene in the predetermined set of genes that are individually analyzed, assigning a point to the gene if the measured gene expression value matches the criteria for classifying the sample as a sample associated with a patient exhibiting an infection, particularly an ICU-acquired infection, and

[0015] ii. summing the scores obtained for all genes in the predetermined gene set to obtain the transcriptome score of the assay sample, and

[0016] iii. Optionally, provide a transcriptome score as an output value.

[0017] "Assessing the risk of adverse clinical outcomes for patients in a clinical setting" means that the definition of the transcriptome score that is the core of the present invention is applicable to assessing the risk, and according to a specific embodiment, is applicable to assessing the risk. "Adverse clinical outcome" refers to the patient's clinical condition, which can be determined to be more serious or more harmful to the patient's health at the time of observation, or to further deteriorate at a subsequent observation time point, relative to the patient's apparent clinical condition observed when the method is implemented. Alternatively, it can be said that the method is used to assess the risk of a patient suffering from a comorbidity (whose sample has been analyzed by transcriptome scoring), or to assess the risk of a complication (a complication is defined as an unfavorable outcome of a disease, health condition, or treatment). According to a specific aspect, adverse clinical outcomes can mean an increased probability / risk of a patient developing an intensive care unit (ICU)-acquired infection (AI) (i.e., typically abbreviated as ICU-AI in this specification). According to a non-limiting list, ICU-AI can be ICU-acquired pneumonia, catheter-related bloodstream infection, or urinary tract infection. ICU-AI typically includes so-called "hospital infections," which are defined as infections that a patient acquires (or may have acquired) when admitted to a health care facility, and which typically develop after 48 hours, i.e., 48 hours or later after admission. Common nosocomial infections are urinary tract infections, respiratory pneumonia, surgical site wound infections, bacteremia, gastrointestinal infections, and skin infections. According to another aspect, an adverse clinical outcome can be a longer ICU hospital stay. More generally, an adverse clinical outcome can be a longer hospital stay in the hospital, which can also increase the patient's chance of being exposed to pathogens that cause nosocomial infections. Examples of longer ICU hospital stays are provided in this specification, and the skilled person can easily and comprehensively obtain these examples. According to another aspect, an adverse clinical outcome can be considered as a higher ICU mortality rate, that is, it can be manifested by the death of the patient.

[0018] Thus, "adverse clinical outcome" generally refers to any change in clinical condition that is detrimental to a patient's health. According to certain embodiments, "adverse clinical outcome" refers to the occurrence of ICU-AI. This condition may occur after a TScore as described herein is determined for a patient, or may be diagnosed for a patient after a TScore as described herein is determined for the patient.

[0019] A "patient in a clinical setting" generally refers to a patient in a nursing facility, i.e., a patient admitted to a nursing facility. According to more specific embodiments, a patient in a clinical setting refers to a patient who has been admitted to a resuscitation unit and / or an intensive care unit or a continuing care unit. According to specific embodiments applicable throughout this specification, a patient in a clinical setting is a patient admitted to an intensive care unit (ICU).

[0020] For the purposes of this method, and to provide the required gene expression values ​​for the method, one or more patient samples are removed from the patient's body beforehand and then measured. This means that the method is not performed directly on humans, i.e., it is an in vitro method, provided that gene expression values ​​are measured as part of the method. If the method is computer-implemented, the measured gene expression values ​​can be provided as input to the method at step a. of the method described above. The expression "retrieval of gene expression values" means that, if the method is computer-implemented, the gene expression values ​​are provided as input data to the computer-implemented method.

[0021] With respect to step a of the method for determining a transcriptome score (TScore) described above and herein, the "patient sample or a sample population previously obtained from a patient" may be from a patient as defined above. According to a specific embodiment, the "patient sample or a sample population previously obtained from a patient" is from a patient admitted to an intensive care unit (ICU).

[0022] It is noteworthy that, since the patients of the method for determining transcriptome scores (TScore) described above and herein are patients admitted to nursing institutions, particularly intensive care units (ICUs), these patients may have potential pathology or conditions, including severe forms of pathology or conditions. As described in the examples, according to a non-limiting list, patients may be susceptible to or diagnosed with sepsis, trauma (including severe trauma) or burns at different stages. The experimental section provides some examples, but these examples are not limited to the conditions encountered in intensive care units. Transcriptome scores (TScore) can be defined without considering the initial condition of the patient, nor the therapy, such as antibiotics, that the patient is currently receiving for the condition.

[0023] According to one specific non-limiting embodiment, the assay biological sample is obtained from a patient susceptible to, susceptible to, or diagnosed with, sepsis.

[0024] With regard to step a. of the method for determining a transcriptome score (TScore) described above and herein, according to a specific embodiment, the gene expression value is measured by mRNA expression level. In a specific embodiment, the mRNA expression level is measured in particular by a multiplex assay involving a plurality of assay genes. According to different specific embodiments, the gene expression value is measured by molecular methods (e.g. amplification methods), by sequencing (e.g. next generation sequencing) or by hybridization methods (e.g. hybridization microchips, All these methods are well known to the skilled person.

[0025] The skilled person will readily appreciate that in step a., "measuring the gene expression values ​​of a sample or sample group previously obtained from a patient" is preferably performed at a given time, particularly the same time for all assay genes whose expression values ​​are measured. The use of multiplex assays facilitates this configuration.

[0026] According to one specific non-limiting embodiment, the assay biological sample is taken from a patient admitted to an intensive care unit (ICU), preferably between day 5 and day 7 of ICU hospitalization.

[0027] According to a specific embodiment, gene expression values ​​are measured by RT-PCR, in particular RT-qPCR, more in particular nested RT-qPCR, in particular multiplex RT-PCR or RT-qPCR or nested RT-qPCR, e.g. using Technology (bioMérieux) or Biomark TM Platform (Fluidigm). The technical staff can easily implement these methods based on the guidance provided in this area and this paper. For example, the technical staff can easily find the mRNA sequence in NCBI or Ensembl database, and can use commonly available tools such as Geneious or Primer 3 to design the required primers or primer pairs for measuring mRNA expression levels.

[0028] According to a particular embodiment, the assay biological sample is a blood sample, preferably a whole blood sample.

[0029] In the context of step a. of the method of determining a transcriptome score (TScore) described above and herein, gene expression values ​​of at least two different genes selected from a predetermined gene set are measured or measurements are retrieved.

[0030] As detailed in the experimental section herein, for the purpose of determining the TScore, an advantage of the present method is that the genes under consideration are preferably known to be independent or simply assumed to be independent. The experimental section provides guidance on the meaning of such established or assumed independence. Thus, a skilled person can easily determine a subset of genes that can be included in the method, for example based on biomarkers whose correlations have been assessed in the literature, or based on assumptions guided by his / her knowledge. In fact, for the implementation of the method, it is not required that the genes are actually independent, as this artifact can grasp the heterogeneity / complexity of patient characteristics in the clinical setting and turn it into an advantage.

[0031] According to a specific embodiment, the genes of the predetermined gene set of step a. of the method for determining a transcriptome score (TScore) described above and herein are selected from the group consisting of: ADGRE3 (adhesion G protein-coupled receptor E3), ARL14EP (ribosylation factor-like GTPase 14 effector protein), BPGM (diphosphoglycerate mutase), C3AR1 (complement C3a receptor 1), CCNB1IP1 (cyclin B1 interacting protein 1), CD177 (CD177 molecule), CD274 (CD274 molecule), CD3D (CD3d molecule), CD74 (CD74 molecule), CIITA (class II major histocompatibility complex transactivator), CTLA4 (cytotoxic T lymphocyte-associated protein 4), CX3C R1 (C-X3-C motif chemokine receptor 1), GNLY (granulysin), IFNgamma (interferon gamma), IL10 (interleukin 10), IL1R2 (interleukin 1 receptor 2), IL1RN (interleukin 1 receptor antagonist), IL7R (interleukin 7 receptor), IP10 / CXCL10 (interferon gamma-induced protein 10), MDC1 (mediator of DNA damage checkpoint 1), OAS2 (2'-5'-oligoadenylate synthetase 2), S100A9 (S100 calcium-binding protein A9), TAP2 (transporter 2, ATP-binding cassette subfamily B member), TDRD9 (Tudor domain-containing protein 9), TNF (tumor necrosis factor), and ZAP70 (zeta chain of T-cell receptor-associated protein kinase 70). Those skilled in the art can find the chromosomal location and sequence of the predetermined gene set in the NCBI or Ensembl databases.

[0032] According to different specific embodiments, the number of genes measured is at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25 or 26. According to different specific embodiments, the number of genes measured is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25 or 26. According to other different specific embodiments, the number of genes measured is at least 2, 3, 4, 5, 6, 7 or 8, in particular any one of 2, 3, 4, 5, 6, 7 or 8.

[0033] According to step b. of the method for determining a transcriptome score (TScore) as described above and herein, and for each gene in the predetermined gene set, and for each sample, it is proposed to compare the measurement value obtained or retrieved in step a. with a predetermined threshold value, said predetermined threshold value identifying the assay sample as being associated with a patient exhibiting an infection or not exhibiting an infection (particularly an ICU-acquired infection), and / or to determine a receiver operating characteristic (ROC) curve intended to represent the diagnostic ability of the gene expression values ​​measured for said gene, to identify the assay sample as being associated with a patient exhibiting an infection or not exhibiting an infection (particularly an ICU-acquired infection).

[0034] "Comparing the measurement value obtained or retrieved in step a. with a predetermined threshold value" means comparing the measurement value with a threshold value for distinguishing between samples associated with patients exhibiting infection (particularly ICU-acquired infection) and samples associated with patients not exhibiting infection (particularly ICU-acquired infection).

[0035] The expression "a sample associated with a patient not exhibiting an infection" may be considered synonymous with the expression "a sample not associated with a patient exhibiting an infection" and, according to this rule, may be used indiscriminately within the meaning of the present invention.

[0036] "Determining receiver operating characteristic (ROC) curve" refers to calculating such a curve according to the knowledge of those skilled in the art, or depicting or drawing such a curve, the latter particularly referring to when the method is computer-implemented. The ROC curve is a well-known graphical method for displaying the accuracy of distinguishing between markers (diagnostic tests) for distinguishing between two populations. It can be used to study the effectiveness of diagnostic markers in distinguishing between sick and healthy individuals. Typically, if the marker value of the test is greater than a given predetermined threshold, the person is assessed as sick (positive), otherwise the subject is diagnosed as healthy (negative). The accuracy of any given threshold can be measured by the probability (sensitivity) of a true positive and the probability (specificity) of a true negative. The ROC curve is a graph of the sensitivity (Se(c)) vs. 1-specificity (1-Sp(c)) of all possible thresholds (c) of the marker under consideration. In the present invention, the marker is the expression level of a gene defined according to any embodiment described herein, and the diagnosis sought is the difference between a sample associated with a patient showing infection (particularly ICU-acquired infection) and a sample associated with a patient not showing infection (particularly ICU-acquired infection).

[0037] On the one hand, the threshold level can generally be determined by considering acquired or known data (gene expression levels) of healthy individuals (i.e., not exhibiting infection (particularly ICU-acquired infection)), as this is known for any diagnostic test. Such a level can be provided as a predetermined threshold value for the method of the present invention.

[0038] On the other hand, in order to obtain training test and related validation data, such as reference transcriptome scores and decision values, such as optimal thresholds (which will be discussed below), or even, according to a specific embodiment, thresholds of gene expression levels in the context of training test, the measurement values ​​or retrieved measurement data of step a. of the method for determining a transcriptome score (TScore) described above and herein are provided for the assay sample population, so that the ROC curve can be drawn in step b.

[0039] In this case, the patients of the training set from which samples are extracted before the implementation of the method of the present invention include both patients who exhibited infection at the time of measurement and patients who did not exhibit infection at the time of measurement, i.e., patients for whom it is known whether they have an infection, particularly an infection that occurred after they entered a care facility (according to the definition provided above) within the timeframe discussed herein, more particularly, patients for whom it is known whether they have had at least one episode of ICU-acquired infection (ICU-AI), particularly since they entered an intensive care unit (ICU). Guidance in this regard is provided in the experimental section of this article.

[0040] In contrast, when samples are tested against a predetermined threshold, such as for classification purposes, then a single sample can be assayed.

[0041] When the method for determining a transcriptome score (TScore) described above and herein is performed on a group of measurement samples previously extracted from a patient, wherein it is known whether a portion of the samples were extracted from patients who have exhibited at least one infectious episode since admission to a intensive care unit (particularly, admission to an ICU), this method may further include, after step b. of the method for determining a transcriptome score (TScore) described above and herein, a step of further selecting genes from the predetermined set of genes to be analyzed, for which the AUC value of the ROC curve drawn in step b. is at least 0.70.

[0042] According to a specific embodiment, the AUC value of the ROC curve drawn in step b. is greater than or equal to: 0.70, 0.71, 0.72, 0.73, 0.74, 0.75, 0.76, 0.77, 0.78, 0.79, 0.80, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86, 0.87, 0.88, 0.89, 0.90, 0.92, 0.93, 0.94, 0.95, 0.96, 0.97, 0.98, 0.99, or 1.

[0043] The area under the curve (AUC) is well known in the art and is a measure of a classifier's ability to distinguish between classes; it can be viewed as a summary of an ROC curve. The higher the AUC, the better the model's ability to distinguish between positive and negative classes.

[0044] For example, such an embodiment enables streamlining of the predetermined gene set for analysis.According to certain embodiments, the updated predetermined gene set may be provided as an output to the method.

[0045] As this is the practice of the skilled person, whenever a ROC curve is available to the skilled person, the AUC value will have specific sensitivity and specificity values ​​which are associated with the test in question and are therefore known to the skilled person.

[0046] According to a specific embodiment, before or after (if performed) selecting genes whose AUC values ​​meet the above requirements, the predetermined gene set used in the method of the present invention as described herein comprises at least 4 genes, preferably at least 6 or 8 genes, and in particular comprises 6 or 8 genes. According to a specific embodiment, the predetermined gene set used in the method of the present invention, whether or not it is streamlined, comprises the following genes: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2, S100A9, TDRD9 and ZAP70. According to a specific embodiment, the predetermined gene set used in the method of the present invention, whether or not it is streamlined, comprises the following genes: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2 and S100A9.

[0047] When the method for determining a transcriptome score (TScore) as described above and herein is performed on a population of assay samples previously extracted from a patient, wherein it is known whether a portion of the samples were extracted from patients who presented with at least one infectious episode since admission to a intensive care unit (particularly admission to an ICU), this method may further comprise, after step b. of the method for determining a transcriptome score (TScore) as described above and herein, and if relevant, after determining the reduced gene set, a step of determining a threshold for each gene of the analyzed (reduced or unreduced) predetermined gene set, in particular by calculating the Youden index for the ROC curve determined in step b. as described herein, more particularly, wherein the threshold is the value that maximizes the Youden index, and optionally providing the threshold for the gene as an output value.

[0048] Such embodiments are applicable to situations where the threshold is defined based on measurements obtained on transcripts of, for example, training samples, rather than by reference to a separately obtainable threshold, which is provided to the method as a predetermined threshold, for example based on data obtained on healthy individuals (e.g., healthy volunteers) who are considered healthy for any medical condition.

[0049] The Youden index, also known in the literature as the "J statistic," is a well-known metric that represents a summary measure of the receiver operating characteristic (ROC) curve for the accuracy of a diagnostic test (see, e.g., Youden WJ. Index for rating diagnostic tests. Cancer. 1950; 3(1): 32-35). It can be defined as follows: Youden index = sensitivity + specificity - 1, or Youden index = [TP / (TP+FN)] + [TN / (FP+TN)] - 1 where TP = true positives, TN = true negatives, FP = false positives, and FN = false negatives. The Youden index ranges between 0 and 1, with a value of 0 indicating that the diagnostic test gives the same proportion of positive results for groups with and without the disease. A value of 1 indicates no false positives or false negatives. The Youden index can be used in conjunction with a receiver operating characteristic (ROC) analysis, where the index is defined for all points of the ROC curve. The maximum value of the index can then be used as a cutoff value for a numerical diagnostic test, as suggested in the embodiments described in detail in the previous paragraphs.

[0050] According to a specific embodiment, wherein the gene expression value is measured by mRNA expression level, preferably by RT-qPCR, in particular according to the protocol described in the experimental part of the present specification, and the genes of the predetermined gene set are selected from the group consisting of: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2, S100A9, TDRD9 and ZAP70, the predetermined threshold value of step b. of the method for determining a transcriptome score (TScore) described above and herein can be determined, for example, to be the following values ​​for each gene, respectively:

[0051] -C3AR1: 0.360

[0052] -CD177:2.867

[0053] -CX3CR1:-1.679

[0054] -IFNgamma: -6.601

[0055] -IL1R2: 3.894

[0056] -S100A9:-0.189

[0057] -TDRD9: 1.923

[0058] -ZAP70:2.482.

[0059] According to a specific embodiment, wherein the gene expression value is measured by mRNA expression level, preferably by RT-qPCR, in particular according to the protocol described in the experimental part of this specification, and the genes of the predetermined gene set are selected from the group consisting of: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2 and S100A9, the predetermined threshold value of step b. of the method for determining a transcriptome score (TScore) described above and herein can be determined as follows for each gene, respectively:

[0060] -C3AR1:0.371

[0061] -CD177:3.199

[0062] -CX3CR1:-1.678

[0063] -IFNgamma: -5.172

[0064] -IL1R2: 3.912

[0065] -S100A9: 0.294.

[0066] According to the convention commonly used in the art, the symbol "-" reflects the fact that the expression of some genes may be reduced relative to a reference pool of samples taken from patients not exhibiting infection, particularly ICU-acquired infection.

[0067] It will be observed that it is within the knowledge of the skilled person to readily adjust the comparison relative to the threshold, taking into account whether the expression of the gene is increased or decreased relative to the sample reference pool as described above. Furthermore, the skilled person can readily incorporate the margin of error associated with the instrument used to measure the mRNA expression level, regardless of the actual measurement method used.

[0068] According to step c. of the method for determining a transcriptome score (TScore) described above and herein, the transcriptome score is calculated as follows: for the assay sample or each patient sample in a population of assay samples, a transcriptome score is calculated, wherein:

[0069] i. for each gene in the predetermined set of genes that are individually analyzed, assigning a point to the gene if the measured gene expression value matches the criteria for classifying the sample as a sample associated with a patient exhibiting an infection, particularly an ICU-acquired infection, and

[0070] ii. summing the scores obtained for all genes in the predetermined gene set to obtain the transcriptome score of the assay sample, and

[0071] iii. Optionally, provide a transcriptome score as an output value.

[0072] “The measured gene expression value matches the criteria for classifying the sample as a sample associated with a patient exhibiting an infection” means that the determined gene expression value is compared with a threshold value that is predetermined and provided or obtained by a receiver operating characteristic (ROC) curve and using the Youden index, including by considering whether the gene expression is increased or decreased in patients with the diagnosis under consideration (exhibiting infection, in particular ICU-AI) relative to patients who are healthy for the diagnosis under consideration.

[0073] It should be understood that the maximum value of the transcriptome score depends on the number of genes used to construct the score. For example, if the expression values ​​of 8 genes are used to construct the transcriptome score, the TScore can range from 0 to 8. For example, if the expression values ​​of 6 genes are used to construct the transcriptome score, the TScore can range from 0 to 6. Guidance can be found in the experimental section.

[0074] For the TScore of the present invention, so-called "relevant validation data" can also be determined. For example, when a group of assay samples previously extracted from a patient is analyzed, wherein a portion of the samples is known to be extracted from patients who have shown at least one infection onset since being admitted to a ward, particularly an ICU, the method of the present invention can also include determining the optimal threshold value associated with the transcriptome score calculated in step c.ii of the method for determining TScore described herein by calculating the Youden index of the transcriptome score, in particular, wherein the optimal threshold value is the value that maximizes the Youden index. According to such an embodiment, such an optimal threshold value (for TScore) is calculated based on the ROC curve of the transcriptome score. The above considerations apply equally to the Youden index. This optimal threshold value or the corresponding Youden index or both can be provided to the method as output values.

[0075] The definition of an optimal threshold for a TScore allows for easier use of such a TScore as a diagnostic tool for the condition under examination, ie, for assessing the risk of adverse clinical outcome in a patient in a clinical setting according to any of the definitions provided herein, as will be seen below.

[0076] The present invention also relates to a method for classifying samples previously extracted from patients, in particular samples previously extracted from patients admitted to an intensive care unit (ICU), into groups (sample groups) reflecting the risk of poor clinical outcomes of patients in a clinical setting in such a setting, said method comprising the steps of:

[0077] a. performing a method for determining a transcriptome score (TScore) according to any embodiment described herein on a sample previously extracted from a patient (to be classified) in order to calculate a transcriptome score (of said sample), and

[0078] b. comparing the transcriptome score calculated in step a. of the present classification method with a predetermined threshold value, in particular with a predetermined threshold value determined as an optimal threshold value by calculating the Youden index of the transcriptome scores obtained for a reference or training population, for example comparing the transcriptome score calculated in step a. of the classification method with an optimal threshold value predetermined using the method defined separately in the above paragraph, and

[0079] c. assigning the sample taken from the patient to a group reflecting the risk of adverse clinical outcomes in a clinical setting based on the comparison in step a. of the present classification method, and

[0080] d. Optionally, provide the classification result obtained in step c. of the present classification method as an output value.

[0081] According to a specific embodiment, it should be noted that the experimental conditions under which the sample to be classified is measured are similar (if not identical) to the experimental conditions used to define the predetermined threshold value (particularly the optimal threshold value) used as a reference in step b. of the classification method. In particular, it should be noted that the same gene or the same measurement method is used in order to make a reliable comparison. For this purpose, the experimental conditions of the reference or training population of the determination sample mentioned in step b. of the classification method can be used.

[0082] For example, as described in the experimental section herein, when the transcriptome score calculated for a patient's assay sample is equal to or greater than 3 (the optimal threshold for TScore is set at 2), the determination of the transcriptome score allows for the determination of the risk of an adverse clinical outcome, wherein the following genes have been assayed: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2, S100A9, TDRD9, and ZAP70. As described in the experimental section herein, this is also the case when the transcriptome score includes the following genes: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2, and S100A9.

[0083] It can be seen that assigning a sample extracted from a patient to a sample group reflecting the risk of an adverse clinical outcome in a clinical setting means classifying the sample into a subgroup of patient samples with a so-called higher risk of an adverse clinical outcome in a clinical setting, or into a subgroup of patient samples with a so-called lower risk of an adverse clinical outcome in a clinical setting, where the definition of "adverse clinical outcome" is as described above and in detail herein. In this article, the use of the adjectives "high" and "low" defines the fact that it is above or below a defined threshold value, which is considered to provide a relevant cutoff line for distinguishing samples, in particular based on the fact that the patients so distinguished may have a specific background of immune dysfunction. The use of the adjectives "high" and "low" relative to the threshold value discussed in the previous sentence also means that the method has a cutoff line, which, taking into account the context in which it is used, is considered to be relevant to the parameter, such as the sensitivity or specificity of the test or method under consideration.

[0084] Depending on the specific embodiment, the so-called "classification" approach may also be referred to as a "hierarchical" approach, which is a term commonly used in the art.

[0085] According to a specific embodiment, the method described herein is computer-implemented, i.e., executed partially or entirely by a computer device. In this case, the possibility of inputting or outputting the data supplied or obtained by the disclosed method is provided, for example, by storing them in the form of variables in a memory or displaying them as parameters on an output visualization system.

[0086] The present invention also relates to a method for identifying patients at risk of adverse clinical outcomes in a clinical setting, comprising subjecting a sample previously extracted from the patient to a classification (or stratification) method as disclosed in any of the embodiments herein, and identifying the patient as having a risk of adverse clinical outcomes in a clinical setting based on the group to which the sample is assigned. The conclusion of risk is based on a conclusion associated with classifying the assay sample into a group that is so-called having a higher or lower risk of adverse clinical outcomes in a clinical setting. According to any disclosed embodiment, the definition of "adverse clinical outcome" is as provided elsewhere in this specification. Therefore, as long as the defined risk applies to the health condition framed by the disease definition, this method qualifies as a diagnostic method. In addition, the method of the present invention is a classification method based on biological parameters.

[0087] According to certain embodiments, the adverse clinical outcome is:

[0088] - Development of complications, such as ICU-acquired infection, particularly intensive care unit-acquired infection (ICU-AI), particularly nosocomial infection, and / or

[0089] - Prolonged stay in a intensive care unit (particularly the ICU), and / or

[0090] -die.

[0091] According to a preferred embodiment, the adverse clinical outcome is the occurrence of a complication, preferably an intensive care unit-acquired infection, particularly an intensive care unit-acquired infection (ICU-AI), particularly a nosocomial infection.

[0092] Thus, the present invention also qualifies as a method for determining the survival rate of a patient in a clinical setting, according to any of the definitions provided for that term in this specification.

[0093] The present invention also relates to a method as disclosed in any of the embodiments herein for:

[0094] a. in vitro or ex vivo diagnosis or prediction of whether a patient is at risk for adverse clinical outcomes in a clinical setting, particularly whether a patient is at risk for acquiring an intensive care unit-acquired infection (particularly a nosocomial infection), and / or

[0095] b. in vitro or ex vivo monitoring of the evolution of the patient's health status over time, particularly when the method is performed on a plurality of different samples taken from the patient at a plurality of different time points, monitoring of the evolution comprises comparing the results obtained at the different sampling times, and / or

[0096] c. Classifying a patient in vitro or ex vivo as belonging to a group showing immune alterations in circulating cells, and / or diagnosing or predicting in vitro or ex vivo whether a patient has a background of immune dysfunction.

[0097] Classifying a patient as belonging to a group that exhibits immune alterations in circulating cells means screening (by in vitro or ex vivo sample determination) the patient's immune status with the goal of making the patient, depending on their immune background, more or less responsive to certain immunomodulatory treatments or drugs. For example, as shown in the experimental section, the inventors found that patients classified as "high risk" for further infection had an immunosuppressive biological profile with low levels of monocyte HLA-DR and high levels of the anti-inflammatory cytokine IL-10.

[0098] The present invention also relates to an in vitro or ex vivo method for screening a drug for its ability to alleviate a patient's adverse clinical outcome as defined in any of the embodiments described herein, the method comprising the steps of:

[0099] - determining whether the patient is at risk for a poor clinical outcome in a clinical setting by performing a method as disclosed in any of the embodiments herein on a sample previously extracted from the patient at a first time point, and

[0100] - if the patient is determined to be at risk for an adverse clinical outcome in a clinical setting, repeating the method disclosed in any embodiment herein (preferably the same method as used at the first time point) on a sample previously extracted from the patient at the second time point later than the first time point, at a second time point subsequent to the first time point, and after treating the patient with a drug believed to have the ability to mitigate the adverse clinical outcome, and

[0101] - Optionally, draw conclusions about the drug's ability to alleviate adverse clinical outcomes.

[0102] Conclusions can be easily drawn by observing any changes that occur after treatment or investigation.

[0103] A "drug that is believed to have the ability to mitigate adverse clinical outcomes" refers, for example, to a drug that would be beneficial for a patient who is classified as a "high risk" patient according to the present invention.

[0104] Throughout this specification, it is observed that if the method described in any embodiment detailed herein is computer-implemented, it can be implemented partially or fully by computer means. Therefore, the present invention also relates to a computer-implemented method comprising the steps of the method according to the present invention according to any embodiment disclosed in this application, in particular, a method or one or more steps being implemented by a computer or computer means or a device commanded by a computer.

[0105] To this end, the present invention further provides a data processing device, comprising:

[0106] a. a device for performing the computer-implemented method according to the invention, in particular a device for retrieving gene expression values ​​and / or a device for inputting reference values, and / or a device for determining a ROC curve and / or a device for determining whether a decision condition is met or calculating a decision condition and / or a device for calculating a transcriptome score, and / or a device for providing an output value, optionally, when the device is executed or driven by a computer, a device for measuring gene expression values ​​of a biological sample previously extracted from a patient, or

[0107] b. comprising a processor adapted / configured to perform the computer-implemented method according to the invention, in particular a processor adapted / configured to perform the steps of the computer-implemented method according to the invention.

[0108] According to a particular embodiment, such a data processing device comprises:

[0109] a. an input interface for receiving the levels of gene expression values ​​defined in step a. of the method for determining TScore as detailed in any embodiment herein, in particular the levels of at least two genes as defined herein,

[0110] b. a memory for storing at least instructions of a computer program, said program comprising instructions for causing a computer to perform the method according to claim 16 when the program is executed by a computer or a processor, optionally a memory for storing data and decision rules,

[0111] c. a processor for accessing a memory to read the aforementioned instructions and executing a computer-implemented method according to the present invention,

[0112] d. Output interface, for providing output values, in particular output values ​​defined in any embodiment described in this specification, and / or for outputting conclusions drawn from any method claims described herein.

[0113] According to another aspect, the present invention also relates to a computer or a computer system configured to perform the operations described in the steps of the computer-implemented method according to the present invention according to any embodiment disclosed in this application.

[0114] The invention also relates to a computer program comprising instructions causing a computer to perform the computer-implemented method according to the invention when said program is executed by a computer or a processor.

[0115] The present invention also relates to a computer-readable medium, in particular a computer-readable non-transitory recording medium, on which a computer program as disclosed in any embodiment of the present invention is stored, in particular when the computer program is executed by a computer or a processor, performs the computer-implemented method according to the present invention. According to another aspect, such a computer-readable medium (which may be a computer-readable non-transitory recording medium) is a computer-readable medium that configures a computer to perform the operations described in the steps of the computer-implemented method according to any embodiment of the present invention disclosed in this application.

[0116] The present invention also relates to the use of a kit comprising means for amplifying and / or detecting gene expression values ​​of at least two different genes, said genes being selected from the group consisting of the genes defined in any embodiment herein, in particular genes selected from the group consisting of: C3AR1, CD177, CD3D, CD74, CIITA, CTLA4, CX3CR1, IFNgamma, IL1R2, TAP2, S100A9, TDRD9 and ZAP70, for performing the method disclosed herein, in particular by measuring gene expression values ​​of at least two different genes selected from the group consisting of the above genes in a biological sample previously taken from a human patient, in particular a human patient who has been admitted to a hospital in a clinical setting, in particular a human patient susceptible to being diagnosed as being at risk for sepsis or a patient susceptible to having a condition that may develop in the event of sepsis.

[0117] According to a specific embodiment, a kit suitable for carrying out the method of the invention as defined in any of the embodiments disclosed herein comprises:

[0118] at least one pair, in particular two pairs, of specific oligonucleotide primers that specifically hybridize to mDNA encoding two or more of C3AR1, CD177, CD3D, CD74, CIITA, CTLA4, CX3CR1, IFNgamma, IL1R2, TAP2, S100A9 TDRD9 and ZAP70, respectively, and optionally one or more of the following reagents,

[0119] - nucleotides (e.g. dATP, dCTP, dGTP, dUTP),

[0120] - reverse transcriptase,

[0121] - a DNA polymerase, in particular a thermostable DNA polymerase, such as Taq DNA polymerase,

[0122] - at least one dye for staining nucleic acids, in particular a dye detectable in a real-time PCT device,

[0123] - optionally, a buffer solution,

[0124] - optionally, reagents required for hybridization of the primers to their targets,

[0125] - optionally, a reference dye, and

[0126] - optionally, a data processing apparatus, in particular as described in detail in any of the embodiments described herein, - optionally, a computer program, in particular as described in detail in any of the embodiments described herein,

[0127] - Optionally, a computer readable medium, in particular as described in detail in any of the embodiments described herein, - Optionally, providing instructions for use and notification of expected values ​​for interpretation of the results.

[0128] As used herein, the term "comprising" is synonymous with "including" or "containing" and is open ended and does not exclude additional, unrecited elements, ingredients, or method steps, whereas the term "consisting of is a closed term that excludes any additional elements, steps, or ingredients not expressly recited.

[0129] The term "consisting essentially of" is a partially open-ended term that does not exclude additional, unrecited elements, steps, or ingredients as long as the additional elements, steps, or ingredients do not materially affect the basic and novel characteristics of the present application.

[0130] The term "comprising" therefore includes the term "consisting of, as well as the term "consisting essentially of. " Therefore, in this application, the term "comprising" is meant to more specifically include the term "consisting of, as well as the term "consisting essentially of.

[0131] To aid readers of this application, this specification is divided into different paragraphs or sections. These divisions should not be considered to separate the content of one paragraph or section from the content of another paragraph or section. Instead, this specification encompasses all contemplated combinations of sections, paragraphs, and sentences.

[0132] All references cited herein are expressly incorporated by reference for their respective relevant disclosures.

[0133] The above and other features of the present invention will become apparent upon reading the examples and accompanying drawings, which illustrate experiments performed by the inventors and supplement the features and definitions given in this specification. The following examples are provided for illustrative purposes. However, these examples are not intended to limit the invention described. BRIEF DESCRIPTION OF THE DRAWINGS

[0134] Figure 1. Characteristics and methods of the discovery cohort

[0135] A: Flowchart of the discovery cohort, B: Schematic diagram describing the method used to identify candidate genes.

[0136] Figure 2. Identification process of genes in the discovery cohort

[0137] A: ROC (receiver operating characteristic) curves of multiple genes (C3AR1: complement C3a receptor 1, CD177: CD177 molecule, CX3CR1: C-X3-C motif chemokine receptor 1, IFNγ: interferon gamma, IL1R2: interleukin 1 receptor 2, S100A9: S100 calcium-binding protein A9, TDRD9: Tudor domain-containing protein 9, ZAP70: T-cell receptor-associated protein kinase 70 zeta chain) with an AUC (area under the curve) greater than 0.7, B: Normalized RNA values ​​of identified genes in patients with (red dots) or without (green) ICU-acquired infection (ICU-AI).

[0138] Figure 3. Clinical Outcomes of High- and Low-Risk Patients in the Discovery Cohort Using TScore

[0139] A: Heat map of unsupervised hierarchical clustering of patients with or without ICU-acquired infection using individual gene scores, B: ROC (receiver operating characteristic) curves with AUC (area under the curve) values, representing the ability of the Tscore obtained on days 5-7 to distinguish whether patients will develop at least one ICU-acquired infection during their ICU stay, C: Proportion of intensive care unit-acquired infection (ICU-AI) among patients with a Tscore between 0 and 2 (low risk) or higher than or equal to 3 (high risk), D: Distribution of types of ICU-AI (UTI: urinary tract infection) among patients with a Tscore between 0 and 2 (low risk) or higher than or equal to 3 (high risk), E: Median and quartiles of intensive care unit length of stay among patients with a Tscore between 0 and 2 (low risk) or higher than or equal to 3 (high risk), F: ICU mortality among patients with a Tscore between 0 and 2 (low risk) or higher than or equal to 3 (high risk).

[0140] Figure 4 Statistical and immunological validity of TScore in the discovery cohort

[0141] Expression levels of multiple immunological parameters (monocyte HLA-DR, IL-10, and immature neutrophil proportion) measured in the blood of patients with a Tscore between 0 and 2 (low risk) or greater than or equal to 3 (high risk)

[0142] Figure 5. Clinical outcomes of high-risk and low-risk patients in the validation cohort using TScore.

[0143] A: Proportion of intensive care unit-acquired infection (ICU-AI) among patients with a Tscore between 0 and 2 (low risk) or higher than or equal to 3 (high risk), B: Distribution of types of ICU-AI (UTI: urinary tract infection) among patients with a Tscore between 0 and 2 (low risk) or higher than or equal to 3 (high risk), C: Median and quartiles of intensive care unit length of stay among patients with a Tscore between 0 and 2 (low risk) or higher than or equal to 3 (high risk), D: intensive care unit mortality among patients with a Tscore between 0 and 2 (low risk) or higher than or equal to 3 (high risk). DETAILED DESCRIPTION

[0144] Example

[0145] method

[0146] Overview: As a training cohort, the inventors used a gene expression dataset obtained from 176 critically ill patients who participated in the REALISM study (NCT02638779) with various etiologies and were still hospitalized in the intensive care unit (ICU) on days 5-7. Based on the independent expression of each gene to identify patients who presented with ICU-acquired infection (ICU-AI) after days 5-7, the inventors established an unweighted score assuming the independence of each gene. The inventors then determined the performance of this score to identify a subgroup of patients at high risk of developing ICU-AI and assessed the ICU length of stay and mortality in this high-risk group. Finally, they validated the validity of the score in a retrospective cohort of 257 patients with sepsis.

[0147] Patients and environment

[0148] In the training cohort, patients and data were collected from the previously published REALISM (REAnimation Low Immune Status Marker) cohort study (7). Briefly, this single-center observational cohort study included critically ill patients with sepsis, trauma, and burns. Inclusion criteria were: patients aged >18 years, clinical diagnosis of sepsis as defined by the SEPSIS-3 consensus guidelines (1), severe trauma with an Injury Severity Score (ISS) >15, or total burn area >30%. Exclusion criteria were any of the following: preexisting conditions or treatments that could affect the patient's immune status, pregnancy, hospitalization, or inability to obtain informed consent. Although data and biosamples were collected on days 1 or 2 (D1-2), D3 or D4 (D3-4), and D5, D6, or D7 (D5-7), we focused on days 5-7. Longitudinal follow-up was performed for 90 days. The IRB (Comité de Protection des Personnes Sud-Est II, Southeastern Commission for the Protection of Human Rights II) approved the study (ref. 2015-42-2) and this study is registered with clinicaltrials.gov (NCT 02638779).

[0149] The validation cohort was derived from the MIPrea (Marqueurs Immunitaires Pronostiques en Réanimation) study (8), which enrolled patients aged >18 years with an expected length of stay >2 days who met SIRS criteria (9) admitted to six French ICUs between December 2009 and June 2011 (CIC IRB#5044 approval). This validation cohort included only patients with sepsis. In this study, we also analyzed biodata from days 5–7.

[0150] Expectation Management

[0151] Patients with sepsis were treated according to the Surviving Sepsis Campaign guidelines (10). Patients received intravenous broad-spectrum antibiotics, with the specific regimen depending on the presumed site of infection, previous antibiotic therapy, and known colonization with antibiotic-resistant bacteria. After identification of the causative pathogen, antimicrobial therapy was deescalated to a narrower spectrum. Source control measures, such as surgery or removal of infected equipment, were performed as necessary.

[0152] After initial stabilization with fluid resuscitation using the “Parkland formula” (11), burn patients underwent a rapid assessment to assess the need for debridement, escharectomy, and fasciotomy to prevent compartment syndrome. In addition to hypnotics, opioids were used for pain control and sedation. Depending on the local condition, patients underwent various surgical procedures for debridement, removal of nonviable tissue, skin grafting, and / or amputation.

[0153] Initial management of severe trauma is aimed at controlling post-traumatic hemorrhage and includes surgical intervention for damage control and medical treatment of coagulopathy, as well as hemodynamic stabilization with vasopressors if needed. If the patient has altered consciousness, intubation and mechanical ventilation are performed to prevent hypoxia. A restrictive transfusion strategy is employed with a target hemoglobin of 70–90 g / L. Pain management is based on the administration of opioids, combined with local anesthesia (if appropriate), or hypnotic medication to sedate the patient. Depending on the site of the trauma, the patient may undergo a variety of surgical procedures (12).

[0154] definition

[0155] The severity of sepsis on admission was assessed using the Simplified Acute Physiology Score 2 and Sequential Organ Failure Assessment (SOFA) scores, while trauma patients were assessed using the ISS score (13,14). ICU-acquired infection was defined as any new episode of probable or definite infection occurring after 48 hours of ICU admission. Only the first episode of ICU-acquired infection was analyzed. ICU-acquired pneumonia was diagnosed according to the American Thoracic Society criteria (15). Patients with clinical suspicion of ventilator-associated pneumonia usually underwent tracheobronchial aspirate and semiquantitative culture. The diagnosis of catheter-related bloodstream infection required the growth of the same pathogen from peripheral blood and catheter tip cultures, or from blood cultures obtained from catheter and venipuncture, with a positive differentiation time of >120 minutes. Urinary tract infection (mostly catheter-related) was diagnosed based on the presence of systemic manifestations of infection and ≥105 Infections were diagnosed by a combination of positive urine bacterial culture and CFU / mL. Infections were reviewed by an independent adjudication committee of three clinicians who were not involved in the recruitment or care of the study patients.

[0156] Immuno-multiplex molecular tools

[0157] Multiplex molecular tools were run as previously reported (9). Briefly, one PAXgene blood RNA tube (PreAnalytix, Hilden, Germany) was sampled at each time point, stabilized at room temperature for at least 2 h after collection, and frozen at −80°C according to the manufacturer's recommendations. mRNA expression was quantified using the BioMérieux 2.0 instrument (bioMérieux), which was used for automated mRNA reverse transcription, amplification, and further quantitative nested PCR of 26 genes (ADGRE3, ARL14EP, BPGM, C3AR1, CCNB1IP1, CD177, CD274, CD3D, CD74, CIITA, CTLA4, CX3CR1, GNLY, IFNgamma, IL10, IL1R2, IL1RN, IL7R, IP10 / CXCL10, MDC1, OAS2, S100A9, TAP2, TDRD9, TNF, and ZAP70).

[0158] As a separate embodiment, the same protocol was used for automated mRNA reverse transcription, amplification, and further quantitative nested PCR of 11 genes (C3AR1, CD177, CD3D, CD74, CIITA, CTLA4, CX3CR1, IFNG, IL1R2, S100A9, and TAP2). Where relevant, the results of this separate embodiment are also reported herein after the results based on the above-mentioned 26 genes. Where this information is not specified, the data set used is the data set obtained starting from the nested PCR of the 26 genes.

[0159] Calculation and statistical analysis of transcriptome scores

[0160] The normalized expression of individual transcripts for each gene assessed by multiple molecular tools was first compared between healthy volunteers and patients with and without ICU-acquired infection. The receiver operating characteristic (ROC) curve for each gene was then calculated to distinguish patients with and without ICU-acquired infection. For genes with an area under the curve (AUC)>0.70, we then used the Youden method to calculate the optimal cutoff value as follows: sensitivity (%)+specificity (%)-100. The cutoff value for each gene was then used to define a threshold, above or below which (depending on its increase or decrease in patients with ICU-acquired infection) each gene was assigned or not assigned a point. Thus, the transcriptome score is the sum of all values ​​for a given patient.

[0161] result

[0162] Training Queue

[0163] In the REALISM cohort, 324 patients had blood samples drawn on days 5-7. Of these, 176 patients did not exhibit ICU-acquired pneumonia before blood sampling on days 5-7 and remained in the ICU at that time point. Most patients in this training set were admitted for trauma (n=72, 41%) and sepsis (n=68, 39%). The overall ICU mortality rate was 8%.

[0164] Overall, 47 patients presented with at least one ICU-acquired infection, while 129 patients did not. The majority of ICU-acquired infections were ICU-acquired pneumonia (n=21, 37%) ( Figure 1A Patients who presented with ICU-acquired infection were more likely to have diabetes mellitus (21% vs. 14%) and chronic lung disease (19% vs. 9%) compared with those without ICU-acquired infection (Table 1). Patients who presented with ICU-acquired infection had a longer median duration of mechanical ventilation (ICU-AI 15 days (95% IQR: 1–29) vs. 1 day (0-3), respectively) compared with those without ICU-acquired infection.

[0165] Table 1: Clinical characteristics of the discovery cohort

[0166]

[0167] Hematological malignancies 0(0) 0(0) <![CDATA[Body mass index, kg / m 2 > 25(22–28) 25(22–29) Parameters at ICU admission SOFA score 6(2–9) 7(5–9) Lactate, mmol / L 0,8(0–2,6) 1,2(0–2,6) Mechanical ventilation 69(53) 37(79) Use of norepinephrine or epinephrine 88(68) 40(85) result Duration of mechanical ventilation, days 1(0–3) 15(1–29) Death in the ICU 7(5) 8(17)

[0168] Using our method ( Figure 1B ), among the 26 genes included in the multiplex molecular tool, 8 transcripts had AUC values ​​greater than 0.70 to classify patients with or without ICU-acquired infection, 8 transcripts namely C3AR1, CD177, CX3CR1, IFNγ, IL1R2, S100A9, TDRD9 and ZAP70 ( Figure 2A With the exception of CX3CR1, IFNγ, and ZAP70, normalized mRNA transcript expression was increased in patients with ICU-acquired infection ( Figure 2B Table 2 summarizes the main diagnostic performance and cutoff value of each of the eight transcripts.

[0169] According to another embodiment, when 11 genes included in the multiplex molecular tool have been used, 6 transcripts were identified with AUC values ​​higher than 0.75 to classify patients with or without ICU-acquired infection, namely C3AR1, CD177, CX3CR1, IFNγ, IL1R2, and (data not shown). Except for CX3CR1 and IFNγ, the normalized mRNA transcript expression was increased in patients with ICU-acquired infection (data not shown).

[0170] Table 2: Table summarizing the diagnostic performance of the identified genes in the discovery cohort.

[0171]

[0172]

[0173] We constructed a transcriptome score (TScore) using the thresholds obtained on the ROC of each transcript using the Youden technique. For each individually analyzed gene, patients with values ​​above the threshold (for C3AR1, CD177, IL1R2, S100A9, and TDRD9) or below the threshold (for CX3CR1, IFNγ, and ZAP70) were assigned one point. In contrast, genes that did not meet these criteria were assigned a value of 0. The TScore varies between 0 and 8 and is the sum of the values ​​of each of the 8 genes ( Figure 3A The AUC value of this TScore on days 5–7 was 0.86 ((0.80–0.92)), which was used to distinguish patients with or without at least one episode of ICU-acquired infection during their ICU stay ( Figure 3B Using the Youden index, the optimal value of the score is 2. Therefore, we defined patients with a TScore higher than 2 as a high-risk group and patients with a TScore lower than or equal to 2 as a low-risk group.

[0174] When applied to the entire population of the training set, patients with a TScore between 3 and 8 exhibited a higher proportion of ICU-acquired infections (n=87, 49%) compared to patients with a TScore less than or equal to 3 (49% vs. 4%, respectively) ( Figure 3C Among these high-risk patients, ICU-acquired infection was more commonly pneumonia (n=19, 49%) compared with low-risk patients (n=2, 11%) ( Figure 3DIn addition, they experienced a longer median ICU length of stay compared to patients in the low-risk TScore group (13 days, 25-75 IQR: 8-30 vs. 7 days, 25-75 IQR: 6-9, p<0.001) ( Figure 3E ). The ICU mortality rate in the high-risk TScore group was higher than that in the low-risk TScore group (15% vs. 2%, p < 0.001) ( Figure 3F ). Table 3 summarizes all parameters.

[0175] We also constructed a transcriptome score TScore for 6 transcripts selected from 11 genes included in the multiplex molecular tool, using the threshold value obtained on the ROC of each transcript using the Youden technique according to another embodiment reported herein. For each individually analyzed gene, patients whose values ​​were above the threshold (for C3AR1, CD177, IL1R2, and S100A9) or below the threshold (for CX3CR1 and IFNγ) were assigned one point. In contrast, genes that did not meet these criteria were assigned a value of 0. The TScore therefore varies between 0 and 6 and is the sum of the values ​​of each of the 6 genes (data not shown). The AUC value of the TScore was 0.84 (0.77–0.89) and was used to distinguish patients with or without at least one ICU-acquired infection on days 5-7 (data not shown). Interestingly, a combination of six transcripts, C3AR1, CD177, IL1R2, S100A9, CX3CR1, and IFNγ, has previously been shown to have an AUC value of approximately 0.79 for assessing the risk of ICU-acquired infection (see WO2022 / 008828, page 156). Therefore, applying the TScore method using the same six transcripts can better distinguish patients with and without at least one episode of ICU-acquired infection.

[0176] Using the Youden index, the optimal value of the score is 2. Therefore, we defined patients with a TScore higher than 2 as a high-risk group and patients with a TScore lower than or equal to 2 as a low-risk group. When applied to the entire population of the training set, patients with a TScore between 3 and 6 showed a higher proportion of ICU-acquired infections (n=82, 46%) compared with patients with a TScore lower than or equal to 2 (59% vs. 9%, respectively) (data not shown). Among these high-risk patients, ICU-acquired infections were more commonly pneumonia (n=20, 49%) compared with low-risk patients (n=1, 11%) (data not shown). In addition, they experienced a longer median ICU stay than patients in the low-risk TScore group (15 days, 25-75 IQR: 9-31 vs. 7 days, 25-75 IQR: 6-9, p<0.001) (data not shown). Thus, in this configuration, the ICU mortality rate was higher in the high-risk TScore group than in the low-risk TScore group (16% vs. 2%, p < 0.001) (data not shown).

[0177] Table 3: Table summarizing clinical characteristics and outcomes of patients with a Tscore between 0 and 2 (low risk) or greater than or equal to 3 (high risk) (SOFA: Sequential Organ Failure Assessment).

[0178]

[0179]

[0180] In the training set, a Wilcoxon-Mann-Whitney test was performed to identify potentially relevant clinical and biological variables for predicting the occurrence of ICU-acquired infection. Of the variables analyzed, only the SOFA score and TScore on day 6 showed statistically significant differences between ICU-acquired and non-ICU-acquired infections. In the multivariable model, the TScore remained independently associated with the occurrence of ICU-acquired infection (Table 4). A type I ANOVA was performed, varying the order of introduction of the explanatory variables in the model, to determine the individual contributions of the day 6 explanatory variables TScore and SOFA to minimizing the bias, as well as their joint contribution. The resulting model explained 25.4% of the total bias, of which 81% was explained by the TScore alone, 16.8% by the day 6 TScore and SOFA together, and 1.7% by the day 6 SOFA alone. Thus, the TScore only slightly overlaps with the information already explained by the main clinical variables in the REALISM database, and given the low proportion of bias explained by these variables in the model, their addition to the TScore appears to be of little relevance.

[0181] Table 4: Univariate analysis of clinical parameters associated between low-risk and high-risk patients in the discovery cohort, followed by multivariate analysis including parameters that were statistically significant in the univariate analysis

[0182]

[0183] Finally, we investigated the expression levels of several biomarkers previously reported to be associated with poor prognosis in ICU patients (16). Interestingly, we found that high-risk patients had reduced levels of HLA-DR expression on monocytes. Similarly, high-risk patients had increased concentrations of the anti-inflammatory cytokine IL10. Only the proportion of immature neutrophils was similar in both groups ( Figure 4 Finally, when we applied our model to the training cohort according to the initial attack (sepsis, burns, trauma, and surgery), we found that TScore had the same ability in identifying groups at high risk for complications (Table 5).

[0184] Table 5: Clinical parameters and outcomes of patients with a Tscore between 0 and 2 (low risk) or greater than or equal to 3 (high risk), subclassified according to their subgroups

[0185]

[0186] Verification Queue

[0187] For the validation cohort, we evaluated the diagnostic performance of the TScore in sepsis patients from the MIPrea cohort. Compared with the REALISM cohort, the MIPrea cohort had a higher proportion of high-risk patients with a TScore between 3 and 8 (62% and 49%, respectively). High-risk patients were the most severely ill at ICU admission, with higher SOFA scores (10, 25th-75th IQR: 7-13 vs. 8, 25th-75th IQR: 5-11) and higher rates of mechanical ventilation (87% vs. 75%) (Table 6). Consequently, the MIPrea cohort had a higher proportion of patients with a TScore between 3 and 6 compared with the REALISM cohort (56% and 46%, respectively). High-risk patients were also the sickest at ICU admission, with higher SOFA scores (11, 25th-75th IQR: 8-13 vs. 8, 25th-75th IQR: 6-10) and higher rates of mechanical ventilation (89% vs. 77%) (data not shown).

[0188] Table 6: Table summarizing clinical characteristics and outcomes of patients with a Tscore between 0 and 2 (low risk) or greater than or equal to 3 (high risk) (SOFA: Sequential Organ Failure Assessment)

[0189]

[0190]

[0191] ICU-acquired infections were more common in high-risk patients (30% vs. 18%, p = 0.06) ( Figure 5A ), while the proportion of urinary tract infection is lower ( Figure 5B The median ICU length of stay was longer in high-risk patients (12 days, 25-75 IQR: 8-20 vs. 9 days, 25-75 IQR: 7-14, p = 0.003) ( Figure 5C ICU mortality was also higher in high-risk patients (20% vs. 13%, p = 0.006) ( Figure 5D ).

[0192] discuss

[0193] Using a transcriptomic score based on the expression levels of six or eight genes, two groups of patients with different risks for adverse outcomes, including ICU-acquired infection, mortality, and ICU length of stay, were identified. Relative to low-risk patients, patients in the high-risk group had an increased risk of ICU-acquired infection, longer ICU length of stay, and higher mortality, even after adjustment for the SOFA score. High-risk assignment on days 5-7 was associated with lower mHLA-DR values ​​and higher circulating IL-10 values.

[0194] Many patients with sepsis now survive the early stages of sepsis due to early identification of sepsis and rapid initiation of antibiotics and organ support therapy. However, they are therefore exposed to a higher risk of secondary infections that are associated with multiple markers of immune dysfunction, such as downregulation of HLA-DR expression on monocytes (17). Despite a sound pathophysiological rationale, several randomized clinical trials over the past decades have failed to show a beneficial effect of immunomodulatory therapies (18). One of the assumptions underlying this failure relies on the significant heterogeneity of patients with sepsis and suggests stratifying patients based on their immune status to identify those who are more likely to benefit from these treatments (19). This stratification requires reliable, relevant, rapid, and decisive monitoring tools.

[0195] To this end, assessment of global differential gene expression in blood has provided encouraging results in the early stages of sepsis (20). However, only a few genes have been reported as relevant biomarkers for predicting ICU-acquired infections (23), regardless of whether most studies have contributed to mortality (21, 22). In fact, it is difficult to identify genes associated with a high risk of ICU-acquired infections, which can be explained by several hypotheses. First, blood cell composition and gene expression change rapidly over time, depending on patient-related and management-related parameters. Second, in most cases, the onset of sepsis remains unknown. Therefore, it seems an interesting approach to evaluate the expression of several genes simultaneously in order to obtain a signal at any time of sampling (9). To this end, we took advantage of multiplex molecular platforms such as bioMérieux The advent of systems that allow for rapid and reliable assessment of the expression of multiple genes has been crucial. We performed our analysis 5–7 days after ICU admission, a timeframe we believe may deviate from abnormalities potentially associated with initial heterogeneity in sepsis, but this analysis could also be performed at another time after ICU admission. Furthermore, transcriptomic results are not affected by the absence of ICU-acquired infection at this time point.

[0196] Over the past few years, machine learning (ML) algorithms have gained increasing attention for classification, unsupervised clustering, or dimensionality reduction tasks on large datasets. ML tools use data-driven algorithms and statistical models to analyze datasets and then make inferences from identified patterns or predictions based on them. However, ML algorithms based on gene expression profiles still perform similarly to conventional severity scores in predicting ICU-acquired infections (24). These poor results may be related to a mismatch between two main assumptions of ML algorithms and the characteristics of the clinical and biological data from which they are derived. First, ML algorithms assume that there is no covariate shift between the training and test sets. Given the important heterogeneity in patient characteristics, this assumption is often violated in the ICU. Second, ML models inherently generate different feature importances across dependent variables in order to fit the data as well as possible. This assumption may be a cause of overfitting of the data, thereby limiting the generalizability of the models.

[0197] Based on these observations, we developed the TScore, which takes into account that 1) each gene is independently associated with ICU-acquired infection and has equal importance in the score without any weighting, and 2) the genes included in the score are independent of each other. This may allow for the capture of different pathophysiological mechanisms reflecting alterations in the immune system. This simplistic approach does not account for possible interactions between genes in signaling pathways that may be involved. However, given the background differences between studies and the heterogeneity of sepsis patients, we believe this approach is suitable for capturing relevant information in heterogeneous patients.

[0198] We found that the TScore was associated with the development of ICU-acquired infection, independent of clinical variables typically associated with it. Furthermore, we identified a biological signature of immunosuppression in high-risk patients, with low monocyte HLA-DR levels and high levels of the anti-inflammatory cytokine IL-10. These data suggest that the TScore can identify patients at risk for ICU-acquired infection based on circulating immune changes. Therefore, our results suggest that the TScore can serve as a stratification tool to identify patients who may benefit from immunotherapy.

[0199] The score was developed using a cohort of critically ill patients admitted to the ICU for a variety of etiologies, including severe infection, burns, trauma, and surgery. Based on previously published results, we hypothesized that immune changes in circulating cells are common across a variety of medical conditions (10). We were then able to validate this approach on a group of patients who presented only with severe infection, demonstrating the robustness of the score.

[0200] This study acknowledges several limitations. First, we performed validation in a historical, independent cohort rather than a prospectively collected cohort. Second, even if these results are replicated in the validation cohort, our results need to be evaluated in a larger cohort and conducted in a prospective manner to confirm the performance of this test. Third, our results are based on whole-blood leukocyte populations, which may vary from patient to patient. Therefore, we cannot rule out that our results reflect differences in leukocyte populations rather than differences in intracellular gene expression. Fourth, multiplex PCR was performed on days 5–7 after ICU admission. This limits the diagnostic performance of the test to the time period in which it was implemented. TScore must be validated and / or adjusted at several time points, which should be possible due to the flexibility and plasticity of the platform.

[0201] in conclusion

[0202] Using a multiplex molecular platform, we established a transcriptomic score based on simultaneous assessment of the expression of six genes on days 5-7 after ICU admission, which identified a subgroup of patients at high risk for developing ICU-acquired infection. The host responses detected in high-risk patients correlated with known biomarkers of immune dysfunction and provide a useful and reliable adjunctive diagnostic tool for further development and evaluation of immunomodulatory drugs in sepsis.

[0203] Transcriptome scoring provides a useful and reliable auxiliary diagnostic tool to further develop immunomodulatory drugs in sepsis in the context of personalized medicine.

[0204] References

[0205] 1.Singer M,Deutschman CS,Seymour CW,Shankar-Hari M,Annane D,Bauer M,et al.The Third International Consensus Definitions for Sepsis and SepticShock(Sepsis-3).JAMA.23févr 2016;315(8):801.

[0206] 2.van Vught LA,Klein Klouwenberg PMC,Spitoni C,Scicluna BP,Wiewel MA,Horn J,et al.Incidence,Risk Factors,and Attributable Mortality of SecondaryInfections in the Intensive Care Unit After Admission forSepsis.JAMA.12avr2016;315(14):1469-79.

[0207] 3.Llitjos JF,Bounab Y,Rousseau C,Dixneuf S,Rimbault B,Chiche JD,etal.Assessing the Functional Heterogeneity of Monocytes in Human SepticShock:a Proof-of-Concept Microfluidic Assay of TNFαSecretion.FrontImmunol.2021;12:686111.

[0208] 4.Llitjos JF,Gassama A,Charpentier J,Lambert J,de Roquetaillade C,Cariou A,et al.Pulmonary infections prime the development of subsequent ICU-acquired pneumonia in septic shock.Ann Intensive Care.15mars 2019;9(1):39.

[0209] 5.Llitjos JF,Bredin S,Lascarrou JB,Soumagne T,Cojocaru M,Leclerc M,etal.Increased susceptibility to intensive care unit-acquired pneumonia insevere COVID-19patients:a multicentre retrospective cohort study.AnnIntensive Care.29janv 2021;11(1):20.

[0210] 6.Tawfik DM,Vachot L,Bocquet A,Venet F,RimmeléT,Monneret G,etal.Immune Profiling Panel:A Proof-of-Concept Study of a New MultiplexMolecular Tool to Assess the Immune Status of Critically Ill Patients.JInfect Dis.21juill2020;222(Supplement_2):S84-95.

[0211] 7.Venet F,Textoris J,Blein S,Rol ML,Bodinier M,Canard B,et al.ImmuneProfiling Demonstrates a Common Immune Signature of Delayed AcquiredImmunodeficiency in Patients With Various Etiologies of Severe Injury.CritCare Med.16sept 2021;

[0212] 8.Friggeri A,Cazalis MA,Pachot A,Cour M,Argaud L,Allaouchiche B,etal.Decreased CX3CR1 messenger RNA expression is an independent molecularbiomarker of early and late mortality in critically ill patients.Crit CareLond Engl.30juin 2016;20(1):204.

[0213] 9.Bone RC,Balk RA,Cerra FB,Dellinger RP,Fein AM,Knaus WA,etal.Definitions for Sepsis and Organ Failure and Guidelines for the UseofInnovative Therapies in Sepsis.Chest.juin 1992;101(6):1644-55.

[0214] 10.Rhodes A,Evans LE,Alhazzani W,Levy MM,Antonelli M,Ferrer R,etal.Surviving Sepsis Campaign:International Guidelines for Management ofSepsis and Septic Shock:2016.Crit Care Med.mars 2017;45(3):486-552.

[0215] 11.Baxter CR.Fluid volume and electrolyte changes of the earlypostburn period.Clin Plast Surg.oct 1974;1(4):693-703.

[0216] 12.Spahn DR,Bouillon B,Cerny V,Duranteau J,Filipescu D,Hunt BJ,etal.The European guideline on management of major bleeding and coagulopathyfollowing trauma:fifth edition.Crit Care.déc 2019;23(1):98.

[0217] 13.Le Gall JR.A New Simplified Acute Physiology Score(SAPS II)Basedon a European / North American Multicenter Study.JAMA J Am Med Assoc.22 déc1993;270(24):2957.

[0218] 14.Vincent JL,de A,Cantraine F,Moreno R,Takala J,Suter PM,etal.Use of the SOFA score to assess the incidence of organ dysfunction / failurein intensive care units:results of a multicenter,prospective study.Workinggroup on 《sepsis-related problems》of the European Society ofIntensive CareMedicine.Crit Care Med.nov 1998;26(11):1793-800.

[0219] 15.Kalil AC,Metersky ML,Klompas M,Muscedere J,Sweeney DA,Palmer LB,etal.Management of Adults With Hospital-acquired and Ventilator-associatedPneumonia:2016 Clinical Practice Guidelines by the Infectious DiseasesSociety of America and the American Thoracic Society.Clin Infect Dis.1sept2016;63(5):e61-111.

[0220] 16.Tremblay JA,Peron F,Kreitmann L,Textoris J,Brengel-Pesce K,Lukaszewicz AC,et al.A stratification strategy to predict secondary infectionin critical illness-induced immune dysfunction:the REALIST score.AnnIntensive Care.17 2022;12(1):76.

[0221] 17.Landelle C,Lepape A,Voirin N,Tognet E,Venet F,BohéJ,et al.Lowmonocyte human leukocyte antigen-DR is independently associated withnosocomial infections after septic shock.Intensive Care Med.nov 2010;36(11):1859-66.

[0222] 18.Cavaillon J,Singer M,Skirecki T.Sepsis therapies:learning from 30years of failure of translational research to propose new leads.EMBO Mol Med[Internet].7avr 2020[cité7 juin 2022];12(4).Disponible sur:https: / / onlinelibrary.wiley.com / doi / 10.15252 / emmm.201810128

[0223] 19.van de Veerdonk FL,Giamarellos-Bourboulis E,Pickkers P,Derde L,Leavis H,van Crevel R,et al.A guide to immunotherapy for COVID-19.NatMed.janv2022;28(1):39-50.

[0224] 20.Davenport EE,Burnham KL,Radhakrishnan J,Humburg P,Hutton P,MillsTC,et al.Genomic landscape of the individual host response and outcomes insepsis:a prospective cohort study.Lancet Respir Med.avr 2016;4(4):259-71.

[0225] 21.Sweeney TE,Perumal TM,Henao R,Nichols M,Howrylak JA,Choi AM,etal.A community approach to mortality prediction in sepsis via geneexpressionanalysis.Nat Commun.déc 2018;9(1):694.

[0226] 22.Scicluna BP,van Vught LA,Zwinderman AH,Wiewel MA,Davenport EE,Burnham KL,et al.Classification of patients with sepsis according to bloodgenomic endotype:a prospective cohort study.Lancet Respir Med.oct 2017;5(10):816-26.

[0227] 23.Peronnet E,Venet F,Maucort-Boulch D,Friggeri A,Cour M,Argaud L,etal.Association between mRNA expression of CD74 and IL10 and risk ofICU-acquired infections:a multicenter cohort study.Intensive Care Med.juill2017;43(7):1013-20.

[0228] 24.Brakenridge SC,Starostik P,Ghita G,Midic U,Darden D,Fenner B,etal.A Transcriptomic Severity Metric That Predicts Clinical Outcomes inCritically Ill Surgical Sepsis Patients.Crit Care Explor.14 oct 2021;3(10):e0554

Claims

1. A method of determining a transcriptome score (TScore) suitable for assessing the risk of adverse clinical outcome for a patient in a clinical setting from an assay sample or a population of assay samples previously extracted from a patient, the method comprising: a. measuring gene expression values ​​of a patient sample or a population of patient samples previously obtained from a patient, or If the method is computer-implemented, retrieving such gene expression values ​​for at least two different genes selected from a predetermined set of genes, and b. for each gene in the predetermined gene set, and for each sample, comparing the measurement value obtained or retrieved in step a. with a predetermined threshold value, said predetermined threshold value identifying the measurement sample as being associated with a patient exhibiting an infection or not exhibiting an infection, in particular an ICU-acquired infection, and / or determining a receiver operating characteristic (ROC) curve based on the measurement values ​​obtained or retrieved in step a., said ROC curve being intended to represent the diagnostic ability of the gene expression values ​​measured for said gene to identify the measurement sample as being associated with a patient exhibiting an infection or not exhibiting an infection, in particular an ICU-acquired infection, and c. for each patient sample in the assay sample or assay sample population, calculating a transcriptome score, in: i. for each gene in the predetermined set of genes that are individually analyzed, assigning a point to said gene if the measured gene expression value matches the criteria for classifying the sample as a sample associated with a patient exhibiting an infection, particularly an ICU-acquired infection, and ii. summing the scores obtained for all genes in the predetermined gene set to obtain the transcriptome score of the assay sample, and iii. Optionally, provide a transcriptome score as an output value.

2. The method according to claim 1, wherein the method is performed on a group of measurement samples previously extracted from patients, wherein for a portion of the samples, it is known whether they were extracted from patients who have exhibited at least one infectious episode since admission to a intensive care unit, particularly an ICU, the method further comprises: a. before step c., further selecting genes having an AUC value of the ROC curve of step b. of at least 0.70 from the predetermined gene set of a., and optionally providing the updated predetermined gene set as an output value, and / or b. Before step c., a step of determining a threshold value for each gene of the predetermined gene set being analyzed, in particular by calculating the Youden index for the ROC curve determined in step b., more particularly, wherein the threshold value is the value that maximizes the Youden index, and optionally providing the threshold value of the gene as an output value.

3. The method according to claim 1 or claim 2, wherein the gene expression value is measured by mRNA expression levels, in particular by a multiplex assay comprising a plurality of assay genes, or by sequencing.

4. The method according to any one of claims 1 to 3, wherein the gene expression values ​​are measured by RT-PCR, in particular RT-qPCR, more in particular nested RT-qPCR, in particular multiplex RT-PCR or RT-qPCR or nested RT-qPCR.

5. The method according to any one of claims 1 to 4, wherein: a. The genes of the predetermined gene set of step a. of claim 1 are selected from the group consisting of ADGRE3, ARL14EP, BPGM, C3AR1, CCNB1IP1, CD177, CD274, CD3D, CD74, CIITA, CTLA4, CX3CR1, GNLY, IFNgamma, IL10, IL1R2, IL1RN, IL7R, IP10 / CXCL10, MDC1, OAS2, S100A9, TAP2, TDRD9, TNF, and ZAP70, and / or b. step a. of claim 1 or step a. of claim 2, wherein the predetermined gene set comprises at least 4 genes, preferably at least 6 or 8 genes, in particular 6 or 8 genes, and / or c. The predetermined gene set of step a. or step a. of claim 2 comprises the following genes: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2, S100A9, TDRD9 and ZAP70 or comprises the following genes: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2 and S100A9.

6. The method according to any one of claims 1 to 5, wherein the biological sample for measurement is extracted from a patient admitted to a resuscitation department, an intensive care unit or a continuous care unit, and / or a patient susceptible to or diagnosed with sepsis or suffering from sepsis, preferably extracted from a patient admitted to an intensive care unit (ICU), more preferably a patient between the 5th and 7th day of hospitalization in the ICU.

7. The method according to any one of claims 1 to 6, wherein the gene expression value is measured by mRNA expression level, and the genes in the predetermined gene set are selected from the group consisting of: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2, S100A9, TDRD9 and ZAP70, wherein the predetermined threshold value of step b. of claim 1 is respectively: -C3AR1: 0.360 -CD177:2.867 -CX3CR1:-1.679 -IFNgamma: -6.601 -IL1R2: 3.894 -S100A9:-0.189 -TDRD9: 1.923 -ZAP70:2.

482.

8. The method according to any one of claims 1 to 7, wherein the assay biological sample is a blood sample, preferably a whole blood sample.

9. The method according to any one of claims 1 to 8, wherein the method is performed on a group of measurement samples previously extracted from a patient, wherein for a portion of the samples, it is known whether they were extracted from patients who have exhibited at least one infectious episode since admission to a intensive care unit, particularly an ICU, the method further comprises: a. determining an optimal threshold associated with the transcriptome score calculated in step c.ii of claim 1 by calculating the Youden index of the transcriptome score, in particular, wherein the optimal threshold is the value that maximizes the Youden index, b. Optionally, providing the optimal threshold or the Youden index or both as output values.

10. A method for classifying samples previously extracted from patients, in particular samples previously extracted from patients admitted to an intensive care unit, into groups reflecting the risk of adverse clinical outcomes in a clinical setting, the method comprising the steps of: a. performing the method for determining a transcriptome score (TScore) according to any one of claims 1 to 9 on a sample previously extracted from a patient to calculate a transcriptome score, and b. comparing the transcriptome score calculated in step a. with a predetermined threshold, in particular with a predetermined threshold value determined as an optimal threshold value calculated by the Youden index of the transcriptome scores obtained for a reference or training population of assay samples, for example comparing the transcriptome score calculated in step a. with an optimal cutoff value predetermined by the method as defined in claim 9, and c. assigning the sample extracted from the patient to a sample group reflecting the risk of adverse clinical outcomes in a clinical setting based on the comparison in step a., and d. Optionally, provide the classification result obtained in step c. as an output value.

11. A method for identifying a patient at risk for a poor clinical outcome in a clinical setting, comprising performing the classification method according to claim 10 on a sample previously drawn from the patient and identifying the patient as being at risk for a poor clinical outcome in the clinical setting based on the group to which the sample was assigned.

12. The method according to any one of claims 1 to 11, wherein the adverse clinical outcome is: - Development of complications, such as ICU-acquired infection, particularly intensive care unit-acquired infection (ICU-AI), particularly nosocomial infection, and / or - Prolonged stay in a intensive care unit (particularly the ICU), and / or -die.

13. The method according to any one of claims 1 to 12, wherein: a. in vitro or ex vivo diagnosis or prediction of whether a patient is at risk for adverse clinical outcomes in a clinical setting, particularly whether a patient is at risk for acquiring an intensive care unit-acquired infection, particularly a nosocomial infection, and / or b. In vitro or ex vivo monitoring of the evolution of the patient's health status over time, particularly when the method is performed on a plurality of different samples taken from the patient at a plurality of different time points, monitoring the evolution comprises comparing the results obtained at the different sampling times, and / or c. in vitro or ex vivo classification of patients into groups showing immune alterations in circulating cells, and / or In vitro or ex vivo diagnosis or prediction of whether a patient has an immune dysfunction background.

14. The method according to any one of claims 1 to 13, wherein when the following genes are assayed: C3AR1, CD177, CX3CR1, IFNgamma, IL1R2, S100A9, TDRD9 and ZAP70, when the transcriptome score calculated for the patient assay sample is equal to or greater than 3, there is a risk of a poor clinical outcome.

15. An in vitro or ex vivo method for screening a drug for its ability to alleviate adverse clinical outcomes in a patient, in particular adverse clinical outcomes as defined in claim 12, comprising the steps of: - determining whether a patient is at risk for a poor clinical outcome in a clinical setting by performing a method according to any one of claims 1 to 11 on a sample previously extracted from the patient at a first time point, and - if the patient is determined to be at risk for an adverse clinical outcome in a clinical setting, repeating the method according to any one of claims 1 to 11 on a sample previously extracted from the patient at a second time point later than the first time point, at a second time point subsequent to the first time point, and after treating the patient with a drug believed to have the ability to mitigate the adverse clinical outcome, and - Optionally, draw conclusions about the drug's ability to alleviate adverse clinical outcomes.

16. The method according to any one of claims 1 to 15, which is computer-implemented.

17. A data processing device comprising: a. A device for performing the method according to claim 16, in particular for retrieving a base means for determining an expression value and / or means for inputting a reference value, and / or means for determining a ROC curve and / or means for determining whether a decision condition is met or a decision condition is calculated and / or means for calculating a transcriptome score, and / or means for providing an output value, Optionally, when the device is executed or driven by a computer, a device for measuring gene expression values ​​of a biological sample previously extracted from a patient, or b. comprising a processor adapted / configured to perform the method according to claim 16, in particular a processor adapted / configured to perform the steps of the method according to claim 16.

18. A computer program comprising instructions which, when said program is executed by a computer or a processor, cause the computer to carry out the method according to claim 16.

19. Use of a kit comprising means for amplifying and / or detecting gene expression values ​​of at least two different genes selected from the group consisting of the genes defined in claim 5, in particular genes selected from the group consisting of: C3AR1, CD177, CD3D, CD74, CIITA, CTLA4, CX3CR1, IFNgamma, IL1R2, TAP2, S100A9, TDRD9 and ZAP70, for performing a method according to any one of claims 1 to 16, in particular by measuring gene expression values ​​of at least two different genes selected from the group consisting of the genes defined in claim 5 in a biological sample previously taken from a human patient, in particular a human patient who has been admitted to a hospital in a clinical setting, in particular a human patient susceptible to being diagnosed as being at risk for sepsis or a patient susceptible to having a condition that may develop in the event of sepsis.

Images