METHOD FOR PREDICTING THE RESPONSE OF A CANCER PATIENT TO IMMUNOTHERAPY TREATMENT

Abstract

The invention relates to an in vitro method for predicting the early progression of cancer in a subject undergoing immunotherapy. This method comprises determining, in a blood sample isolated from the subject, the total concentration of circulating cell-free DNA and the concentration of large circulating cell-free DNA fragments, combining these data in a mathematical function to obtain a score, and comparing this score to a predetermined reference value. It concludes that the subject exhibits early progression during immunotherapy treatment when the score obtained is less than or equal to the reference value.

Description

Title of the invention: METHOD FOR PREDICTING THE RESPONSE OF A CANCER PATIENT TO IMMUNOTHERAPY TREATMENT

[0001] The present invention falls within the general field of oncology.

[0002] More particularly, the present invention relates to an in vitro method for predicting the early progressor and / or non-early progressor status of a subject with cancer during the administration of immunotherapy treatment.

[0003] Cancer treatment has been the focus of much research for decades. In recent years, a particular effort has been made to develop more individualized therapeutic strategies.

[0004] The approach of stimulating the immune system rather than directly attacking the tumor is now recognized as one of the most effective therapeutic strategies for treating a wide range of cancer types. Immunotherapy is used routinely in certain cancers, with spectacular results, particularly in small cell lung cancer. A meta-analysis revealed that 25% of patients treated with immunotherapy experienced a survival-prolonging response. Multivariate analysis also showed that durable responses were more frequent in patients treated with immunotherapy, meaning that their overall survival was improved (Pons-Tostivint et al., 2019, JCO Precision Oncology, 3(3): 1-10).

[0005] However, immunotherapy treatments are not effective in all patients. Furthermore, they can cause adverse effects in some. For these patients, conventional chemotherapy would have been a better option. These patients are referred to as non-responders to immunotherapy treatment. This information is often obtained late, several weeks after the start of immunotherapy treatment, weeks which can lead to a deterioration in the patient's health and significantly reduce their life expectancy. Identifying patients for whom immunotherapy will be effective is now a recognized necessity within the entire immunotherapy community.

[0006] No method currently exists that allows, before the start of immunotherapy treatment, the identification / classification of cancer patients who will respond or not respond to this treatment, more specifically, the identification of patients for whom this treatment will, or will not, prevent progression of the tumor. A reliable, standardized biomarker that could be used in clinical practice to predict treatment response and assess future efficacy has not yet been identified. However, the existence of such a predictive marker would improve the effectiveness of patient care and its underlying ethics, preventing patients from experiencing missed opportunities for recovery and unnecessary side effects. Furthermore, new immunotherapy treatments are sold at very high prices. The cost of CAR-T cell therapy is approximately €350,000 per patient, and that of immunomodulatory therapy is approximately €75,000 per year. These costs obviously pose problems of access and coverage, especially since these treatments are only effective in a fraction of patients and do not always provide a significant benefit.

[0007] To date, two markers are used to predict whether patients will respond to immunotherapy treatments, but only as an indication because the predictive value of these markers is limited: - the combined positive score (CPS), which assesses the expression of PD-L1 (“programmed death ligand-1”) in tumor cells and immune cells - this score is equal to the total number of PD-L1-labeling cells x 100, divided by the number of viable tumor cells; - and the Tumor Proportion Score (TPS), which only takes into account the expression of PD-L1 by tumor cells - this score is equal to the number of tumor cells marking PD-L1) x 100, divided by the number of viable tumor cells.

[0008] These CPS and TPS markers nevertheless require the implementation of protocols often specific to each care center, and no standardization has been carried out to date, which leads to differences in the values ​​of the scores and thresholds applied in the different care centers.

[0009] The present invention aims to provide a method for predicting, with high accuracy, how a cancer patient will respond, more specifically whether their tumor will progress early or not, when they are subjected to immunotherapy treatment, before the start of such treatment.

[0010] Additional objectives of the invention are that this method is easy to implement, using equipment commonly available in analytical laboratories, and that it can be applied to the greatest possible number of, on the one hand, types of cancer, and on the other hand, types of immunotherapy treatments.

[0011] Aiming to develop such a process, the present inventors focused on the circulating free deoxyribonucleic acid (DNA) present in the blood of patients. This circulating free DNA in blood plasma is currently the subject of intense clinical research in oncology, as mutations in genomic DNA can be found there. present in tumor cells, which guide the therapy to be administered to the patient. For convenience, this circulating free deoxyribonucleic acid will be referred to in this description as "circulating cell-free DNA" or cfDNA.

[0012] It is well established that the blood carries a small amount of circulating cell-free DNA resulting from the release of genetic material from tissues. This cell-free circulating DNA is in the form of double-stranded DNA with an average size of 150-180 base pairs (bp), corresponding to the wrapping of DNA within a nucleosome. Its lifespan is less than two hours before it is filtered and eliminated from the bloodstream by the spleen, liver, and kidneys. All studies agree on a lower quantitative detection rate of circulating cell-free DNA in healthy individuals and an increase associated with various clinical situations such as strokes and myocardial infarctions, intense muscular exercise, acute renal failure, hepatic cytolysis, trauma, surgery, cancer, and the presence of a fetus during gestation.

[0013] In the case of cancer, recent advances in molecular biology and DNA sequencing have made it possible to identify numerous mutations in genes involved in oncogenesis from a particular type of circulating DNA, circulating tumor DNA (ctDNA). It has notably been shown that the ctDNA level is correlated with tumor burden and cancer stage. Generally speaking, studies have focused on the use of ctDNA as a tumor biomarker to monitor the patient's positive or negative response to treatment (immunotherapy or other therapies).

[0014] Studies correlating circulating DNA with patient outcomes during immunotherapy treatment have been described in the literature. Most of these studies focus on post-treatment follow-up, monitoring tumor burden either by measuring ctDNA using standard techniques such as polymerase chain reaction (PCR) or sequencing, or by direct examination of the tumor using imaging.

[0015] From a more general perspective, studies on circulating DNA focus on small DNA fragments, less than 250 base pairs, which are considered to carry the main information, and more specifically on sizes between 70 and 150 base pairs, which are enriched in ctDNA. ctDNA is thus often proposed as an indicator of disease progression.

[0016] It has now been discovered by the present inventors that the analysis of the size profile of the total circulating cell-free DNA of subjects with cancer makes it possible to predict their future response to immunotherapy treatments, more specifically to determine whether or not the subject will be an early progressor when such treatment is administered to him, prior to this administration.

[0017] In the present description, the term "non-early progressor" or "non-early progressor" is conventionally used in the medical field to refer to a cancer patient for whom the first image of the tumor acquired by medical imaging, particularly by X-ray CT scan, after the start of immunotherapy treatment, compared to the image acquired by the same medical imaging technique before the start of this treatment, shows an increase of less than 20% in the tumor lesions detected by imaging; that is, the tumor lesions have decreased, remained stable, or increased by less than 20% (in surface area). This is more commonly referred to as progression of less than 20% of the cancerous tumor at the first imaging assessment after the start of immunotherapy treatment.

[0018] The timing of the first medical imaging acquisition after the start of immunotherapy treatment is determined according to the conventional protocol for managing cancer patients in use in the field, and depends on the specific type of cancer concerned. Typically, this time is 6 weeks after the start of immunotherapy treatment for lung cancer, and 12 weeks for other cancers.

[0019] In contrast, an early-progressing subject, or a subject with early progression characteristics, is understood, also conventionally in the medical field, to be a cancer patient for whom the first image of the tumor acquired by medical imaging, in particular by X-ray CT scan, after the start of immunotherapy treatment, compared to the image acquired by the same medical imaging technique before the start of this treatment, shows an increase of more than 20% in the tumor lesions detected by imaging; that is to say, the tumor lesions have increased by 20% or more (in surface area). This is more commonly referred to as progression of more than 20% of the cancerous tumor at the first imaging evaluation after the start of immunotherapy treatment.

[0020] As indicated above, circulating cell-free DNA also refers to extracellular DNA present in the patient's blood plasma. This circulating cell-free DNA includes circulating tumor cell-free DNA and circulating non-tumor cell-free DNA.

[0021] More specifically, the present inventors have discovered that patients with a non-early progressor profile, in whom the tumor does not progress, progresses very little, or even regresses, during the initial period of immunotherapy treatment, in the first weeks of this treatment, have a higher concentration of large circulating cell-free DNA fragments than patients with an early progressor profile. Thus, the concentration, particularly the relative concentration, of circulating cell-free DNA fragments in a cancer patient with a size greater than 500 base pairs constitutes a parameter that can be used in a method for detecting to determine whether this patient will be an early or non-early progressor upon administration of immunotherapy treatment. Nothing in the prior art, which only focuses on small circulating cell-free DNA fragments, well below 500 base pairs, suggested such a result.

[0022] Thus, the present invention proposes an in vitro method for predicting whether a cancer patient will be an early progressor and / or a non-early progressor during immunotherapy treatment, that is, when immunotherapy treatment is administered. This method falls under the category of statistical analysis methods. It comprises the following steps: a / determination, in an isolated blood sample obtained from said subject, preferably before said immunotherapy treatment: - the total concentration of circulating free deoxyribonucleic acid in said blood sample, referred to as the total cfDNA concentration, - at least one concentration, in said blood sample, of circulating free deoxyribonucleic acid fragments whose size is included within a predetermined size range, said predetermined size range, which will be designated in this description by the expression "analytical size range", being included in sizes greater than 500 base pairs, the concentration thus determined being called in this description "large cfDNA concentration", b / combination in a mathematical function of at least said large cfDNA concentration, and said total cfDNA concentration, so as to obtain a score, this score reflecting the probability that the subject exhibits or does not exhibit an early progressor trait and / or an early non-progressor trait during immunotherapy treatment, And : cl / comparison of this score to a first predetermined reference value, and dl / conclusion that the subject presents an early progressor characteristic during immunotherapy treatment when this score is less than or equal to this first reference value, and / or c2 / comparison of this score to a second predetermined reference value, and d2 / conclusion that the subject presents an early non-progressor characteristic during immunotherapy treatment when the score is greater than this second reference value.

[0023] In particular embodiments of the invention, the method is an in vitro method for predicting the early progression of a cancer subject during immunotherapy treatment, which comprises the steps of: a / determination, in an isolated blood sample obtained from said subject, preferably before said immunotherapy treatment: - the total concentration of circulating free deoxyribonucleic acid in said blood sample, referred to as the total cfDNA concentration, - at least one concentration, in said blood sample, of circulating cell-free deoxyribonucleic acid fragments whose size is included within a predetermined size range, said predetermined size range, which will be designated in this description by the expression "analytical size range", being within sizes greater than 500 base pairs, the concentration thus determined being called in this description "large cfDNA concentration", b / combination in a mathematical function of at least said large cfDNA concentration, and said total cfDNA concentration, so as to obtain a score, this score reflecting the probability that the subject exhibits or does not exhibit an early progressor characteristic during immunotherapy treatment, cl / comparison of this score to a first predetermined reference value,and / conclusion that the subject presents an early progressor profile during immunotherapy treatment when this score is less than or equal to this first reference value.

[0024] In particular embodiments of the invention, the method is an in vitro method for predicting the non-progressor status of a cancer subject during immunotherapy treatment, which comprises the steps of: a / determination, in an isolated blood sample obtained from said subject, preferably before said immunotherapy treatment: - the total concentration of circulating free deoxyribonucleic acid in said blood sample, referred to as the total cfDNA concentration, - at least one concentration, in said blood sample, of circulating free deoxyribonucleic acid fragments whose size is included within a predetermined size range, said predetermined size range, which will be designated in this description by the expression "analytical size range", being included in sizes greater than 500 base pairs, the concentration thus determined being called in this description "large cfDNA concentration", b / combination in a mathematical function of at least said large cfDNA concentration, and said total cfDNA concentration, so as to obtain a score, this score reflecting the probability that the subject exhibits or does not exhibit an early non-progressor trait during immunotherapy treatment, c2 / comparison of this score to a second predetermined reference value, and d2 / conclusion that the subject exhibits an early non-progressor characteristic during immunotherapy treatment when the score is greater than this second value of reference.

[0025] The method according to the invention can otherwise be an in vitro method for predicting whether a subject with cancer is an early progressor or a non-early progressor during immunotherapy treatment. It then comprises, after the scoring step: - a step cl / of comparing this score to the first predetermined reference value, and a step dl / of concluding that the subject presents an early progressor characteristic during immunotherapy treatment when this score is less than or equal to this first reference value, And - a step c2 / of comparing this score to the second predetermined reference value, and a step d2 / of concluding that the subject presents a non-progressor early characteristic during immunotherapy treatment when the score is greater than this second reference value.

[0026] As stated above, the method according to the invention, which takes into account the content of large cfDNA in the blood sample taken from the subject, makes it possible to predict, with high accuracy, whether a cancer patient will be a non-early progressor and / or an early progressor with respect to immunotherapy treatment, even before such treatment has begun. This method is particularly advantageous because it allows, for example, avoiding the administration of immunotherapy treatment to patients who have been predicted to be early progressors and / or who have not been predicted to be non-early progressors, for whom this treatment will not be effective.

[0027] The accuracy of prediction of the method according to the invention was verified in particular in a prospective observational study conducted on a cohort of 51 patients suffering from different types of cancer, and having been subjected to different types of immunotherapy treatments (the method according to the invention having been implemented on blood samples taken from the patients before the start of these treatments).

[0028] In a classical manner, the term precision here refers to the proportion of patients who have been correctly classified by the process according to the invention.

[0029] The performance of statistical prediction methods is generally evaluated by plotting a receptor function characteristic curve (ROC curve) and measuring the area under the curve (AUC). The ROC curve is established by plotting sensitivity versus (1-specificity) after classifying patients, based on the results obtained by the prediction method. The closer the AUC value is to 1, the higher the sensitivity and specificity of the method, and the better the performance of the method. In the aforementioned patient cohort study, a method using a method of implementation The work of the invention made it possible to predict an early progressor character of patients with an AUC as high as 0.833, which clearly demonstrates its good performance.

[0030] “Sensitivity” here means, in a classical way, the probability of the process identifying "Specificity" refers to the probability that the process will not identify as positive subjects those who are early progressors (or, as the case may be, those who are not early progressors), that is, to identify true positives.

[0031] Furthermore, in this description, "positive predictive value" (PPV) means the probability that a subject will be an early progressor (or, as the case may be, a non-early progressor) when the outcome of the process predicts that subject will be as such. "Negative predictive value" (NPV) means the probability that a subject will be a non-early progressor (or, as the case may be, an early progressor) when the outcome of the process predicts that subject will be as such.

[0032] Each of the reference values ​​used in the method according to the invention is preferably a threshold value, also commonly referred to as a "cut-off value." It is within the competence of a person skilled in the art to establish such reference values, particularly such threshold values, for the prediction method according to the invention, depending on the level of sensitivity and specificity required for each prediction. Such a determination can be made either empirically or theoretically.

[0033] In particular embodiments of the invention, said first reference value and said second reference value are identical, and are designated by the expression "common reference value". Then, the method includes a conclusion step such that: - if the score is less than or equal to the common reference value, the subject presents as an early progressor during immunotherapy treatment, and / or - if the score is higher than the common reference value, the subject presents an early non-progressor characteristic during immunotherapy treatment.

[0034] In alternative embodiments of the invention, the first reference value and the second reference value are different, the first reference value being lower than the second reference value. The method may then include a conclusion step such that: - if the score is less than or equal to the first reference value, the subject presents as an early progressor during immunotherapy treatment. - if the score is higher than the second reference value, the subject presents a early non-progressor characteristic during immunotherapy treatment, - if the score is greater than the first reference value and less than or equal to the second reference value, the method does not allow for a satisfactory prediction of the subject's response to immunotherapy treatment.

[0035] The prediction method according to the invention, the results of which are based on statistical analysis, advantageously allows, in particular: - a prediction-stratification before therapy of patients who are not early progressors and / or patients who are early progressors during immunotherapy treatment, thus allowing for personalized care; - for the patient, a saving of time increasing the chances of success of their therapy, by avoiding side or induced effects; - simplicity and speed of implementation by analysis carried out from a blood sample already taken from the patient for other purposes; - a very significant reduction in the costs of immunotherapy treatments, which can be targeted only at patients predicted as non-early progressors by the method according to the invention.

[0036] The blood sample used in the method according to the invention was preferably isolated from the subject by a blood sample taken before the start of immunotherapy treatment. The invention does not, however, preclude this blood sample from being taken at the time of administration of this treatment, or afterwards, preferably immediately afterwards, within a few hours or a few days following the start of treatment administration.

[0037] The method according to the invention does not in itself involve any step applied to the patient's body. The analytical steps it involves are carried out using whole blood samples that have been previously taken from the patient, in a conventional manner.

[0038] The subject to which the method according to the invention is applied is preferably a mammal. It is preferably a human.

[0039] The method according to the invention may also meet one or more of the characteristics described below, implemented in isolation or in each of their technically operative combinations.

[0040] The mathematical function implemented in the process according to the invention can be any type of multivariate function.

[0041] In addition to the variables of large cfDNA concentration and total cfDNA concentration, it can use other variables. In particular, it can use, as variables, several different large cfDNA concentrations, each corresponding to a predetermined analytical size range different from the others. It should be noted that, for the purposes of the present invention, the expression "Analytical size range" means any size range within sizes greater than 500 base pairs. It may also, or otherwise, use other clinical or biological information about the patient as variables.

[0042] Each of the variables contained in the mathematical function according to the invention can be weighted by a coefficient which is specific to it and which can be equal to 1, or different from 1.

[0043] In particular embodiments of the invention, the mathematical function is the ratio between at least said concentration of large cfDNA, and said total cfDNA concentration.

[0044] This refers in particular to the ratio between a large cfDNA concentration and the total cfDNA concentration. This function can then be described as the relative concentration of cfDNA fragments whose size falls within the predetermined analytical size range, present in the blood sample of the subject under study. As indicated above, the present inventors have discovered that such a relative concentration constitutes an indicator / biomarker that effectively predicts the response (in terms of early or late tumor progression) of a cancer patient undergoing immunotherapy treatment.

[0045] The mathematical function implemented according to the invention can otherwise be equal to the sum of different concentrations of large cfDNA, each being associated with a predetermined analytical size range, this sum being divided by the total cfDNA concentration.

[0046] It can otherwise combine the above variables in any other way, and in particular be part of linear regression.

[0047] In particular embodiments of the invention, each predetermined size range, or analytical size range, is defined solely by its minimum value, that is to say, it is a range defined as including sizes greater than, or greater than or equal to, a given minimum value.

[0048] Thus, in particular embodiments of the invention, at least one predetermined size range / analytical size range is the range of sizes greater than 500 base pairs, preferably the range of sizes greater than or equal to 580 base pairs, more preferably the range of sizes greater than or equal to 600 base pairs, and even more preferably the range of sizes greater than or equal to 1500 base pairs. Preferably, it is the range of sizes greater than or equal to 1600 base pairs, or even greater than or equal to 1650 base pairs, or even greater than or equal to 1700 base pairs.

[0049] In particular alternative embodiments of the invention, each predetermined size range, or analytical size range, is defined by its value minimum and its maximum value, that is to say, it is a range defined as including sizes greater than, or greater than or equal to, a given minimum value, and less than, or less than or equal to, a given maximum value. In such a configuration, larger and / or wider size ranges are particularly preferred within the scope of the invention, compared to smaller and / or narrower size ranges.

[0050] In particular embodiments of the invention, at least one predetermined size range / analytical size range is the size range between 580 and 1649 base pairs.

[0051] The mathematical function used according to the invention can combine all combinations of the variables associated each with a range of analytical size which is its own, which are mentioned above.

[0052] Among the predetermined size ranges particularly preferred within the framework of the invention, in the context in which the mathematical function implemented is the ratio between a concentration of large cfDNA (i.e. cfDNA fragments in the predetermined size range) and the total cfDNA concentration, examples of analytical size ranges include the size range greater than or equal to 1650 base pairs, and the size range from 580 to 1649 base pairs.

[0053] The score obtained by the method according to the invention advantageously constitutes a generic indicator of the early progressor and / or non-early progressor status of a subject studied, independent of both the type of cancer from which the subject has, and the type of immunotherapy treatment that could be administered to him.

[0054] In particular, the method according to the invention advantageously makes it possible to predict the early progressor and / or non-early progressor character of a subject with any type of cancer, and in particular, but not limited to, metastatic tumor pathology, melanoma, kidney cancer, in particular clear cell carcinoma, urothelial carcinoma of the bladder, squamous cell carcinoma of the head and neck, or small cell or non-small cell lung cancer, or a plurality of such cancers.

[0055] Immunotherapy treatment aimed at determining whether the subject is an early progressor and / or a non-early progressor can be of any type. It may, in particular, involve nivolumab, ripilimumab, pembrolizumab, and / or one or more anti-PD-L1 antibodies such as atezolizumab, avelumab, and / or durvalumab.

[0056] The prediction by the method according to the invention is also effective with regard to configurations in which immunotherapy treatment is carried out in conjunction with other therapeutic treatment methods, such as a chemotherapy treatment and / or targeted therapy when possible.

[0057] Obtaining the subject's blood sample, from which the process according to the invention is applied, may have been carried out in any conventional manner. For example, it may have been carried out by collecting a blood sample from the subject in tubes containing a buffer based on ethylenediaminetetraacetic acid (EDTA) or in specific tubes for obtaining circulating DNA, such as the tubes marketed by Streck under the name Cell-Free DNA BCT® (Streck) or by Roche Diagnostic under the name "Cell-free DNA collection tube", according to the supplier's recommendations.

[0058] In particular embodiments of the invention, the determination of said total cfDNA concentration and the determination of said at least one large cfDNA concentration are carried out by direct analysis of the isolated blood sample from the subject, i.e. a whole blood sample.

[0059] In preferred embodiments of the invention, the process includes a preliminary step of obtaining a plasma sample from this whole blood sample.

[0060] The step of selectively obtaining the plasma sample from the whole blood sample isolated from the patient by blood collection is preferably carried out within a few days of the blood collection, for example within 6 hours after the blood collection if EDTA buffer tubes are used, and within 7 days after the blood collection if specific tubes for obtaining circulating cell-free DNA are used, such as the tubes mentioned above.

[0061] This step can be carried out in any conventional way in itself, for example by centrifugation.

[0062] In order to limit the release of cellular DNA contained in circulating cells present in the whole blood sample, which would cause dilution of the circulating cell-free DNA, a protocol including two centrifugations is preferably preferred according to the invention. Preferably, the step of obtaining the plasma sample from a whole blood sample obtained from the patient, according to the process of the invention, comprises the following: - a first gentle centrifugation, for example at a speed between 1200 and 1600 g, preferably at room temperature, i.e. at 20°C + / - 5°C; - the recovery of the plasma (supernatant), without removing the layer of cells separating the plasma and the red blood cells; - a second centrifugation at a faster speed, for example between 3000 and 16000 g, preferably also at room temperature; - and the recovery of the plasma (supernatant), for example by aspiration, without collecting the pellet formed.

[0063] Regardless of the aforementioned tubes, the plasma sample thus obtained can typically be stored at -20°C for a period less than or equal to 1 month, or at -80°C for periods greater than 1 month, before its analysis for the determination of the different concentrations of cfDNA necessary for the implementation of the process according to the invention.

[0064] The determination of the total cfDNA concentration and the determination of the concentration(s) of large cfDNA in the blood sample can be carried out by any method that is conventional in itself for those skilled in the art. Depending on the particular technique used, they can be performed directly on the whole blood sample or on the plasma sample obtained from this blood sample, or after a step of extracting cfDNA from one or the other of these samples.

[0065] Thus, in particular embodiments of the invention, which are particularly advantageous in terms of the speed and simplicity of implementation of the process according to the invention, the determination of said total cfDNA concentration and the determination of said at least one large cfDNA concentration are carried out by direct analysis of the plasma sample obtained from the whole blood sample isolated from the subject, that is to say, they are carried out directly on this sample itself, without a prior DNA extraction step.

[0066] In particular alternative embodiments of the invention, the process includes a step of extracting circulating free deoxyribonucleic acid from the blood sample, or where applicable, from the plasma sample obtained from that blood sample, so as to obtain an extract containing the circulating free deoxyribonucleic acid contained therein. The determination of the total cfDNA concentration and the determination of said at least one concentration of large cfDNA are then carried out by analyzing the extract of circulating free deoxyribonucleic acid thus obtained.

[0067] The step of extracting cfDNA from the blood sample, or where applicable from the plasma sample, can be carried out in any conventional way in itself, for example by a technology using magnetic beads, or by means of a kit marketed for this purpose, by the companies IDSolutions (IDXtract Kit), Promega (Maxwell® RSC ccfDNA plasma Kit) or Qiagen (QIAamp Circulating Nucleic Acid Kit) for example.

[0068] The analysis of the blood sample, plasma sample, or cfDNA extract obtained according to the invention, to determine the total cfDNA concentration and one or more large cfDNA concentration(s) (each associated with a different analytical size range) can be carried out by any technique known to those skilled in the art for this purpose.

[0069] It can, for example, be performed by polymerase chain reaction (PCR), or by capillary electrophoresis technology, using a commercially available device, such as the Fragment Analyzer and System 7100 CE from Agilent Technologies, the QIAxcel offered by Qiagen, or the GenomeLab system from Sciex. These techniques require prior extraction of cell-free circulating DNA from the blood or plasma sample.

[0070] Preferably, the determination of the different concentrations of cfDNA is carried out by a technique based on the so-called pLas technology, as described in particular in documents WO 2016 / 016470, WO 2014 / 020271, or the publication by Ranchon et al., Lab Chip, 2016, 16(7): 1243-53. Schematically, the pLas technology operates in two steps, respectively concentration and separation, performed online. First, the DNA is concentrated via a capillary system formed by the junction of a small capillary and another capillary of larger cross-section. The solution containing the DNA is made to flow laminarly into the large capillary, and an electric field is used to slow the migration. Because of the shearing caused by the laminar flow, this counter-electrophoresis creates a transverse force, dependent on the size of the DNA, which pushes the DNA towards the walls.The change in flow velocity and electric field at the constriction allows the DNA to be stopped and concentrated into a disc. Upstream of the constriction, the flow and counter-electrophoresis are slow, resulting in a weak transverse force. The DNA is therefore within the bulk of the flow and moves towards the constriction. Conversely, downstream of the constriction, the flow and counter-electrophoresis are rapid, strongly pressing the DNA against the wall where the laminar flow is very weak and counter-electrophoresis dominates. The DNA then moves back towards the constriction, by counter-electrophoresis, along the wall. This disc is then released by the gradual decrease in the electric field, which also allows the separation operation to be carried out according to the size of the fragments.

[0071] The analysis of the blood sample, plasma sample, or cfDNA extract obtained according to the invention, to determine the total cfDNA concentration and one or more concentrations of large cfDNA (each associated with a different analytical size range), can in particular be carried out by the method, based on the pLas technology described above, as described in document FR 3128231 or in the publication by Boutonnet et al., Analytical Chemistry, 2023, 95(24): 9263-70. This method has the advantage of being able to be implemented directly on a plasma sample, without prior extraction of the circulating DNA contained therein, and it advantageously allows for the easy and economical determination of the cfDNA size profile characterization with a sen increased sensitivity compared to other analytical techniques. This method can notably be implemented using the device described in documents WO 2017 / 009566 and Andriamanampisoa et al., Analytical Chemistry, 2018, 90(6): 3766-74, and marketed under the name Biabooster by Adelis Technologies.

[0072] In essence, this method involves at least one iteration of an alternation: - a laminar flow step of the sample to be analyzed in a capillary in a first direction of flow, the capillary being provided with at least one local restriction of its cross-section and comprising an analysis buffer, during which the sample is subjected to a first electrical potential difference whose action on the nucleic acid molecules is opposite to the first direction of flow and causes the retention of nucleic acid molecules in the capillary, - a laminar flow stage of the sample, in a second flow direction opposite to the first flow direction, and, after the last iteration of the alternation, a step of separation by laminar flow of the sample in the capillary in the first direction of flow, during which the sample is subjected to an electrical potential difference less than or equal to the first potential difference, the action of which on the nucleic acid molecules is opposite to the first direction of flow and causes partial retention of nucleic acid molecules in the capillary. This method further includes, during or after the separation step, a step for measuring a temporal fluorescence profile of fluorescent nucleic acid molecules and a step for converting the temporal fluorescence profile into a concentration profile of nucleic acid molecules of different lengths. This conversion is achieved by applying a fluorescence profile to a standard sample, the concentration of which is known for each length of fluorescent nucleic acid molecule present in the sample. The nucleic acid molecules were made fluorescent by a technique known to those skilled in the art, for example, by adding an intercalating fluorophore to the analysis buffer—that is, a molecule that fluoresces weakly in its free state but fluoresces strongly when intercalated between the DNA bases. The guaranteed technical uncertainty of such a method, based on pLas technology, is advantageously 5 to 10%. The repeatability in size is better than 1%. In particular, this method advantageously allows the determination of the total cfDNA concentration and the concentration(s) of large cfDNA in the plasma sample with a very high degree of accuracy.

[0073] It has further been discovered by the present inventors that the score obtained in the prediction method according to the invention also constitutes a good statistical indicator of the overall progression-free survival prognosis of the subject with cancer treated with immunotherapy, as well as his 3-year survival.

[0074] The features and advantages of the invention will become more apparent in the light of the following implementation examples, provided by way of illustration only and in no way limiting the invention, with the support of Figures 1 to 8, in which:

[0075] [Fig-1] Fig.1 shows a curve representing the fluorescence intensity profile rescence as a function of time, obtained by analysis, by a method based on pLas technology, of a plasma sample from a cancer patient containing cfDNA.

[0076] [Fig.2] Fig.2 shows a curve representing the fluorescence intensity profile rescence as a function of time, obtained by analysis, using a method based on pLas technology, of a standard sample containing DNA fragments of known size and concentration.

[0077] [Fig.3] Figure [Fig.3] shows a curve representing the DNA concentration profile free circulating, expressed in picograms per pl per base pair (pb), as a function of size, obtained from the curve of [Fig.1].

[0078] [Fig.4] Fig.4 shows a histogram representing the distribution of the circulating free DNA concentration for different size ranges, obtained from the curve in [Fig.3].

[0079] [Fig. 5] Fig. 5 represents ROC curves relating to the prediction of the characteristic of early progressor of a cancer patient during immunotherapy treatment, obtained for an indicator according to the invention, the relative concentration of cfDNA fragments of size greater than or equal to 1650 bp in a plasma sample of said patient, in a / for a training lot, in b / for a test lot, and in c / for the total cohort; for each curve, the area under the curve (“AUC”) is shown in the figure.

[0080] [Fig. 6] [Fig. 6] represents ROC curves relating to the prediction of the characteristic of early progressor of a cancer patient during immunotherapy treatment, obtained for an indicator according to the invention, the relative concentration of cfDNA fragments of size between 580 and 1649 bp in a plasma sample of said patient, in a / for a training lot, in b / for a test lot, and in c / for the total cohort; for each curve, the area under the curve (“AUC”) is shown in the figure.

[0081] [Fig.7] Fig.7 represents the box plots obtained for a cohort of 51 patients for two indicators according to the invention, in a / the relative concentration of cfDNA fragments of size greater than or equal to 1650 bp in a plasma sample of said patient, in b / the relative concentration of cfDNA fragments of size between 580 and 1649 bp in this plasma sample.

[0082] [Fig.8] Figure [Fig.8] shows a graph representing, as a function of time, the curve of progression-free survival of cancer patients who underwent immunotherapy treatment, determined by the Kaplan-Meier method, according to whether they have a score above (light curve) or below (dark curve) the threshold value of 0.034, this score being predicted by a method according to the invention taking into account the relative concentration of cfDNA fragments of size greater than or equal to 1650 bp in the plasma samples of the patients; the number of patients ("individuals at risk") from the initial cohort included in the study at each time of evaluation, respectively with a score less than 0.034 (top line) and a score greater than 0.034 (bottom line) is indicated in the table below the graph.

[0083] This study is based on a cohort of 51 patients with 5 types of cancerous pathologies: melanoma (20%), ENT (squamous cell carcinoma of the head and neck) (20%), lung (non-small cell) (21%), kidney (18%), bladder (21%).

[0084] The response to immunotherapy treatment of patients refers to the response to the treatments used which are: nivolumab, ipilimumab and / or pembrolizumab, these treatments being able to be combined with targeted therapy, administration of anti-PD-L1 antibodies or chemotherapy treatment.

[0085] The characteristics of the patients in the cohort at the beginning of the study are summarized in Table 1.

[0086] [Tables 1] Characteristics All Bladder ENT Kidney Melanoma Lung Number of patients 51 11 10 9 10 11 Mean age (years) 65 68 60 64 66 66 Male 39 8 8 7 7 9 Female 12 3 2 2 3 2 Treatment performed Nivolumab 13 0 9 0 1 3 Pembrolizumab 13 9 0 2 0 2 Nivolumab + ipilimumab 10 0 0 1 9 0 Nivolumab + targeted therapy 2 1 1 0 0 0 Pembrolizumab + targeted therapy 6 1 0 5 0 0 Pembrolizumab + anti-PD-L1 1 0 0 1 0 0 Pembrolizumab + chemotherapy 5 0 0 0 0 5 Pembrolizumab + chemotherapy + targeted therapy 1 0 0 0 0 1 Number of metastases: 0 12 1 6 0 2 3 1 12 4 2 4 1 1 2 16 4 1 4 5 2 3 6 1 1 1 1 2 4 3 1 0 0 0 2 6 1 0 0 0 0 1 7 1 0 0 0 1 0

[0087] Table 1 - Characteristics of patients in the cohort

[0088] The evaluation of patient progression or lack thereof during immunotherapy treatment is performed by analyzing tumor images acquired by CT scan using the standard iRECIST methodology (“iRECIST: guidelines for Response criteria for use in trials testing immunotherapeutics,” Seymour et al., LANCET Oncology, vol 18, issue 3, el43-el52, 2017), 6 to 12 weeks after the start of treatment, depending on the type of cancer (6 weeks for lung cancer, 12 weeks for other cancers). The clinical data monitored are: serum lactate dehydrogenase (LDH) level, level of metastases, age, sex, smoker or non-smoker, tumor proportion score (TPS).

[0089] 1 / Preparation of plasma samples

[0090] The samples used according to the invention are prepared from whole blood taken from the individual. For this purpose, whole blood is collected in Cell-Free DNA Collection Tubes according to the recommendations of the supplier Roche Diagnostics.

[0091] Each sample thus collected is subjected to two successive centrifugations: - a first gentle centrifugation at 1600 g, at room temperature (20°C + / - 5°C), for 10 min, this first centrifugation being carried out within a maximum period of 7 days after collection; the plasma (supernatant) is recovered without collecting the cell pellet, - and a 2nd centrifugation at a faster speed, at 4500 g, also at room temperature (20°C + / - 5°C) and for 10 minutes; the plasma is aspirated without collecting the pellet formed.

[0092] The plasma thus recovered can be used immediately, or it can be stored at -20°C for a period less than or equal to 1 month, or -80°C for periods greater than 1 month, before its analysis.

[0093] 2 / Analysis of plasma samples for circulating cell-free DNA concentrations in different size ranges

[0094] The total concentration of circulating cell-free DNA in the plasma samples, and the concentration profiles as a function of size, are determined by a method based on pLas microfluidic technology. More specifically, the protocol implemented is as described in the publication by Boutonnet et al., Analytical Chemistry, 2023, 95(24): 9263-70.

[0095] The equipment used is as follows: - Agilent G7100A CE electrophoresis system equipped with a capillary incorporating pLAS technology, the diameter of which has a local restriction, - Picometrics fluorescence detector, Zetalif® LED 480.

[0096] The protocol implemented is schematically as follows.

[0097] Plasma samples are first pretreated with proteinase K in the presence of detergent to release nucleic acids from the vesicles and nucleoprotein complexes in which they are most often trapped. For this purpose, they are incubated with an aqueous solution of proteinase K (2 mg / ml) and NP-40 non-ionic surfactant (1%) at 56°C for 2 hours with stirring. vigorous (900 rpm), then recovered by centrifugation.

[0098] Each plasma sample is injected into the electrophoresis system, the temperature of which is maintained at 25 °C. The various ingredients and solutions used for the analysis are as follows: - RNASE: RNase Away (Biolab), - BSA water: a solution of BSA at 50 mg / ml in purified water, - 0.1 M HCl: 0.1 M hydrochloric acid solution in water, - PVA 2%: polyvinyl alcohol solution in water (800 mg in 40 ml of water), - TAE buffer: Tris-Acetate-EDTA (TAE) 0.5X + bovine serum albumin (BSA) 0.5 mg / mL, pH 8, polyvinylpyrrolidone (PVP) 360kDa 5% (w / v), water, - TAE buffer + dye: TAE buffer + SybrGreen 2X (Sigma), - PIPES buffer: Bis-tris 30 mM, piperazine-N,N'-bis(2-ethanesulfonic acid) (Pipes) 10 mM, pH 6.5, EDTA 1 mM, PVP 360 kDa 5% (w / v), water, - PIPES stamp + dye: PIPES stamp + SYBR Green 2X.

[0099] The exact program implemented in the electrophoresis system is detailed in Tables 2 to 4.

[0100] [Tables2] Inlet Outlet Pressure (bar) Voltage (kV) Time (min) Preconditioning Bin RNASE 10 0 4 RNASE Bin -9.9 0 4 WATER BSA Bin 10 0 4 HCl 0.1 M Bin 10 0 2 PVA 2% Bin 10 0 3 TAE Buffer Bin 10 0 6 Injection Sample Bin 5 0 1.33 (80s) Transfer Buffer TAE Buffer 5 0 0.17 (10s)

[0101] Table 2 - Electrophoresis system program - Preconditioning, Injection and Transfer

[0102] [Tables3] Inlet Outlet Pressure (bar) Voltage (kV) Time (min) Concentration with returns TAE buffer + dye (pH 8) TAE buffer (pH 8) 11 25 0 11 25 1 -11 0 1.01 -11 0 1.75 11 25 1.76 11 25 2.76 -11 0 2.77 -11 0 3.52 11 25 3.53 11 25 4.53 -11 0 4.54 -11 0 5.29 11 25 5.3 11 25 7.3 -11 0 7.31 -11 0 8.06 11 25 8.07 11 25 11.07 -11 0 11.08 -11 0 11.58 11.25 11.59 11.25 16 -11 0 16.01 -11 0 16.51 11.25 16.52 11.25 20.52 -11 0 20.53 -11 0 21.03

[0103] Table 3 - Electrophoresis system program - Concentration with feedback

[0104] [Tables4] Inlet Outlet Pressure (bar) Voltage (kV) Time (min) Buffer change and end of concentration at 7 bar Buffer PIPES + dye Buffer PIPES 7 30 0 7 30 9 Separation Buffer PIPES + dye Buffer PIPES 7 30 9 7 10 9.40 7 5 9.80 7 0.6 13 7 0.1 24 7 0 24.1

[0105] Table 4 - Electrophoresis system program - Buffer change and end of concentration, separation

[0106] Following the analysis, a fluorescence intensity profile as a function of time is obtained for each plasma sample analyzed. An example of such a profile is shown in [Fig. 1].

[0107] The fluorescence profile is then converted into a concentration profile using the fluorescence intensity profile of a standard sample comprising DNA fragments whose respective sizes, ranging from 100 base pairs (“bp”) to 1500 base pairs, are known and for which, for each size of DNA fragment, the migration time, fluorescence intensity and concentration are known (allowing the fluorescence-to-concentration conversion factor to be known for each migration time of the peaks of the standard sample).

[0108] This standard sample has the composition shown in Table 5:

[0109] [Tables5] DNA Size (bp) 100 150 200 300 400 500 600 700 800 1000 1500 Concentration (pg / pl) 9 3 5 4 3 1.7 1.4 0.7 1.4 1.4 1.4

[0110] Table 5 - Composition of the standard sample (Conc. denotes concentration)

[0111] The migration profile of this standard sample in the device used is shown on the [Fig.2].

[0112] These data are used to perform, on the one hand, a calibration of fluorescence, and, on the other hand, a calibration of migration times.

[0113] For fluorescence calibration, the area of ​​each peak of the standard sample is measured (in RFU.min) and then divided by the concentration of the corresponding fragment. To obtain an instantaneous conversion factor, the result is divided by the unit of time (1 min), thus yielding, for each peak in [Fig. 2], a fluorescence value per pg.pL for each migration time of the standard sample. Virtual points can be added outside the curve expressing the conversion factor (RFU / pg / pL.min) as a function of the migration time thus obtained, for both short and long migration times. Linear interpolation is performed between the points of this curve, in a conventional manner, to determine the fluorescence-to-concentration conversion factor for all migration times between the first and last peaks of the standard sample.A slight linear extrapolation is performed on both sides to slightly broaden the migration time range within which the conversion factor is known. Here, the extrapolation is made between 75 and 100 bp and between 1500 and 1650 bp. This results in a curve composed of straight segments, which is smoothed using second-degree splines, in a conventional manner. This smoothed curve is used to convert the fluorescence value into a concentration value throughout the analysis. Thus, the fluorescence-time curve for each given sample is converted into a concentration-time curve.

[0114] For calibrating migration times, the fragments of the standard sample provide "migration time / DNA size" pairs. A linear interpolation is performed, in a conventional manner, between each point to obtain a conversion of migration times into DNA size. Additional points can be added to the curve representing DNA size versus migration time thus obtained, the migration times of these additional points being calculated relative to the nearest experimental DNA fragments in the standard sample, in order to improve the accuracy of DNA size calibration without having to increase the number of DNA fragments in the standard sample. After a further linear interpolation, a curve representing DNA size versus migration time is obtained.

[0115] Using these two calibrations, that of fluorescence versus concentration, and that of migration time versus DNA size, the fluorescence curve as a function of time is converted, for each plasma sample analyzed, into a concentration curve as a function of size.

[0116] Thus, for each plasma sample analyzed, a profile of cfDNA concentration intensity as a function of size is obtained. An example of such a profile is shown on the [Fig.3].

[0117] By this method, concentrations are only quantitatively determined between 75 and 1650 base pairs. Thus, for fragments smaller than 75 base pairs and fragments larger than 1650 base pairs, a relative value of the concentration, in arbitrary units, is calculated. This relative value is, however, highly representative of the true value.

[0118] To determine the cfDNA concentration in each desired size range, the minimum and maximum sizes of the interval are converted into minimum and maximum times according to the time / size calibration, and then the integral of the time / concentration curve is calculated between these minimum and maximum times. This gives the absolute concentration of cfDNA in each desired size range, expressed in pg / pL.

[0119] For each plasma sample analyzed, from the individual concentrations of each size, a histogram of the percentages of the following different size ranges is constructed: size less than 75 bp, from 75 to 239 bp, from 240 to 369 bp, from 370 to 579 bp, from 580 to 1649 bp, greater than or equal to 1650 bp. An example of such a histogram is shown in [Fig. 4].

[0120] For each plasma sample analyzed, the total cfDNA concentration is also calculated. In this example, this total concentration is approximated to the DNA concentration between sizes 75 bp and 1650 bp, the value thus approximated being very close to the true value.

[0121] Finally, for each plasma sample analyzed, the relative concentration of DNA is calculated in each of the above size ranges, this relative concentration being equal to the absolute concentration of DNA in that size range, determined as indicated above, divided by the total concentration of DNA.

[0122] 3 / Statistical analysis

[0123] Preliminary statistical tests including classification by univariate logistic regression, univariate and multivariate COX regression and Kaplan-Meier estimation to classify early progressors and non-early progressors on a first set of 25 patients with a distribution of pathologies equivalent to that of the entire cohort of 51 patients were first carried out, for the following potential indicator panels: relative concentrations of cfDNA fragments in plasma, for the following respective size ranges: size less than 75 bp, from 75 to 239 bp, from 240 to 369 bp, from 370 to 579 bp, from 580 to 1649 bp, greater than or equal to 1650 bp.

[0124] The two indicators that characterize the patient's response (early progressors or non-early progressors under immunotherapy treatment) are the relative concentrations of cDNA fragments in the plasma for the following size ranges: size greater than or equal to 1650 base pairs, and size between 580 and 1649 base pairs.

[0125] The mathematical functions (indicators) used for statistical analysis, named PI and P2, are therefore as follows:

[0126] [Math.l] Concentration of cfDNA fragments of size greater than or equal to 1650 bp Total cfDNA concentration

[0127] [Math.2] Concentration of cfDNA fragments with sizes between 580 and 1649 bp; Total cfDNA concentration

[0128] The cohort of 51 patients is divided into two data sets: a training set (70% - 35 patients) and a test set (30% - 16 patients). This distribution between the training and test sets is standard and corresponds to the standard methodology for this type of statistical analysis. Both sets (training and test) show a balanced distribution in terms of patients with progressive disease, pathology, and sex.

[0129] Statistical data are established in relation to clinical data obtained by the iRECIST methodology 6 to 12 weeks after the start of immunotherapy treatment, depending on the type of cancer.

[0130] For each indicator, a ROC curve is generated on the training set to determine the threshold that minimizes the distance from the optimal point (0,1). The method used to set the optimal threshold value is the bootstrap method. For 100 experiments, patients are randomly sampled with replacement between the training set ("in bag") and the test set ("out of bag").

[0131] For each indicator, the threshold is optimized and defined as the median of the 100 thresholds obtained by the BOOTSTRAP method.

[0132] These thresholds are then used on the training batch and then the test batch, and the following metrics are determined: - area under the curve (AUC, for the English "Area Under Curve"), - classification accuracy, - sensitivity, - specificity, - Positive predictive value (PPV): the probability that the patient predicted to be an early progressor is actually an early progressor. - negative predictive value (NPV, for the English "Negative Predictive Value"): probability that the patient predicted as an early non-progressor is an early non-progressor.

[0133] The thresholds determined by the BOOTSTRAP method are further used for calculate a 95% confidence interval of the threshold values ​​by extracting the quantiles at 2.5% and 97.5%.

[0134] The ROC curves obtained for each indicator, respectively for the training group of 35 patients, the test group of 16 patients, and the total cohort, are shown in [Fig. 5] for the PI indicator and in [Fig. 6] for the P2 indicator. The area under the curve (AUC) values ​​are shown in the figures for each curve. This value is observed to be high for both indicators, and is higher for the PI indicator ([Fig. 5]) than for the P2 indicator ([Fig. 6]). For comparison, for example, for the indicator of the relative concentration of cfDNA fragments in the size range greater than or equal to 240 bp, the total cohort obtained an AUC of only 0.65.

[0135] Figure 7 shows the box plots obtained for each of the PI indicators. and P2, for the entire cohort. We observe that the results are statistically different between early progressors and non-early progressors.

[0136] All the results obtained, for the training lot (“A”) and for the test lot (“T”) are summarized in Table 6, for each of the PI and P2 indicators.

[0137] [Tableauxô] PI Indicator P2 Threshold 0.034 0.072 Quantile 2.5% 0.028 0.062 Quantile 97.5% 0.043 0.122 AUC A 0.830 0.693 AUC T 0.833 0.817 Accuracy A 0.800 0.771 Accuracy T 0.875 0.812 Sensitivity A 0.733 0.533 Sensitivity T 0.833 0.500 Specificity A 0.850 0.950 Specificity T 0.900 1.000 PPV A 0.786 0.889 PPVT 0.833 1.000 NPV A 0.810 0.731 NPVT 0.900 0.769

[0138] Table 6 - Characteristic statistical values ​​for the PI and P2 indicators

[0139] For the PI indicator, we obtain a sensitivity of 0.833 and a specificity of 0.900, for a threshold (score below which the patient is considered an early progressor) of 0.034 (confidence interval of 0.028 to 0.043).

[0140] The P2 indicator, for its part, exhibits excellent specificity of 1 when considering a threshold of 0.072 (confidence interval of 0.062 to 0.122). This indicator thus has good predictive power regarding tumor progression at the first examination: if it is low, then there will very likely (better than 90%) be disease progression at the first examination. Conversely, if it is high, its predictive power for the absence of early progression is moderate, lower than that of the PI marker, which offers the best compromise between sensitivity and specificity.

[0141] For the test group, progression-free survival curves and survival probabilities were estimated using the Kaplan-Meier method and compared to the prediction results using the PI indicator according to the invention, with a threshold value of 0.034. The results are shown in [Fig. 8]. It can be observed that the curves are well correlated and characteristic of a long response, and therefore of a good prognosis for progression-free survival. Thus, the PI indicator constitutes a marker for predicting progression-free survival.

Claims

Demands

1. An in vitro method for predicting the status of early progressor and / or non-early progressor during immunotherapy treatment of a subject with cancer, comprising the steps of: a / determining, in an isolated blood sample obtained from said subject before said immunotherapy treatment: - the total concentration of circulating free deoxyribonucleic acid in said blood sample, referred to as the total cfDNA concentration, - at least one concentration, in said blood sample, of circulating free deoxyribonucleic acid fragments whose size is included within a predetermined size range, said predetermined size range being within sizes greater than 500 base pairs, referred to as the large cfDNA concentration, b / combining in a mathematical function at least said large cfDNA concentration and said total cfDNA concentration,so as to obtain a score, and: c1 / comparison of said score to a first predetermined reference value, and d1 / conclusion that said subject exhibits the characteristics of an early progressor during said immunotherapy treatment when said score is less than or equal to said first reference value, and / or c2 / comparison of said score to a second predetermined reference value, and d2 / conclusion that said subject exhibits the characteristics of a non-early progressor during said immunotherapy treatment when said score is greater than said second reference value.

2. A method according to claim 1, wherein said first reference value and said second reference value are identical.

3. A method according to claim 1, wherein said first reference value is less than said second reference value.

4. A method according to any one of claims 1 to 3, wherein said mathematical function is the ratio between at least said concentration of large cfDNA, and said total cfDNA concentration.

5. A method according to any one of claims 1 to 4, wherein said predetermined size range is the range of sizes greater than 500 base pairs, preferably greater than or equal to 600 base pairs, and preferably even greater than or equal to 1500 base pairs.

6. A method according to any one of claims 1 to 4, wherein said predetermined size range is the size range greater than or equal to 1650 base pairs.

7. A method according to any one of claims 1 to 4, wherein said predetermined size range is the size range between 580 and 1649 base pairs.

8. A method according to any one of claims 1 to 7, comprising a step of obtaining a plasma sample from said blood sample.

9. A method according to claim 8, wherein the determination of said total cfDNA concentration and the determination of said at least one large cfDNA concentration are carried out by direct analysis of said plasma sample.

10. A method according to any one of claims 1 to 9, comprising a step of extracting circulating free deoxyribonucleic acid from said blood sample, or where appropriate from said plasma sample, and wherein the determination of said total cfDNA concentration and the determination of said at least one large cfDNA concentration are carried out by analysis of the circulating free deoxyribonucleic acid extract thus obtained.

11. A method according to any one of claims 1 to 10, wherein said subject is a human.

12. A method according to any one of claims 1 to 11, wherein said cancer is a metastatic tumor pathology, a melanoma, a kidney cancer, a urothelial carcinoma of the bladder, a squamous cell carcinoma of the head and neck, or a non-small cell lung cancer.

13. A method according to any one of claims 1 to 12, wherein said immunotherapy treatment uses nivolumab, ripilimumab, pembrolizumab, atezolizumab, avelumab and / or durvalumab.

Images