Methods of predicting the risk of recurrence and / or death of patients suffering from a hepatocellular carcinoma (HCC)

Abstract

Circulating tumor DNA (ctDNA) is a promising non-invasive biomarker in cancer. The inventors aim to assess the dynamic evolution of ctDNA in patients with hepatocellular carcinoma (HCC). For that, they analyzed 772 plasmas from 173 patients with HCC collected at the time of diagnosis or treatment (n=502), 24h after locoregional treatment (n=154), and during follow-up (n=116). All samples were analyzed for cell free DNA (cfDNA) concentration, and for mutations in TERT promoter, CTNNB1, TP53, PIK3CA and NFE2L2 by sequencing and droplet-based digital PCR. Results were compared to 232 corresponding tumor samples. In active HCC, the inventors identified 27-5% of mutations in TERT promoter, 21 -3% in TP53, 13- 1% in CTNNB1, 0-4% in PIK3CA and 0-2% in NFE2L2, most of the times similar to those identified in the corresponding tumor. In 103 patients treated by percutaneous ablation, the presence and number of mutations in the ctDNA before treatment were associated with higher risk of death (p=0 001 ) and recurrence (p<0 001). In patients treated by atezolizumab / bevacizumab, persistence of mutation in ctDNA was associated with radiological progression (63-6% versus 36-4% for disappearance, p=0 019). Thus, circulating tumor DNA offers dynamic information reflecting tumor biology and represents a non-invasive tool useful to guide HCC clinical management.

Description

[0001] METHODS OF PREDICTING THE RISK OF RECURRENCE AND / OR DEATH OF PATIENTS SUFFERING FROM A HEPATOCEEEUEAR CARCINOMA (HCC)

[0002] FIELD OF THE INVENTION:

[0003] The present invention is in the field of medicine, in particular oncology.

[0004] BACKGROUND OF THE INVENTION:

[0005] Circulating tumor DNA (ctDNA) consists of small fragments of DNA, 100 to 150 base pairs in length, which are released in biological fluids from tumor cells1. Mutations can be detected in ctDNA, providing insight into the mutational landscape of tumors and can be utilized as a liquid biopsy. CtDNA shows promise as a biomarker and has been tested in various malignancies for cancer detection, prognosis prediction, residual disease monitoring, and identification of mechanisms of resistance to systemic treatments2. Furthermore, it has also been proposed as a tool for detecting tumor heterogeneity, as different mutations from various regions of the tumors can be identified in the same blood sample3.

[0006] In hepatocellular carcinoma (HCC), most of the somatic mutations and alterations belong to the following pathways: telomere maintenance TERT promoter, 60%), cell cycle (TP53, 30%; CDKN2A, 10%), Wnt / p-catenin (CTNNB1, 20%; AXIN1, 10%), oxidative stress (NFE2L2, 5%), epigenetic modifications (ARID1A, 15%; ARID2, 5%; BAP1, 5%), and Akt / mTOR and MAP kinase (RPS6KA3, 8%; PIK3CA, 2%)4 6. Several pilot studies have highlighted the potential to detect ctDNA in patients with I IC C7 0. However, these studies are often limited by the small number of samples that have been analyzed and the absence of analysis of concomitant tumor. Additionally, the utility of ctDNA for detecting small HCC remains a topic of controversy11,12. Deeper investigations are also needed to determine the role of ctDNA in monitoring treatments such as locoregional or systemic therapies, including atezolizumab / bevacizumab, the current standard of care in advanced HCC.

[0007] SUMMARY OF THE INVENTION:

[0008] The present invention is defined by the claims. In particular, the present invention relates to methods of predicting the risk of recurrence and / or death of patients suffering from a hepatocellular carcinoma (HCC). DETAILED DESCRIPTION OF THE INVENTION:

[0009] Circulating tumor DNA (ctDNA) is a promising non-invasive biomarker in cancer. The inventors aim to assess the dynamic evolution of ctDNA in patients with hepatocellular carcinoma (HCC). For that, they analyzed 772 plasmas from 173 patients with HCC collected at the time of diagnosis or treatment (n=502), 24h after locoregional treatment (n=154), and during follow-up (n=116). For controls, 56 plasmas from patients with chronic liver disease without HCC were analyzed. All samples were analyzed for cell free DNA (cfDNA) concentration, and for mutations in TERT promoter, CTNNB1, TP53, PIK3CA and NFE2L2 by sequencing and droplet-based digital PCR. Results were compared to 232 corresponding tumor samples. In patients with active HCC, 40-2% of the ctDNA was mutated versus 14-6% in patients with inactive HCC and 1 -8% in controls (p<0-001). In active HCC, the inventors identified 27 5% of mutations in TERT promoter, 21 3% in TP53, 13 1% in CTNNB1, 0 4% in PIK3CA and 0-2% in NFE2L2, most of the times similar to those identified in the corresponding tumor. CtDNA mutation rate increased with advanced tumor stages (p<0-001). In 103 patients treated by percutaneous ablation, the presence and number of mutations in the ctDNA before treatment were associated with higher risk of death (p=0 001) and recurrence (p<0-001). Interestingly, cfDNA concentration and detectable mutations increased 24 hours after a loco- regional treatment. Among 356 plasmas collected in 53 patients treated by systemic treatments, the inventors detected mutations at baseline in 60-4% of the cases. In patients treated by atezolizumab / bevacizumab, persistence of mutation in ctDNA was associated with radiological progression (63’6% versus 36 4% for disappearance, p=0 019). In two patients progressing under systemic treatments, the inventors detected occurrence of mutations of CTNNB1 in the plasma, not present at baseline. Thus, circulating tumor DNA offers dynamic information reflecting tumor biology and represents a non-invasive tool useful to guide HCC clinical management.

[0010] Main definitions:

[0011] As used herein, the term “hepatocellular carcinoma” or “HCC” has its general meaning in the art and refers to a malignant tumor of hepatocellular origin that may develop in patients with risk factors that include alcohol abuse, viral hepatitis, and metabolic liver disease and often on cirrhosis, or a chronic liver disease. HCC is a type of primary liver cancer. Vascular invasion, particularly of the portal system, is common. Aggressive HCC can cause hepatic rupture and hemoperitoneum. As used herein, the term "patient" or “subject” refers to any vertebrate including, without limitation, humans and other primates (e.g., chimpanzees and other apes and monkey species), farm animals (e.g., cattle, sheep, pigs, goats and horses), domestic mammals (e.g., dogs and cats), laboratory animals (e.g., rodents such as mice, rats, and guinea pigs), and birds (e.g., domestic, wild and game birds such as chickens, turkeys and other gallinaceous birds, ducks, geese, and the like). In some embodiments, the subject may be a mammal, preferably a human.

[0012] As used herein, the term “mutation” has its general meaning in the art and includes any form of substitution(s), deletion(s), rearrangements) and / or insertion(s) in the coding and / or noncoding region of the locus, alone or in various combination(s) in DNA. Mutations more specifically include point mutations. Deletions may encompass any region of two or more residues in a coding or non-coding portion of the gene locus, such as from two residues up to the entire gene or locus. Typical deletions affect smaller regions, such as domains (introns) or repeated sequences or fragments of less than about 50 consecutive base pairs, although larger deletions may occur as well. Insertions may encompass the addition of one or several residues in a coding or non-coding portion of the gene locus. Insertions may typically comprise an addition of between 1 and 50 base pairs in the gene locus. Rearrangement includes inversion of sequences. The mutation may result in the creation of stop codons, frameshift mutations, amino acid substitutions, particular RNA splicing or processing, product instability, truncated polypeptide production, etc. The alteration may result in the production of a polypeptide with altered function, stability, targeting or structure. The mutation may also cause a reduction in protein expression or, alternatively, an increase in said production.

[0013] As used herein, the term 'TERT' has its general meaning in the art and refers to the gene encoding for the telomerase reverse transcriptase. The term is also known as TP2; TRT; CMM9; EST2; TCS1; hTRT; DKCA2; DKCB4; hEST2; and PFBMFT1. The genomic NCBI reference sequence is NG 009265.1 and the NCBI Gene ID is 7015. The promoter region of the TERT gene is well characterized. Telomerase reverse transcriptase TERT) is thus a gene that encodes an enzymatic protein that possesses reverse transcriptase activity. The protein also functions as an RNA component that serves as a template for the telomere repeat TTAGGG (SEQ ID NO:1). Point mutations such astdeletions and insertions in the promoter, missense mutations, nonsense mutation and silent mutations are observed in cancers such as cancers of the central nervous system, hepatocellular carcinoma, and cancers of the urinary tract. As used herein, the term ^TP53 has its general meaning in the art and refers to the gene encoding for the tumor protein p53. The term is also known as BCC7, BMFS5, LFS1, P53, and TRP53. The genomic NCBI reference sequence is NG 017013.2 and the NCBI Gene ID is 7157. Tumor protein p53 (TP53) is thus a gene that codes for a tumor suppressor protein, cellular tumor antigen p53. The protein regulates expression of genes involved in cell cycle arrest, apoptosis, senescence, DNA repair, and changes in metabolism. In cancer, TP53's normal roles are not fulfilled, leading to cell survival, DNA damage, and cell proliferation. TP 53 is the most frequently mutated gene in cancer; it is mutated in about half of all cancers.

[0014] As used herein, the term “CTNNBT” has its general meaning in the art and refers to the gene encoding for the catenin beta 1 protein. The term is also known as CTNNB, EVR7, MRD19, NEDSDV, and armadillo. The genomic NCBI reference sequence is NG_013302.2 and the NCBI Gene ID is 1499. CTNNB 1 (catenin (cadherin-associated protein), beta 1, 88kDa) is thus a gene that encodes catenin beta-1 protein (also known as beta-catenin). beta-catenin is part of a complex of proteins that form adherens junctions, which are important for the establishment and maintenance of epithelial cell layers by regulating cell growth and adhesion between adjacent cells (PMID: 17854762). Mutant beta-catenin has been implicated in the pathogenesis of several cancers including melanoma, colorectal cancer, hepatocellular carcinoma, and ovarian cancer (PMID: 12781368).

[0015] As used herein, the term NFE2L2 " has its general meaning in the art and refers to the NFE2 like bZIP transcription factor 2. The term is also known as HEBP1, IMDDHH, NRF2, and Nrf- 2. The NCBI Gene ID is 4780. Nuclear factor, erythroid 2-like 2 (NFE2L2) is thus a gene that encodes a transcription factor protein that is a member of a family of leucine zipper proteins. The protein functions in the regulation of antioxidant response elements that promote immunological responses such as inflammation. Missense mutations, nonsense mutations, silent mutations, and in-frame insertions and deletions are observed in cancers such as esophageal cancer, hepatocellular carcinoma, and lung cancer.

[0016] As used herein, the term "PIK3CA' has its general meaning in the art and refers to the phosphatidylinositol-4,5-bisphosphate 3 -kinase catalytic subunit alpha. The term is also known as CCM4, CLAPO, CLOVE, CWS5, MCAP, MCM, MCMTC, PI3K, PI3K-alpha, and pl 10- alpha. The genomic NCBI reference sequence is NG 012113.2 and the NCBI Gene ID is 5290. PIK3CA (phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit alpha) is thus a gene that encodes the protein phosphatidylinositol 4, 5 -bisphosphate 3-kinase catalytic subunit alpha isoform, a subunit of the PI3K protein. Phosphatidyl 3-kinases (PI3K) are a family of lipid kinases involved in many cellular processes, including cell growth, proliferation, differentiation, motility, and survival.

[0017] Methods of the present invention:

[0018] The present invention relates to a method of predicting the risk of recurrence and / or death of a patient suffering from a hepatocellular carcinoma (HCC) comprising the steps of i) detecting at least one mutation in TERT, TP53, and CTNNB1 (i.e. determining whether the TERT, TP53, and CTNNB1 genes are mutated) in a sample obtained from the patient, ii) determining the number of mutated genes wherein said number correlates with the risk of recurrence and / or death.

[0019] In some embodiments, the method of the present invention further comprises the step of detecting at least one mutation in NFE2L2 and PIK3CA genes (i.e. determining whether the NFE2L2 and PIK3CA genes are mutated).

[0020] Thus, in some embodiments, the present invention relates to a method of predicting the risk of recurrence and / or death of a patient suffering from a hepatocellular carcinoma (HCC) comprising the steps of i) detecting at least one mutation in TERT, TP53, CTNNB1, NFE2L2 and PIK3CA in a sample obtained from the patient, ii) determining the number of mutated genes wherein said number correlates with the risk of recurrence and / or death.

[0021] In some embodiments, the method of the present invention is particularly suitable for predicting the survival time of the HCC patient. In particular, the method of the present invention is particularly suitable for predicting the duration of the overall survival (OS), progression-free survival (PFS) and / or the disease-free survival (DFS) of the HCC patient. More particularly, the method of the present invention is particularly suitable for predicting the disease-free survival.

[0022] As used herein, the term “survival time” includes “progression-free survival”, “death-free survival” and “overall survival”. As used herein, the term “progression-free survival” or “PFS” in the context of the invention refers to the length of time during and after treatment during which, according to the assessment of the treating physician or investigator, the patient's disease does not become worse, i.e. , does not progress. As the skilled person will appreciate, a patient's progression-free survival is improved or enhanced if the patient experiences a longer length of time during which the disease does not progress as compared to the average or mean progression free survival time of a control group of similarly situated patients.

[0023] As used herein, the term “disease-free survival” or “DFS” has its general meaning in the art and is defined as the time from randomization to recurrence of tumor or death, and it is typically used in the adjuvant treatment setting. The term is also known as “relapse-free survival”.

[0024] As used herein, the term “overall survival” or “OS” in the context of the invention refers to the average survival of the patient within a patient group. As the skilled person will appreciate, a patient's overall survival is improved or enhanced, if the patient belongs to a subgroup of patients that has a statistically significant longer mean survival time as compared to another subgroup of patients. Improved overall survival may be evident in one or more subgroups of patients but not apparent when the patient population is analyzed as a whole.

[0025] As used herein, the expression “short survival time” indicates that the subject will have a survival time that will be lower than the median (or mean) observed in the general population of subjects. When the subject will have a short survival time, it is meant that the subject will have a “poor prognosis”. Inversely, the expression “long survival time” indicates that the subject will have a survival time that will be higher than the median (or mean) observed in the general population of subjects. When the subject will have a long survival time, it is meant that the subject will have a “good prognosis”.

[0026] In some embodiments, the method of the present invention is implemented for patients having a tumor that exceeds 20mm.

[0027] In some embodiments, the method of the present invention is implemented for patients that are at BCLC stage B, Stage C or Stage D. The Barcelona Clinic Liver Cancer (BCLC) staging system is well known in the art and includes 5 stages (Reig, Maria, et al. "BCLC strategy for prognosis prediction and treatment recommendation: The 2022 update. "Journal of hepatology 76.3 (2022): 681-693). Thus, in some embodiments, the method of the present invention is not implemented for patients at BCLC Stage 0 (Very early stage).

[0028] According to the present invention, at least one mutation in TERT gene is detected. More particularly, at least one mutation in TERT promoter is detected. Mutations in TERT genes, and in particular in TERT promoter are well known in the art and typically include those described in Quaas A, Oldopp T, Tharun L, Klingenfeld C, Krech T, Sauter G, et al. Frequency of TERT promoter mutations in primary tumors of the liver. Virchows Arch. 2014 Sep 30. In some embodiments, the mutations are located 124bp and 146bp upstream of the translation start site and typically include C228T and C250T.

[0029] According to the present invention, at least one mutation in TP53 gene is detected. Mutation in TP53 gene are well known in the art and typically cover covering all the TP53 coding sequence (exon 1,3, 4, 5, 6, 7, 8, 9). TP53 loss of function mutations have fully been exemplified in the prior art and thus the skilled man in the art can easily identify TP53 mutations (Petitjean A, Mathe E, Kato S, Ishioka C, Tavtigian SV, Hainaut P, Olivier M. Impact of mutant P53 functional properties on TP53 mutation patterns and tumor phenotype: lessons from recent developments in the IARC TP53 database.Hum Mutat. 2007 Jun;28(6):622-9) (http: / / P53.iarc.fr / )). TP53 mutations are mainly missense mutations. Examples of TP53 mutations include but are not limited to L130H, R175H, Y163C, C242Y, Y243H, G245S, M246V, R248W, R248Q, R249S, R249K, R273H, C277Y and C277F.

[0030] According to the present invention, at least one mutation in CTNNB1 gene is detected. Mutations in CTNNB1 gene are well known in the art and typically include those located in exon 3,7, and 8, and more particularly in exon 3. Those mutations have been fully exemplified in the prior art (see e.g. Kim S, Jeong S. Mutation Hotspots in the f-Catenin Gene: Lessons from the Human Cancer Genome Databases. Mol Cells. 2019 Jan 31;42(1):8-16). Examples of CTNNB1 mutations include but are not limited to S33A, S33C, S37F, S45F, S45A, S45P, T41A and S45P.

[0031] In some embodiment, at least one mutation in NFE2L2 gene is detected. Mutations in NFE2L2 gene are well known in the art and typically include those located in exon 2. The most common alterations in NFE2L2 are L14H, L14P, R34G, E79Q, E79K and T80I. In some embodiment, at least one mutation in PIK3CA gene is detected. Mutations in PIK3CA gene are well known in the art and typically include those located in exon 10 and 21. Those mutations have been fully exemplified in the prior art (see e.g. Lee JW, Soung YH, Kim SY, Lee HW, Park WS, Nam SW, Kim SH, Lee JY, Yoo NJ, Lee SH. PIK3CA gene is frequently mutated in breast carcinomas and hepatocellular carcinomas. Oncogene. 2005 Feb 17;24(8):1477-80). The most common alterations in PIK3CA are H1047R, H1047S and E545K.

[0032] As used herein, the term “sample” refers to any sample obtained from the patient for the purpose of performing the method of the present invention. In some embodiments, the sample is a bodily fluid (e.g. a blood sample), a population of cells, or a tissue. As used herein, the term “blood sample” refers to a whole blood sample, serum sample and plasma sample. A blood sample may be obtained by methods known in the art including venipuncture or a finger stick. Serum and plasma samples may be obtained by centrifugation methods known in the art. The sample may be diluted with a suitable buffer before conducting the assay. As used herein, the term “tumor biopsy sample” refers to a tumor sample that result from a biopsy performed in the primary tumor of the patient or performed in metastatic sample distant from the primary tumor of the patient. Typically tumor biopsy sample can results from a stereotactic needle biopsy.

[0033] According to the present invention, the sample contain nucleic acids (e.g. DNA) so that the mutations can be detected. In particular, samples comprising low levels (e.g., ng quantities or less) of nucleic acids (e.g., plasma samples) can be analyzed by methods described herein.

[0034] In some embodiments, the sample is a ctDNA sample. As used herein, the term “ctDNA” has its general meaning in the art and refers to the circulating tumor DNA, i.e. the free DNA originating from tumor, which suspends in blood of the patient. It is known that among DNA existing in blood only an extremely small amount of ctDNA generally exists as compared with normal cell DNA.

[0035] Generally, the DNA can be extracted from the sample, e.g. by a variety of techniques such as those described by Maniatis, et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, N.Y., pp. 280-281 (1982 ). In some embodiments, the sample can be processed using nucleic acid extraction conditions provided herein to achieve higher concentrations of nucleic acid sequences from the same amount of starting material than can be achieved using other conditions. Such nucleic acid extraction conditions involve isolation of cell-free DNA (cfDNA) from the biological sample, usually blood sample(e.g. plasma sample). In other examples, genomic DNA (gDNA) is extracted from the biological sample using any suitable method known in the art or described herein.

[0036] According to the present invention the mutation are detected by any method well known in the art. In current practice, mutation are detected in tissue-derived tumor DNA (tDNA) by next generation sequencing (NGS) or fluorescent in situ hybridization (FISH). Techniques such as polymerase chain reaction (PCR) and whole genome next-generation sequencing (WGS) are also commonly used to detect mutations. Other methods of sequencing include targeted plasma re-sequencing (TAm-Seq), massively paralleled sequencing (MPS), whole-genome sequencing (WGS), and whole exome sequencing (WES). In particular, mutations in ctDNA can be detected by plasma sequencing as described in the EXAMPLE. Briefly, blood is drawn from the patient, and the plasma DNA is isolated. The shotgun library is then prepared for sequencing. This step takes a total of 24 hours to perform. Shallow sequencing of the plasma DNA using an Illumina MiSeq machine takes about 12 hours to perform. Finally, alignment of the DNA sequences and analysis of the resulting data take between 2 and 3 hours. More recently, droplet digital PCR (ddPCR) was shown to be a fast and sensitive method for detection of mutations in ctDNA (Postel M., Roosen A., Laurent-Puig P., Taly V., Wang-Renault S.F. Droplet-based digital PCR and next generation sequencing for monitoring circulating tumor DNA: a cancer diagnostic perspective. Expert Rev. Mol. Diagn. 2018;18(l):7-17. Jan 2). By using multiplex ddPCR, several mutations can be detected in parallel in one reaction. Thus in some embodiments, the mutations are detected by droplet digital PCR. As used herein, the term “droplet digital PCR" or “ddPCR” refers to a digital PCR assay that measures absolute quantities by counting nucleic acid molecules encapsulated in discrete, volumetrically defined, water-in-oil droplet partitions that support PCR amplification (Hindson, et al, 2011. Analytical Chemistry 83, 8604-8610 ; Pekin, et al, 2011, Lab on a Chip 11, 2156-66 ; Pinheiro, et al, 2012, Anal Chem 84, 1003-1011). A single ddPCR reaction may be comprised of at least 20,000 partitioned droplets per well. A "droplet" or "water-in-oil droplet" refers to an individual partition of the droplet digital PCR assay. A droplet supports PCR amplification of template molecule(s) using homogenous assay chemistries and workflows similar to those widely used for real-time PCR applications. Droplet digital PCR may be performed using any platform that performs a digital PCR assay that measures absolute quantities by counting nucleic acid molecules encapsulated in discrete, volumetrically defined, water-in-oil droplet partitions that support PCR amplification. The strategy for droplet digital PCR may be summarized as follows: a sample is diluted and partitioned into thousands to millions of separate reaction chambers (water-in-oil droplets) so that each contains one or no copies of the nucleic acid molecule of interest. The number of "positive" droplets detected, which contain the target amplicon (i.e., nucleic acid molecule of interest), versus the number of "negative" droplets, which do not contain the target amplicon (i.e., nucleic acid molecule of interest), may be used to determine the number of copies of the nucleic acid molecule of interest that were in the original sample. Examples of droplet digital PCR systems include the QX200™ Droplet Digital PCR System by Bio-Rad, which partitions samples containing nucleic acid template into 20,000 nanoliter-sized droplets; the RainDrop™ digital PCR system by RainDance Technologies (today Biorad), which partitions samples containing nucleic acid template into 5,000,000 to 10,000,000 picoliter-sized droplets and the Crystal Digital PCR™ with the STILLA Naica® system which partitions samples containing nucleic acid template into 25,000 to 30,000 nano liter-sized droplets.

[0037] According to the present invention, the risk of recurrence and / or death increases with the detection of mutated genes. In particular, more genes mutated are detected, higher is risk of recurrence and / or death. Typically, a patient who harbors less than two mutated genes, i.e. a patient who harbors zero mutated gene or one mutated gene has a low risk of recurrence and / or death in comparison with a patient harboring two or more mutated genes. Similarly, a patient who harbors less than two mutated genes, i.e. a patient who harbors zero mutated gene or one mutated gene will have longer survival time in comparison with a patient harboring two or more mutated genes. Thus, in some embodiments, the more genes are mutated, the higher is the risk of recurrence and / or death and similarly. The more genes are mutated, the shorter will be the survival time of the patient. More particularly, a patient who harbors a mutation in two or more genes in ctDNA has a high risk of recurrence and / or death and thus will have a short survival time. Even more particularly, a patient who harbors a mutation in two or more genes in ctDNA exhibits an increased risk of extrahepatic recurrence in comparison to those with a single mutation or no mutations.

[0038] In some embodiments, the method of the present invention involves use of an algorithm. In some embodiments, the method of the present invention comprises the steps of a) assessing a first parameter that is the number of mutated genes, b) implementing an algorithm on data comprising or consisting of the parameter assessed at step a) as to obtain an algorithm output, the implementing step being computer-implemented; and c) determining the risk of recurrence and / or death from the algorithm output obtained at step b).

[0039] As used herein, the term “algorithm” is any mathematical equation, algorithmic, analytical or programmed process, or statistical technique that takes one or more continuous parameters and calculates an output value, sometimes referred to as an “index” or “index value.”

[0040] As used herein, the term “parameter” refers to any characteristic assessed when carrying out the method according to the invention. As used herein, the term “parameter value” refers to a value (a number for instance) associated to a parameter.

[0041] In some embodiments, the algorithm implements one or more additional parameters. Typically, the additional parameters are selected from the group consisting of age of the patient, tumor size and numbers, cfDNA concentration...

[0042] Non-limiting examples of algorithms include sums, ratios, and regression operators, such as coefficients or exponents, biomarker value transformations and normalizations (including, without limitation, those normalization schemes based on clinical parameters, such as gender, age, or ethnicity), rules and guidelines, statistical classification models, and neural networks trained on historical populations. Non- limiting examples of algorithms thus include logistic regression, linear regression, random forests, classification and regression trees (C&RT), boosted trees, neural networks (NN), artificial neural networks (ANN), neuro fuzzy networks (NFN), network structures, perceptrons such as multi-layer perceptrons, multi-layer feedforward networks, support vector machines (e.g., Kernel methods), multivariate adaptive regression splines (MARS), Levenberg-Marquardt algorithms, Gauss-Newton algorithms, mixtures of Gaussians, gradient descent algorithms, learning vector quantization (LVQ), and combinations thereof. Of particular use in combining parameters are linear and non-linear equations and statistical classification analyses to determine the relationship between levels of said parameters and the objective response to the preoperative adjuvant therapy. Of particular interest are structural and syntactic statistical classification algorithms, and methods of risk index construction, utilizing pattern recognition features, including established techniques such as cross-correlation, Principal Components Analysis (PCA), factor rotation, Logistic Regression (LogReg), Linear Discriminant Analysis (LDA), Eigengene Linear Discriminant Analysis (ELDA), Support Vector Machines (SVM), Random Forest (RF), Recursive Partitioning Tree (RPART), as well as other related decision tree classification techniques, Shrunken Centroids (SC), StepAIC, Kth-Nearest Neighbor, Boosting, Decision Trees, Neural Networks, Bayesian Networks, Support Vector Machines, and Hidden Markov Models, among others. Other techniques may be used in survival and time to event hazard analysis, including Cox, Weibull, Kaplan-Meier and Greenwood models well known to those of skill in the art.

[0043] In some embodiments, the method of the present invention comprises the use of a machine learning algorithm. The machine learning algorithm may comprise a supervised learning algorithm. Examples of supervised learning algorithms may include Average One-Dependence Estimators (AODE), Artificial neural network (e.g., Backpropagation), Bayesian statistics (e.g., Naive Bayes classifier, Bayesian network, Bayesian knowledge base), Case-based reasoning, Decision trees, Inductive logic programming, Gaussian process regression, Group method of data handling (GMDH), Learning Automata, Learning Vector Quantization, Minimum message length (decision trees, decision graphs, etc.), Lazy learning, Instance-based learning Nearest Neighbor Algorithm, Analogical modeling, Probably approximately correct learning (PAC) learning, Ripple down rules, a knowledge acquisition methodology, Symbolic machine learning algorithms, Subsymbolic machine learning algorithms, Support vector machines, Random Forests, Ensembles of classifiers, Bootstrap aggregating (bagging), and Boosting. Supervised learning may comprise ordinal classification such as regression analysis and Information fuzzy networks (IFN). Alternatively, supervised learning methods may comprise statistical classification, such as AODE, Linear classifiers (e.g., Fisher's linear discriminant, Logistic regression, Naive Bayes classifier, Perceptron, and Support vector machine), quadratic classifiers, k-nearest neighbor, Boosting, Decision trees (e.g., C4.5, Random forests), Bayesian networks, and Hidden Markov models. The machine learning algorithms may also comprise an unsupervised learning algorithm. Examples of unsupervised learning algorithms may include artificial neural network, Data clustering, Expectation-maximization algorithm, Self-organizing map, Radial basis function network, Vector Quantization, Generative topographic map, Information bottleneck method, and IBSEAD. Unsupervised learning may also comprise association rule learning algorithms such as Apriori algorithm, Eclat algorithm and FP-growth algorithm. Hierarchical clustering, such as Single-linkage clustering and Conceptual clustering, may also be used. Alternatively, unsupervised learning may comprise partitional clustering such as K-means algorithm and Fuzzy clustering. In some embodiments, the machine learning algorithms comprise a reinforcement learning algorithm Examples of reinforcement learning algorithms include, but are not limited to, temporal difference learning, Q-leaming and Learning Automata. Alternatively, the machine learning algorithm may comprise Data Preprocessing.

[0044] The method of the present invention is particularly suitable for discriminating responder from non-responder.

[0045] In particular, the method of the present invention is particularly suitable for determining whether the patient achieves a response after a preoperative therapy, a locoregional therapy and / or a postoperative therapy.

[0046] As used herein the term “responder” in the context of the present disclosure refers to a patient who will achieve a response, i.e. a patient where the cancer is eradicated, reduced or improved. According to the invention, the responders have an objective response and therefore the term does not encompass patients having a stabilized cancer such that the disease is not progressing after the therapy. A “non-responder” or “refractory patient” includes patients for whom the cancer does not show reduction or improvement after the therapy. According to the invention the term “non-responder” thus includes patients having a stabilized cancer. Typically, the characterization of the patient as a responder or non-responder can be performed by reference to a standard or a training set. The standard may be the profile of a patient who is known to be a responder or non-responder or alternatively may be a numerical value. Such predetermined standards may be provided in any suitable form, such as a printed list or diagram, computer software program, or other media. When it is concluded that the patient is a non-responder, the physician could take the decision to stop the therapy to avoid any further adverse sides effects. Typically, the method of the present invention is particularly suitable for predicting the pathological response. As used herein, the term “pathological response” refers to the response to the therapy that is assessed by any pathological method well known in the art that typically includes anatomic and histological assessment of the anti-tumoral response. Response may be recorded in a quantitative fashion or in a qualitative fashion like “no change” (NC), “partial response” (PR), “complete response” (CR) or other qualitative criteria. Thus, in some embodiments, multiple (e.g., at least 2, 3, 4, 5, or more) samples may be collected from a patient, over time or at particular time intervals, for example, to assess the disease progression or monitor the efficacy of a treatment.

[0047] Typically, the increase of mutated genes or the appearance of new mutated genes along the therapy indicates that the patient does not achieve a response to the therapy. Conversely, when a decrease of the number of mutated genes or even more the disappearance of mutated gene indicates that the patient achieves a response to the therapy. Thus, the persistence of mutations during the therapy is linked to a higher risk of progression in comparison with cases where mutations disappear.

[0048] Thus, in some embodiments, the method comprises the steps of a) detecting the number of mutated genes in a first sample obtained from the patient before the therapy, b) detecting the number of mutated genes in a second sample obtained from the patient after the therapy and c) comparing the numbers of mutated genes between the two samples wherein when the number of mutated genes in the second sample is equal or higher than the number of mutated genes in the first sample, it is concluded that the patient does not achieve a response to the therapy and whereas when the number of mutated genes in the second sample is lower than the number of mutated genes in the first sample of becomes null, it is concluded that the patient achieves a response to the therapy.

[0049] In some embodiments, the second sample is obtained from the patient about 24 hours after the therapy is administered.

[0050] In some embodiments, the second sample is obtained after 1, 2, 3 or more cycles of therapy. As used herein, the term “cycle” refers to a period of time during the therapy is administered to the patient. Typically, in cancer therapy a cycle of therapy is followed by a rest period during which no treatment is given. Following the rest period, one or more further cycles of therapy may be administered, each followed by additional rest periods.

[0051] As used herein, the term “therapy” refers to a timely sequential or simultaneous administration of anti-tumor, and / or anti vascular, and / or anti stroma, and / or immune stimulating or suppressive, and / or blood cell proliferative agents, and / or radiation therapy, and / or hyperthermia, and / or hypothermia for cancer therapy. The administration of these can be performed in an adjuvant and / or neoadjuvant mode. The composition of such protocol may vary in the dose of each of the single agents, timeframe of application and frequency of administration within a defined therapy window. Currently various combinations of various drugs and / or physical methods, and various schedules are under investigation.

[0052] As used herein, the term “locoregional therapy” is defined as minimally invasive image- guided liver tumor-directed procedure that can be categorized into ablative therapy, transcatheter therapy and radiation therapy (LlovetJM, De Baere T, Kulik L, et al. Locoregional therapies in the era of molecular and immune treatments for hepatocellular carcinoma. Nat Rev Gastroenterol Hepatol 2021;18:293-313). Typically, there are 2 major types of locoregional therapy used in hepatocellular carcinoma (HCC): percutaneous ablation (either chemical or thermal) and intraarterial chemoembolotherapy. Percutaneous ablation consists of inserting a needle directly into the tumor under image guidance (either ultrasound or computed tomography) to destroy the tumor through heating, freezing, or the application of alcohol. Thermal-based ablations include radiofrequency ablation, microwave ablation, and cryoablation. Non-thermic ablation include irreversible electroporation. Percutaneous ablation is typically reserved for early-stage HCC (Barcelona Clinic Liver Cancer [BCLC] stage A) and is therefore limited to tumors 3 to 4 cm or smaller and less than 3 in number. The other type of locoregional therapy is intraarterial-based and consists of utilizing the hepatic artery to deliver highly concentrated doses of chemotherapy or radiation directly to the tumor, sparing the healthy liver tissue in the process. Indeed, catheters can be manipulated deeply within the arterial liver circulation in close proximity to the tumors. After this step, toxic doses of chemotherapy (in the case of “chemoembolotherapy”, also called “transcatheter arterial chemo embolization” or “TACE”) or high doses of radiation (in the case of radioembolization) can be delivered directly to tumors. These procedures also are performed under image guidance, using fluoroscopy. Chemoembolization has become the mainstay of therapy for patients with unresectable HCC, and is considered the standard of care for patients with intermediate-stage disease (BCLC stage B).

[0053] As used herein, the “preoperative adjuvant therapy” or “neoadjuvant therapy” refers to a preoperative therapy regimen (before resection, ablation or transplantation) consisting of a panel of therapies that can include e.g. chemotherapy, radiotherapy, targeted therapy, hormone therapy and / or immunotherapy which is aimed to shrink the primary tumor, thereby rendering local therapy (e.g. surgery) less destructive or more effective, or enabling conserving surgery or enabling organ preservation.

[0054] As used herein, the “postoperative adjuvant therapy” or “adjuvant therapy” refers to a postoperative therapy regimen (after resection, ablation or transplantation) consisting of a panel of therapies that can include e.g. chemotherapy, radiotherapy, targeted therapy, hormone therapy and / or immunotherapy which is aimed to reduce risk of risk of metastasizing and / or relapse. The aim of such an adjuvant treatment is to improve the prognosis.

[0055] As used herein, the term “radiotherapy” has its general meaning in the art and refers a therapy with ionizing radiation. Ionizing radiation deposits energy that injures or destroys cells in the area being treated (the target tissue) by damaging their genetic material, making it impossible for these cells to continue to grow. One type of radiation therapy commonly used involves photons, e.g. X- rays. Depending on the amount of energy they possess, the rays can be used to destroy cancer cells on the surface of or deeper in the body. The higher the energy of the x-ray beam, the deeper the x-rays can go into the target tissue. Linear accelerators and betatrons produce x-rays of increasingly greater energy. The use of machines to focus radiation (such as x-rays) on a cancer site is called external beam radiation therapy. Gamma rays are another form of photons used in radiation therapy. Gamma rays are produced spontaneously as certain elements (such as radium, uranium, and cobalt 60) release radiation as they decompose, or decay. In some embodiments, the radiation therapy is external radiation therapy. Examples of external radiation therapy include, but are not limited to, conventional external beam radiation therapy; three-dimensional conformal radiation therapy (3D-CRT), which delivers shaped beams to closely fit the shape of a tumor from different directions; intensity modulated radiation therapy (IMRT), e.g., helical tomotherapy, which shapes the radiation beams to closely fit the shape of a tumor and also alters the radiation dose according to the shape of the tumor; conformal proton beam radiation therapy; image-guided radiation therapy (IGRT), which combines scanning and radiation technologies to provide real time images of a tumor to guide the radiation treatment; intraoperative radiation therapy (IORT), which delivers radiation directly to a tumor during surgery; stereotactic radiosurgery, which delivers a large, precise radiation dose to a small tumor area in a single session; hyperfractionated radiation therapy, e.g., continuous hyperfractionated accelerated radiation therapy (CHART), in which more than one treatment (fraction) of radiation therapy are given to a subj ect per day; and hypofractionated radiation therapy, in which larger doses of radiation therapy per fraction is given but fewer fractions.

[0056] As used herein, the term “chemotherapy” has its general meaning in the art and refers to the treatment that consists in administering to the patient a chemotherapeutic agent. As used herein, the term “chemotherapeutic agent” refers to a chemical compound that is (i.e., drug) or becomes (i.e., prodrug), for example, selectively destructive or selectively toxic to malignant cells and tissues, xamples of chemotherapeutic agents include but not limited to: alkylating agents such as thiotepa and cyclophosphamide (CYTOXAN®); alkyl sulfonates such as busulfan, improsulfan, and piposulfan; aziridines such as benzodepa, carboquone, meturedepa, and uredepa; ethylenimines and methylamelamines including altretamine, triethylenemelamine, triethylenephosphoramide, triethylenethiophosphoramide, and trimemylolomelamine; acetogenins, e.g., bullatacin and bullatacinone; a camptothecin, including synthetic analog topotecan; bryostatin, callystatin; CC-1065, including its adozelesin, carzelesin, and bizelesin synthetic analogs; cryptophy cins, particularly cryptophycin 1 and cryptophycin 8; dolastatin; duocarmycin, including the synthetic analogs KW-2189 and CBI- TMI; eleutherobin; 5-azacytidine; pancratistatin; a sarcodictyin; spongistatin; nitrogen mustards such as chlorambucil, chlomaphazine, cyclophosphamide, glufosfamide, evofosfamide, bendamustine, estramustine, ifosfamide, mechlorethamine, mechlorethamine oxide hydrochloride, melphalan, novembichin, phenesterine, prednimustine, trofosfamide, and uracil mustard; nitrosoureas such as carmustine, chlorozotocin, foremustine, lomustine, nimustine, and ranimustine; antibiotics such as the enediyne antibiotics (e.g., calicheamicin, especially calicheamicin gammall and calicheamicin phill), dynemicin including dynemicin A, bisphosphonates such as clodronate, an esperamicin, neocarzinostatin chromophore and related chromoprotein enediyne antibiotic chromomophores, aclacinomycins, actinomycin, authramycin, azaserine, bleomycins, cactinomycin, carabicin, carminomycin, carzinophilin, chromomycins, dactinomycin, daunorubicin, detorubicin, 6-diazo-5-oxo-L-norleucine, doxorubicin (including morpholino-doxorubicin, cyanomorpholino-doxorubicin, 2-pyrrolino- doxorubicin, and deoxydoxorubicin), epirubicin, esorubicin, idarubicin, marcellomycin, mitomycins such as mitomycin C, mycophenolic acid, nogalamycin, olivomycins, peplomycin, porfiromycin, puromycin, quelamycin, rodorubicin, streptonigrin, streptozocin, tubercidin, ubenimex, zinostatin, and zorubicin; anti-metabolites such as methotrexate and 5 -fluorouracil (5-FU); folic acid analogs such as demopterin, methotrexate, pteropterin, and trimetrexate; purine analogs such as cladribine, pentostatin, fludarabine, 6-mercaptopurine, thiamiprine, and thioguanine; pyrimidine analogs such as ancitabine, azacitidine, 6-azauridine, carmofur, cytarabine, dideoxyuridine, doxifluridine, enocitabine, and floxuridine; androgens such as calusterone, dromostanolone propionate, epitiostanol, mepitiostane, and testolactone; antiadrenals such as aminoglutethimide, mitotane, and trilostane; folic acid replinishers such as frolinic acid; radiotherapeutic agents such as Radium-223, 177-Lu-PSMA-617; trichothecenes, especially T-2 toxin, verracurin A, roridin A, and anguidine; taxoids such as paclitaxel (TAXOL®), abraxane, docetaxel (TAXOTERE®), cabazitaxel, BIND-014, tesetaxel; platinum analogs such as cisplatin and carboplatin, NC-6004 nanoplatin; aceglatone; aldophosphamide glycoside; aminolevulinic acid; eniluracil; amsacrine; hestrabucil; bisantrene; edatraxate; defofamine; demecolcine; diaziquone; elformthine; elliptinium acetate; an epothilone; etoglucid; gallium nitrate; hydroxyurea; lentinan; leucovorin; lonidamine; maytansinoids such as maytansine and ansamitocins; mitoguazone; mitoxantrone; mopidamol; nitracrine; phenamet; pirarubicin; losoxantrone; fluoropyrimidine; folinic acid; podophyllinic acid; 2- ethylhydrazide; procarbazine; polysaccharide-K (PSK); razoxane; rhizoxin; sizofiran; spirogermanium; tenuazonic acid; trabectedin, triaziquone; 2,2Z,2 / z-trichlorotriemylamine; urethane; vindesine; dacarbazine; mannomustine; mitobronitol; mitolactol; pipobroman; gacytosine; arabinoside ( “Ara-C” ); cyclophosphamide; thiopeta; chlorambucil; gemcitabine (GEMZAR®); 6-thioguanine; mercaptopurine; methotrexate; vinblastine; platinum; etoposide (VP- 16); ifosfamide; mitroxantrone; vancristine; vinorelbine (NAVELBINE®); novantrone; teniposide; edatrexate; daunomycin; aminopterin; xeoloda; ibandronate; CPT-11; topoisomerase inhibitor RFS 2000; difluoromethylomithine (DFMO); retinoids such as retinoic acid; capecitabine; NUC-1031; FOLFOX (folinic acid, 5 -fluorouracil, oxaliplatin); FOLFIRI (folinic acid, 5 -fluorouracil, irinotecan); FOLFOXIRI (folinic acid, 5-fluorouracil, oxaliplatin, irinotecan), FOLFIRINOX (folinic acid, 5-fluorouracil, irinotecan, oxaliplatin), and pharmaceutically acceptable salts, acids, or derivatives of any of the above. Such agents can be conjugated onto an antibody or any targeting agent described herein to create an antibody-drug conjugate (ADC) or targeted drug conjugate.

[0057] As used herein, the term “targeted therapy” refers to a therapy targeting a particular class of proteins involved in tumor development or oncogenic signaling. In particular, targeted therapy includes administration of tyrosine kinase inhibitors (TKIs). The term “tyrosine kinase inhibitor” (“TKI”) as used herein broadly refers to agents or compounds which are capable of selectively inhibiting tyrosine kinases family of enzymes but do no not target serine or threonine kinases, including those that inhibit MAPK (mitogen-activated protein kinases). The TKI may inhibit tyrosine kinase activity by directly acting on a tyrosine kinase molecule, or it may cooperate with one or more other factors or agents to achieve the desired inhibition. The tyrosine kinase family of enzymes includes both receptor tyrosine kinases and non-receptor tyrosine kinases. For instance, TKIs may target epidermal growth factor receptors (EGFRs) and receptors for fibroblast growth factor (FGF), platelet-derived growth factor (PDGF), and vascular endothelial growth factor (VEGF). Examples of TKIs include, but are not limited to sorafenib, tepotinib, cabozantinib, tivantinib, lenvatinib, afatinib, ARQ-087, asp5878, AZD3759, AZD4547, bosutinib, brigatinib, cediranib, crenolanib, dacomitinib, dasatinib, dovitinib, E-6201, erdafitinib, erlotinib, gefitinib, gilteritinib (ASP-2215), FP-1039, HM61713, icotinib, imatinib, KX2-391 (Src), lapatinib, lestaurtinib, midostaurin, nintedanib, ODM-203, osimertinib (AZD-9291), ponatinib, poziotinib, quizartinib, radotinib, rociletinib, sulfatinib (HMPL-012), sunitinib, and TH-4000. In some embodiments, the targeted therapy comprises at least one anti-VEGF agent. As used herein an "anti- VEGF agent" refers to a molecule that inhibits Vascular endothelial growth factor (VEGF) -mediated angiogenesis. For example, an anti-VEGF therapeutic may be an antibody or a Approved anti-VEGF agents useful in the present invention include but are not limited to bevacizumab (Avastin®, Genentech / Roche) an anti-VEGF monoclonal antibody; ramucirumab (Cyramza®, Eli Lilly), an anti-VEGFR-2 antibody and ziv-aflibercept, also known as VEGF Trap (Zaltrap®; Regeneron / Sanofi). VEGFR inhibitors, such as regorafenib (Stivarga®, Bayer); vandetanib (Caprelsa®, AstraZeneca); axitinib (Inlyta®, Pfizer); and lenvatinib (Lenvima®, Eisai); Raf inhibitors, such as sorafenib (Nexavar®, Bayer AG and Onyx); dabrafenib (Tafinlar®, Novartis); and vemurafenib (Zelboraf®, Genentech / Roche); MEK inhibitors, such as cobimetanib (Cotellic®, Exelexis / Genentech / Roche); trametinib (Mekinist®, Novartis); Bcr-Abl tyrosine kinase inhibitors, such as imatinib (Gleevec®, Novartis); nilotinib (Tasigna®, Novartis); dasatinib (Sprycel®, BristolMyersSquibb); bosutinib (Bosulif®, Pfizer); and ponatinib (Inclusig®, Ariad Pharmaceuticals); Her2 and EGFR inhibitors, such as gefitinib (Iressa®, AstraZeneca); erlotinib (Tarceeva®, Genentech / Roche / Astellas); lapatinib (Tykerb®, Novartis); afatinib (Gilotrif®, Boehringer Ingelheim); osimertinib (targeting activated EGFR, Tagrisso®, AstraZeneca); and brigatinib (Alunbrig®, Ariad Pharmaceuticals); c-Met and VEGFR2 inhibitors, such as cabozanitib (Cometriq®, Exelexis); and multikinase inhibitors, such as sunitinib (Sutent®, Pfizer); pazopanib (Votrient®, Novartis); and ALK inhibitors, such as crizotinib (Xalkori®, Pfizer). As used herein, the term “immunotherapy” has its general meaning in the art and refers to the treatment that consists in administering an immunogenic agent i.e. an agent capable of inducing, enhancing, suppressing or otherwise modifying an immune response. In some embodiments, the immunotherapy consists in administering the patient with at least one immune checkpoint inhibitor. As used herein, the term "immune checkpoint inhibitor" has its general meaning in the art and refers to any compound inhibiting the function of an immune inhibitory checkpoint protein. As used herein the term "immune checkpoint protein" has its general meaning in the art and refers to a molecule that is expressed by T cells in that either turn up a signal (stimulatory checkpoint molecules) or turn down a signal (inhibitory checkpoint molecules). Immune checkpoint molecules are recognized in the art to constitute immune checkpoint pathways similar to the CTLA-4 and PD-1 dependent pathways (see e.g. Pardoll, 2012. Nature Rev Cancer 12:252-264; Mellman et al. , 2011. Nature 480:480- 489). Examples of inhibitory checkpoint molecules include A2AR, B7-H3, B7-H4, BTLA, CTLA-4, CD277, IDO, KIR, PD- 1, LAG-3, TIM-3 and VISTA. Inhibition includes reduction of function and full blockade. Preferred immune checkpoint inhibitors are antibodies that specifically recognize immune checkpoint proteins. A number of immune checkpoint inhibitors are known and in analogy of these known immune checkpoint protein inhibitors, alternative immune checkpoint inhibitors may be developed in the (near) future. The immune checkpoint inhibitors include peptides, antibodies, nucleic acid molecules and small molecules. Examples of immune checkpoint inhibitor includes PD-1 antagonist, PD-L1 antagonist, PD-L2 antagonist CTLA-4 antagonist, VISTA antagonist, TIM-3 antagonist, LAG-3 antagonist, IDO antagonist, KIR2D antagonist, A2AR antagonist, B7-H3 antagonist, B7-H4 antagonist, and BTLA antagonist.

[0058] Examples of immune checkpoint inhibitors that can be used in the preoperative adjuvant immunotherapy include anti-CTLA4 antibodies, anti-PDl antibodies, anti-PDLl antibodies, anti-PDL2 antibodies, anti-TIM-3 antibodies, anti-LAG3 antibodies, anti-IDOl antibodies, anti-TIGIT antibodies, anti-B7H3 antibodies, anti-B7H4 antibodies, anti-BTLA antibodies, and anti-B7H6 antibodies.

[0059] Examples of PD-1 and PD-L1 antibodies are described in US Patent Nos. 7,488,802; 7,943,743; 8,008,449; 8,168,757; 8,217,149, and PCT Published Patent Application Nos: W003042402, WO2008156712, W02010089411, W02010036959, WO2011066342, WO2011159877, WO2011082400, and WO2011161699. In some embodiments, the anti-PDl antibody is selected from the group consisting of Pembrolizumab (Keytruda), Nivolumab (Opdivo), Cemiplimab (Libtayo) and Camrelizumab and Tislelizumab. In some embodiments, the anti- PD-L1 antibody is selected from the group consisting of atezolizumab, durvalumab, avelumab, LY3300054, BGB-A333, SHR-1316, CK-301, and combinations thereof.

[0060] In some embodiments, the patient is administered with a TKI selected from sorafenib (13), tepotinib, cabozantinib, tivantinib, and lenvatinib.

[0061] In some embodiments, the patient is administered with a combination of an anti-VEGF agent with an anti-PD-Ll antibody. In some embodiments, the patient is administered with a combination of bevacizumab and atezolizumab.

[0062] In some embodiments, the patient is administered with a combination of a CTLA-4 antibody and an anti-PD-Ll antibody. In some embodiments, the patient is administered with a combination of tremelimumab and durvalumab.

[0063] In some embodiments, the patient is administered with a combination of cabozantinib and atezolizumab.

[0064] In some embodiments, the patient is administered with a combination of lenvatinib and pembrolizumab.

[0065] In some embodiments, the patient is administered with a combination of rivoceranib and camrelizumab.

[0066] In some embodiments, the above-described method is implemented in many ways, notably using hardware, software or a combination thereof. In particular, each step is implemented by a module adapted to achieve the step or computer instructions adapted to cause the execution of the step by interaction with the system or a specific apparatus comprising the system. It should also be noted that two steps in succession may, in fact, be executed substantially concurrently or in a reverse order depending on the considered embodiments.

[0067] The method of the present invention is particularly suitable for orientating the clinical decisions after preoperative therapy or locoregional therapy. For instance, when it is concluded that the patient will have a high risk of recurrence and / or death, and / or a short survival time (e.g. DFS), a postoperative adjuvant therapy is then decided. The method of the present invention is thus particularly suitable for determining whether or not the patient is eligible to a postoperative adjuvant therapy. In some embodiments, the postoperative adjuvant therapy consists of a radiotherapy, a chemotherapy, a targeted therapy, an immunotherapy or a combination thereof. Said therapies are described above. In some embodiments, when it is concluded that the patient will have a low risk of recurrence and / or death, and in particular long time to recurrence and / or survival time (e.g. DFS), the administration of a postoperative adjuvant therapy may not be decided.

[0068] In some embodiment, in view of the currently limited options for HCC management, the method disclosed herein is useful for identifying patients with poor-prognosis, in particular patients with HCCs that are likely to relapse and metastasize. For instance, the method of the present invention can be used to identify patients in need of frequent follow-up by a physician or clinician to monitor HCC disease progression. Screening patients for identifying patients having a poor prognosis as disclosed herein is also useful to identify patients most suitable or amenable to be enrolled in clinical trial for assessing a therapy for HCC, which will permit more effective subgroup analyses and follow-up studies. Furthermore, the detection of mutated genes as disclosed herein can be monitored in patients enrolled in a clinical trial to provide a quantitative measure for the therapeutic efficacy of the therapy which is subject to the clinical trial. This invention also provides a method for selecting a therapeutic regimen or determining if a certain therapeutic regimen is more appropriate for a patient identified as having a poor prognosis as identified by the methods as disclosed herein. For example, an aggressive anticancer therapeutic regime can be pursued in a patient having a poor prognosis. Accordingly, patient identified with a poor prognosis can be administered a therapy that can comprises 1, 2 or 3 additional lines of treatment. In some embodiments, a patient can be monitored for HCC using the methods as disclosed herein, and if on a first (i.e. initial) testing the patient is identified as having a poor prognosis, the patient can be administered an anti-cancer therapy, and on a second (i.e. follow-up testing), the patient is identified as having a good prognosis, the patient can be administered an anti-cancer therapy at a maintenance dose. The method of the present invention is particularly suitable to determining which patients will be responsive or experience a positive treatment outcome to a treatment. The invention will be further illustrated by the following figures and examples. However, these examples and figures should not be interpreted in any way as limiting the scope of the present invention.

[0069] FIGURES:

[0070] Figure 1: Circulating tumor DNA and prognosis in percutaneous ablation. Overall Survival and Recurrence Free Survival of 103 patients undergoing a first ablation treatment according to (A) and (B) cfDNA concentration (ng / mL), (C) and (D) Detection of at least one gene in ctDNA. For cfDNA survival analysis, patients were divided by the median cfDNA concentration (ng / mL) value in high and low cfDNA level (Kaplan-Meier method, log-rank test). Numbers at risk are reported under the x-axis. (E) and (F) Number of mutated genes in ctDNA of patients who underwent first ablative treatment and developed a recurrence (n=71) according to the Milan criteria for transplant (E) and localization of recurrence (F) (y2tcst for trend). cfDNA: cell-free DNA. ctDNA: circulating tumor DNA. NM: non-mutated. M: mutated.

[0071] Figure 2. Survival analysis according to genes identified mutated in cfDNA. Overall Survival and Recurrence Free Survival, respectively, of 103 patients undergoing a percutaneous ablation for HCC according to presence of (A) and (B) TERT promoter mutations, of (C) and (D) CTNNB1 (E) and of (F) TP53 mutations identified in cfDNA. Results were represented using the Kaplan-Meier method and compared using the log-rank test. Numbers at risk are reported under the x-axis.

[0072] Figure 3. Overall survival according to tumor size in patients treated by percutaneous ablation. Overall Survival of 102 patients undergoing a first percutaneous ablation for HCC according to size of the main nodule (<20 mm or >20 mm, respectively) according to (A) and (B) cfDNA concentration (ng / pL), (C) and (D) detection of at least one mutation in cfDNA, (E) and (F) TERT promoter (G) and (H) CTNNB1 (I) and (J) TP53 mutations in cfDNA. For cfDNA level analysis, patients were divided into two groups according to whether the cfDNA concentration (ng / pL) was above (high) or below (low) the median concentration (ng / pL) value (Kaplan-Meier method and the log-rank test). Numbers at risk are reported under the x-axis. cfDNA: cell-free DNA. NM: non-mutated. M: mutated. T Figure 4: Early variation of circulating tumor DNA after loco-regional treatments. Overall Survival according to mutational status of couple of ctDNA collected at ablation (n=103) in four different groups: “double negativity” no mutations in the ctDNA found at HO and H24, “positive before” mutation at HO but not at H24, “positive after” no mutation at HO but mutation at H24, “double positivity” mutations at HO and H24 (Kaplan-Meier method with the log-rank test). Numbers at risk are reported under the x-axis. cfDNA: cell-free DNA. ctDNA: circulating tumor DNA. NM: non-mutated. M: mutated. HO: the day of treatment. H24: the day after the treatment

[0073] Figure 5. Recurrence free survival according to mutational status at HO and H24 patients treated by ablation. Recurrence Free Survival according to mutational status of couple of cfDNA collected at ablation (n=103) in four different groups: “double negativity” no mutations in the cfDNA found at HO and H24, “positive before” mutation at HO but not at H24, “positive after” no mutation at HO but mutation at H24, “double positivity” mutations both at HO and H24 (Kaplan-Meier method with the log-rank test). Numbers at risk are reported under the x- axis.

[0074] EXAMPLE:

[0075] Methods

[0076] Selection of patients and collection of clinical data

[0077] We included 229 patients (173 with HCC and 56 controls with chronic liver diseases but without HCC) referred to Avicenne Hospital, France between 2013 and 2021. Patients’ clinical and biological features and tumor characteristics at imaging (tumor size and number, macrovascular invasion and metastasis) and type of treatment received were collected at tumor and plasma sampling. Radiological response assessment was performed after four weeks for curative or locoregional treatment and at 12 weeks for systemic treatment and was categorized as complete response (CR), partial response (PR), stable disease (SD) or progressive disease (PD) according to the modified Response Evaluation Criteria in Solid Tumors (mRECIST)13. Tumor recurrence or progression and subsequent treatments received were recorded for each patient. All patients signed an informed consent and an ethic committee approved the current study (CCPPRB Paris Saint-Louis IRB00003835). Tumors

[0078] 232 tumor and 202 corresponding non-tumor samples (biopsy, n=230 and resection specimen, n=2) were collected in 149 patients. Tissues were frozen at -80°C after sampling, nucleic acid extraction, quantification and qualification were performed as previously described5. 232 tumor samples and their corresponding non-tumor tissues were sequenced: 213 for 39 driver genes by MiSeq5and 19 by WES (ten) and WGS (nine) as previously described5. Transcriptomic profiling was performed in 211 out of 232 tumor samples using quantitative RT-PCR of 283 genes as previously reported and classified using the G1-G6 transcriptomic classification and the prognostic 5-gene score14,15.

[0079] Plasma collection and cfDNA extraction

[0080] A total of 772 plasmas were collected in 173 patients with HCC and 56 plasmas in 56 patients with chronic liver diseases without any history of HCC (with a mean follow-up of more than 45.3 months after sampling) (“control” group) (Table 1).

[0081] Patients with HCC had 772 baseline and sequential plasma samples that were divided into three groups according to the timeline at which they were collected:

[0082] - 502 plasmas at the time of diagnosis of HCC, the day of treatment or at the time of radiological evaluation showing active HCC (“active HCC”);

[0083] - 116 plasmas after treatment of HCC without any active tumor at imaging at the time of plasma sampling (“inactive HCC”);

[0084] - 154 plasmas collected 24 hours after locoregional treatment (percutaneous ablation or trans-arterial embolization).

[0085] Plasma analysis by droplet-based digital PCR and MiSeq

[0086] 763 plasmas were tested for the presence of C228T TERT promoter alteration using dropletbased digital PCR (ddPCR) and sequenced for the hotspot mutations in TERT promoter, CTNNB1 (exon 3,7,8), PIK3CA (exon 10, 21) and NFE2L2 (exon 2), and covering all the TP53 coding sequence (exon 1,3, 4, 5, 6, 7, 8, 9) using MiSeq sequencing5.

[0087] Pipeline of analysis of mutations (tumors and plasmas)

[0088] In 149 tumors, 96 mutations of C228T TERT promoter were confirmed by both MiSeq sequencing and ddPCR and 5 mutations were identified by MiSeq only (data not shown). The variant allele frequency (VAF) in mutated samples were highly correlated between the two methods (R=0.886, p<0.001, data not shown). In 711 ctDNA, 84 mutations of C228T TERT promoter were confirmed by both methods, 68 by ddPCR only and 22 mutations by MiSeq only (data not shown). To note, mutations identified by ddPCR only or by MiSeq only have a lower VAF compared to mutations identified by both technics (data not shown). When a tumor was available (n=52 for ddPCR and n=16 for Miseq), the same mutation was observed in the tumor in 79% of the case for mutations identified by ddPCR only and in 75% of the case for mutations identified by Miseq only. In order to increase the sensitivity of detection of TERT promoter mutation, we considered the ctDNA as mutated if the variant was identified in at least one of the two technics. For TP53, the first analysis showed a rate of 14-3% mutations of TP53 in the control group without HCC, 6 9% in the inactive HCC and 29 - 1% in active HCC with a higher VAF in active HCC than in the other groups (data not shown). To be more stringent and decrease the risk of false positive mutations in TP 53 in patients without HCC, we considered the ctDNA as mutated if the variant had a VAF of 0.01 and, was identified, in the same patients, as a somatic mutation in the tumor, or in another plasma, or if the variant was a R249S, a well- known hot spot.

[0089] Statistical analysis

[0090] Continuous variables are described as median and interquartile range, and categorical variables as numbers and percentages. Non-parametric Mann-Whitney, Kruskal- Wallis, Wilcoxon and Jonckheere-Terpstra tests were used to compare quantitative variables and Chi-square, Fisher’s exact, McNemar, Cochrane Q and Cochran- Armitage tests for categorical variables. The correlation analysis was conducted using Spearman’s method. Follow-up of the patients was recorded until November 22, 2022. Overall Survival (OS) was calculated from the date of treatment to the date of death or censored at the time of the last follow-up. Recurrence and / or Progression Free Survival (RFS and PFS respectively) were calculated from the date of treatment to the date of recurrence / or disease progression or to the date of death. Kaplan-Meier curve and log-rank test were used to evaluate OS, RFS and PFS. Univariate and multivariate survival analyses were carried out using Cox model. Univariate and multivariate logistic regression analyses were performed to evaluate variables associated with treatment response and disease progression. A two-sided p value of <0-05 was considered statistically significant. All analyses were performed using R software version 4.1.1 (https : / / www.R-proj ect.org).

[0091] Results We collected 772 plasma samples from 173 patients with HCC, alongside 56 plasma samples from 56 control patients with chronic liver disease devoid of any HCC history (named “control” Table 1 for patient’s description). Patients diagnosed with HCC had a median age of 63 years, were predominantly male (81 -5%), and 73% had cirrhosis. At the time of tumor sampling, a majority of patients were Child-Pugh A (81 -7%), with a median AFP level of 11 pg / L and were classified BCLC 0 (22-8%), A (42-7%), B (20 7%) and C (13 8%). Among the 232 tumors, we identified mutations in one of these genes in 81.5% of the cases: TERT promoter (68-8%), TP 53 (30-6%), CTNNBI (27-2%), NFE2L2 (2- 1%) and PIK3CA (2-1%) (data not shown). The tumors were classified using the transcriptomic G1-G6 classification: 5 2% Gl, 1 9% G2, 16-7% G3, 47-6% G4, 16-7% G5, and 11 -9% G6 (data not shown). Using the 5-gene score, 40-8% of HCC were classified with poor molecular prognosis and 59-2% with good molecular prognosis (data not shown).

[0092] Initially, we analyzed 502 plasmas from patients with active HCC, 116 from patients inactive for HCC but with a prior history of HCC, and 56 from a control group without any current or prior HCC (data not shown). The concentration of cfDNA was significantly higher in “active HCC” (13-50 ng / mL) compared to “inactive HCC” (7 95 ng / mL, p<0.001, data not shown). Furthermore, 40 8% of ctDNA samples in the “active HCC” showed at least one mutation, while this was observed in 14-7% in the “inactive HCC” and 1 -8% for the “control “(p<0 • 001 ) (data not shown). Detection of two or more mutations in ctDNA was never observed in the “control”, found in 1 -7% of “inactive HCC”, and detected in 17- 1% of “active HCC”, with a significant correlation with cfDNA concentrations (data not shown).

[0093] Within “active HCC” samples compared with “inactive HCC”, we detected more frequent mutations in TERT promoter (27-5% vs 3-5%), TP53 <21-3% vs 8-8%), CTNNBI <13 1% vs 4-4%), PIK3CA (0-4% vs 0), and NFE2L2 <0-2% vs 0) (data not shown). Within the 17 mutated ctDNA collected from 11 patients presumed with “inactive HCC” after curative HCC treatment, 64% (seven patients) experienced tumor recurrence after a median follow-up of 11.5 months following the first occurrence of mutated ctDNA. In contrast, four patients remained free of tumor recurrence with a median follow-up of 23 -9 months (data not shown). One patient within the control group had a CTNNBI mutation with a very low VAF of 0-005, yet this patient remained HCC-free throughout 8.5 -month follow-up period. Next, we compared the mutational profile of 134 plasmas with the one of their corresponding tumor samples collected the same day. The mutational prevalence of the five genes in ctDNA of HCC patients was strongly associated with those observed in the corresponding tumor tissues (rho=0 975, p=0 005) (data not shown). Identical mutations were observed in the corresponding tumor in TERT promoter, CTNNB1, and TP53 in 67% to 85% of the cases, demonstrating the biological robustness of the ctDNA analyses (data not shown). An enrichment in G5 / G6 subgroups in tumors was noted in conjunction with CTNNB1 mutations in ctDNA (p=0-020) (data not shown), G3 subgroup in tumors with TP 53 mutations in ctDNA (p<0 • 001) (data not shown), and a poor prognosis 5 -gene score in tumors with TP 53 mutations (data not shown, p=0-002) and detection of two or three gene mutations in ctDNA (p=0-020) (data not shown).

[0094] Focusing on mutations in ctDNA, analyses considering BCLC staging (61 stage 0, 81 stage A, 51 stage B and 68 stage C) were performed taking into account the first plasma available in each stage for each patient (data not shown). We demonstrated a gradual elevation in cfDNA concentrations from BCLC 0 (9-50 ng / mL) to BCLC C (23-6 ng / mL) (p<0-001) (data not shown). The proportion of ctDNA samples harboring at least one mutation similarly exhibited a progressive increase across tumor stages: from BCLC 0 (8%) to BCLC C (71.4%, p value < 0 001) (data not shown).

[0095] Additionally, patients with mutations in two or three genes displayed higher cfDNA concentrations, elevated serum AFP levels, and more advanced tumor stages in contrast to those with only one mutated gene or lacking mutations in the ctDNA (data not shown). Interestingly, presence of ctDNA mutation could be used as a tumor biomarker in patients with normal serum AFP levels in all of BCLC 0 patients, 43 8% of BCLC A, 33-3% of BCLC B, and 27-3% of BCLC C cases (data not shown). The frequency of mutations in TERT promoter and TP53 exhibited a progressive increase from BCLC 0 to BCLC C, while CTNNB1 mutations reached a plateau in BCLC B and C in agreement with more frequent TP53 mutations observed in advanced HCC5(data not shown).

[0096] Next, we focused on 103 patients with “active HCC” treated by percutaneous ablation to evaluate the prognostic significance of ctDNA. The median OS was 45 -9 months, with a tumor recurrence rate of 77.7% within three years. Prior to ablation, mutations in ctDNA were detectable in at least one gene within 20-4% of patients; 7-8% exhibited mutations in two genes, and 1% in three genes. CfDNA concentration values above the median (p=0-006), presence of ctDNA mutations (p<0-001) (Figures 1A and 1C), TERT promoter mutations (p=0-003), and TP53 mutations (p<0 ■ 001 ) were associated with a reduced OS (Figure 2, Table 2). The risk of death increased with the number of detected mutated genes in ctDNA (Figure 1C). Multivariate analysis confirmed that the presence of mutations in ctDNA was independently linked to the risk of death (HR=2 6, 95% CI: 1 -5-4-7, p=0 001) (Table 2). A similar pattern was demonstrated for RFS (Figure IB and ID, Figure 2). The association between ctDNA mutations and an adverse prognosis was exclusively identified in patients with HCC exceeding 20 mm (Figure 3). Among the 71 patients that experienced tumor recurrence after ablation, those harboring a mutation in two or more genes in ctDNA at baseline displayed a higher frequency of tumor recurrence exceeding the Milan Criteria (75%), in contrast to those with a solitary mutation (25%) or no mutations (18 2%) (p=0 001) (Figure IE). Moreover, they exhibited an increased risk of extrahepatic recurrence (62-5%) in comparison to those with a single mutation (0%) or no mutations (12-2%, p<0 001) (Figure IF).

[0097] Then, we evaluated the variation in cfDNA concentration obtained prior (HO, included in "active HCC") and subsequent (H24) to locoregional treatment among 154 couples (data not shown). Among these, 130 were collected at percutaneous ablations, while 24 at trans-arterial embolization. HCC cases were categorized into BCLC 0 (32-5%), BCLC A (40-9%), BCLC B (20- 1%), and BCLC C (6-5%)( data not shown). We observed a significant rise in median cfDNA concentration at H24 (9’35 ng / mL at HO versus 31 -2 ng / mL at H24, p<0 001; data not shown). The cfDNA concentration detected at H24 notably increased in cases of BCLC B and C (data not shown) and was correlated with tumor size at H24 but not at HO (p=0 003; data not shown). Mutations detection in ctDNA increased at H24 (23-4% at HO versus 42-2% at H24, p<0 001) in particular in TERT promoter (13- 1% vs 35-3%), TP53 (13% vs 14%), and CTNNB1 (5- 1% vs 13-1%). Mutations in ctDNA at H24 displayed an increased frequency in relation to the BCLC stage (data not shown). We found identical mutations in the corresponding tumor samples in most of the cases, even though a majority of tumor mutations were still undetectable in the plasma even after the treatment (data not shown). In the subgroup of 103 patients treated by percutaneous ablation, patients harboring ctDNA mutations detected both at HO and at H24 had a reduced OS (p<0-001; Figure 4) and RFS (p=0-003; Figure 5) compared with other patients. Next, we conducted an analysis including 356 plasmas from 53 patients with advanced HCC, collected sequentially at initiation and during systemic treatments (19% BCLC B and 81% BCLC C). The treatment regimens encompassed atezolizumab / bevacizumab (35 patients), sorafenib (13), tepotinib (1), cabozantinib (2), tivantinib (1), and lenvatinib (1) (data not shown). Within the 53 plasmas, obtained just before the beginning of systemic therapy, we observed at least one mutation in 60 4% of the cases involving TERT promoter (45 3%), TP53 (28-3%), CTNNB1 (17-0%), NFE2L2 (0%), or PIK3CA (1 -9%) (data not shown). Among the 35 patients treated with atezolizumab / bevacizumab, the baseline cfDNA concentration or presence of mutations in ctDNA were not correlated with OS, PFS, and radiological response (data not shown). Among the 19 patients with detectable ctDNA mutations at baseline (data not shown), the persistence of mutations during the initial four cycles of atezolizumab / bevacizumab administration was notably linked to a higher rate of radiological progression (63 -6%) compared with cases where mutations disappeared (36-4%, p=0-019, data not shown). In the five patients classified as responders during the first imaging assessment, four demonstrated a decline in VAF of ctDNA mutations over subsequent evaluations, leading to their disappearance following the third cycle of atezolizumab / bevacizumab (data not shown). For the three patients classified as stable in the initial imaging evaluation, mutations became undetectable after the first cycle of atezolizumab / bevacizumab, with two of them showing a delayed radiological response at 6 months.

[0098] Lastly, in two patients we identified the appearance of CTNNB1 mutations in ctDNA during disease progression. A first BCLC-A HCC patient was treated by neoadjuvant nivolumab followed by curative percutaneous irreversible electroporation, followed by a one-year course of adjuvant nivolumab (NCT03630640 Nivolep trial, data not shown). A C228T TERT promoter and L130H TP 53 mutations were identified in the baseline tumor biopsy. No mutations were identified in the cfDNA at baseline or during neoadjuvant treatment, except for the detection of the C228T TERT promoter and L130H TP53 mutation at H24 post- percutaneous treatment. Throughout the adjuvant treatment, no mutations were observable in the ctDNA until the fourteenth injection, at which point a C228T TERT promoter mutation, an L130H TP53 mutation, and an S33A CTNNB1 mutation were detected three weeks prior to radiological progression and during atezolizumab / bevacizumab treatment. In the baseline tumor biopsy, TP53 and TERT promoter mutations were found to be clonal (with VAFs of 80% and 49% respectively), while the CTNNB1 mutation was subclonal (with a VAF of 3%). A second patient with advanced HCC, initially exhibiting an R249S TP53 mutation and MET amplification in the baseline tumor biopsy, experienced progression under sorafenib treatment. This patient received a second-line treatment by tepotinib, a MET inhibitor, being included in a clinical trial (NCT02115373)( data not shown). Notably, a complete radiological response was achieved, accompanied by the disappearance of the TP53 mutation in the ctDNA. Following 14 months of tepotinib treatment, radiological progression occurred, marked by the reappearance of the R249S TP53 mutation in the ctDNA, along with a newly detected G34E CTNNB1 mutation that was not initially identified in the baseline tumor biopsy.

[0099] Discussion:

[0100] In this study, we investigated the presence of ctDNA across different HCC stages and, its role as a prognostic tool in early-stage HCC as well as its utility as a dynamic biomarker to monitor treatment responses. To accomplish this, we utilized a panel of 5 genes to search for mutations in ctDNA. We specifically selected TERT promoter, TP53, and CTNNB1 genes, as they collectively accounted for 81.5% of HCC cases in our series showing at least one mutation in these genes. NFE2L2 and PIK3CA were also included due to presence of recurrent hot spot mutations, making them easily analyzable. Moreover, to improve the detection of low VAF TERT promoter mutations, we combined two detection methods: ddPCR and MiSeq sequencing. During the first analysis, we observed a high frequency of TP53 mutations in the cell free DNA of control patients without HCC and in patients with a history of HCC but without “active HCC” at sampling. This could be explained by ultradeep sequencing of a large gene using MiSeq technology and / or the presence of clonal hematopoiesis16. It also highlights the risk of false positive results when conducting ultradeep sequencing of a large gene in search of very low VAF variants. To ensure rigorous analysis, we decided to include in the study only TP53 mutations that were identified in the tumor or in another sample, as well as the R249S TP53 mutation, which is a known hot spot mutation.

[0101] This approach allowed mutation detection in the ctDNA of 40-2% of patients with active HCC, with an increased frequency of detection in intermediate and advanced stages. One of the strengths of our study is the analysis of corresponding tumors, which revealed good concordance between the mutations observed in the tumor and those found in the ctDNA. However, we observed a lower percentage of mutations in ctDNA in BCLC 0 (8%) and BCLC A (21%), likely due to the small tumor burden and minimal release of tumor DNA into the blood. Since the target of HCC screening is small tumors classified as BCLC 0 and A, our findings suggests that the current method of detecting ctDNA may not be sensitive enough in these cases. New liquid biopsy methods, such as methylation biomarkers or more sensitive approaches to detect circulating tumor DNA, are needed17. Interestingly, among patients with mutations detected in cfDNA, approximately one-third exhibited normal serum AFP levels, implying that ctDNA could provide additional value as a biomarker. However, the detection of mutations in the ctDNA should be interpreted in the clinical context. Notably, some patients with a history of HCC and a detectable mutation in ctDNA collected during “inactive HCC” did not experience tumor recurrence during follow-up.

[0102] Previous studies have reported a potential prognostic role of ctDNA in HCC patients treated by resection but none of them have focused on percutaneous ablation18'19. In our study, we demonstrated that the detection of mutations in the ctDNA before percutaneous treatment is associated with a poor prognosis. Interestingly, the detection of mutations in two or three genes in the ctDNA was linked to recurrence outside the Milan criteria, suggesting that it may be a surrogate marker of tumor aggressiveness. However, this prognostic value was only observed in tumors larger than 2 cm and, consequently, ctDNA could be used as a prognostic biomarker in patients with large tumors in order to stratify clinical trial testing neoadjuvant and adjuvant systemic treatments. We also assessed the early dynamics of ctDNA before (HO) and after (H24) locoregional treatment, revealing increased concentrations and ctDNA mutations are detectable at H24. This suggests that cell necrosis induce by loco-regional treatment enhance the detection of mutations in the blood.

[0103] Since we were able to detect mutations in the ctDNA in 60.4% of patients treated with systemic therapies, we conducted a pilot analysis to monitor the variation of ctDNA in patients treated with atezolizumab / bevacizumab. The persistence of mutations in the ctDNA was significantly associated with radiological progression, whereas the disappearance of mutations was linked to response. However, to validate these findings and conduct a more detailed analysis of the fluctuation of the VAF during treatment, a larger cohort will be required. Finally, in two cases, we demonstrated that mutations in the ctDNA appeared during tumor progression, with additional CTNNB1 mutation that was not initially detectable in the blood but later identified as sub-clonal mutations (VAF of 3%) in the tumor in one patient. These findings suggest that the emergence of new subclones during tumor progression could be detected in the ctDNA, offering the potential for ctDNA to serve as a liquid biopsy tool for longitudinal monitoring of tumor heterogeneity20.

[0104] In conclusion, we highlight that ctDNA provides dynamic information about tumor biology and may serve as a non-invasive tool with the potential to predict prognosis and monitor treatment response in patients with HCC. Its role in guiding the clinical management needs to be adapted according to the tumor stages and to the types of treatments received.

[0105] TABLES:

[0106] Table 1. Characteristics of patients (n=229) and active plasma samples (n=502) included in the study.

[0107] *At sampling. Samples collected at the time of the diagnostic investigation leading to HCC diagnosis at the patient’s clinico-radiological assessments. AFP: alpha-foetoprotein. BCLC: Barcelona Clinic Liver Cancer. NA: not available.#Child Pugh score was reported only in cirrhotic patients. Table 2. Univariate and multivariate Cox regression analysis of variables associated with overall survival in 103 patients at first percutaneous ablation.

[0108] AFP: alpha-foetoprotein. *Child Pugh score was reported only in cirrhotic patients.AHigh means having a cell free DNA concentration (ng / mL) value above the median of the population included in the analysis. Significant p-values are reported in bold and italic. We didn’t include cell free DNA concentration, TERT promoter, CTNNB1 and TP53 mutations in multivariate analysis to avoid collinearity with the variable “mutation”.

[0109] REFERENCES:

[0110] Throughout this application, various references describe the state of the art to which this invention pertains. The disclosures of these references are hereby incorporated by reference into the present disclosure.

[0111] 1. Cescon, D. W., Bratman, S. V., Chan, S. M. & Siu, L. L. Circulating tumor DNA and liquid biopsy in oncology. Nat Cancer 1, 276-290 (2020).

[0112] 2. Wan, J. C. M. et al. Liquid biopsies come of age: towards implementation of circulating tumour DNA. Nat Rev Cancer 17, 223-238 (2017).

[0113] 3. Abbosh, C. et al. Tracking early lung cancer metastatic dissemination in TRACERx using ctDNA. Nature 616, 553-562 (2023).

[0114] 4. Dhanasekaran, R., Nault, J.-C., Roberts, L. R. & Zucman-Rossi, J. Genomic Medicine and Implications for Hepatocellular Carcinoma Prevention and Therapy. Gastroenterology 156, 492-509 (2019).

[0115] 5. Nault, J. et al. Clinical Impact of Genomic Diversity From Early to Advanced Hepatocellular Carcinoma. Hepatology 71, 164-182 (2020).

[0116] 6. Nault, J. C. et al. High frequency of telomerase reverse-transcriptase promoter somatic mutations in hepatocellular carcinoma and preneoplastic lesions. Nat Commun 4, 2218 (2013).

[0117] 7. Muraoka, M. et al. Usefulness of Cell-Free Human Telomerase Reverse Transcriptase Mutant DNA Quantification in Blood for Predicting Hepatocellular Carcinoma Treatment Efficacy. Hepatology Communications 5, 1927-1938 (2021).

[0118] 8. von Felden, J. et al. Mutations in circulating tumor DNA predict primary resistance to systemic therapies in advanced hepatocellular carcinoma. Oncogene 40, 140-151 (2021).

[0119] 9. Howell, J. et al. Identification of mutations in circulating cell-free tumour DNA as a biomarker in hepatocellular carcinoma. European Journal of Cancer 116, 56-66 (2019).

[0120] 10. Lee, H. W. et al. Applications of molecular barcode sequencing for the detection of low - frequency variants in circulating tumour DNA from hepatocellular carcinoma. Liver International 42, 2317-2326 (2022). 11. Von Felden, J., Garcia-Lezana, T., Schulze, K., Losie, B. & Villanueva, A. Liquid biopsy in the clinical management of hepatocellular carcinoma. Gut 69, 2025-2034 (2020).

[0121] 12. Singal, A. G. et al. International Liver Cancer Association (ILCA) white paper on hepatocellular carcinoma risk stratification and surveillance. Journal of Hepatology 79, 226- 239 (2023).

[0122] 13. Lencioni, R. & Llovet, J. Modified RECIST (mRECIST) Assessment for Hepatocellular Carcinoma. Semin Liver Dis 30, 052-060 (2010).

[0123] 14. Nault, J. et al. A Hepatocellular Carcinoma 5-Gene Score Associated With Survival of Patients After Liver Resection. Gastroenterology 145, 176-187 (2013).

[0124] 15. Boyault, S. et al. Transcriptome classification of HCC is related to gene alterations and to new therapeutic targets. Hepatology 45, 42-52 (2007).

[0125] 16. Yaung, S. J. et al. Clonal Hematopoiesis in Late-Stage Non-Small-Cell Lung Cancer and Its Impact on Targeted Panel Next-Generation Sequencing. JCO Precision Oncology 1271— 1279 (2020) doi:10.1200 / P0.20.00046.

[0126] 17. Xu, R. et al. Circulating tumour DNA methylation markers for diagnosis and prognosis of hepatocellular carcinoma. Nature Mater 16, 1155-1161 (2017).

[0127] 18. Tokuhisa, Y. et al. Circulating cell-free DNA as a predictive marker for distant metastasis of hepatitis C virus-related hepatocellular carcinoma. Br J Cancer 97, 1399-1403 (2007).

[0128] 19. Cai, Z. et al. Comprehensive Liquid Profiling of Circulating Tumor DNA and Protein Biomarkers in Long-Term Follow-Up Patients with Hepatocellular Carcinoma. Clinical Cancer Research 25, 5284-5294 (2019).

[0129] 20. Abbosh, C. et al. Tracking early lung cancer metastatic dissemination in TRACERx using ctDNA. Nature 616, 553-562 (2023).

Claims

CLAIMS:

1. A method of predicting the risk of recurrence and / or death of a patient suffering from a hepatocellular carcinoma (HCC) comprising the steps of i) detecting at least one mutation in TERT, TP53, and CTNNB1 (i.e. determining whether the TERT, TP53, and CTNNB1 genes are mutated) in a sample obtained from the patient, ii) determining the number of mutated genes wherein said number correlates with the risk of recurrence and / or death.

2. The method of claim 1 that comprises the steps of i) detecting at least one mutation in TERT, TP53, CTNNB1, NFE2L2 and PIK3CA in a sample obtained from the patient, ii) determining the number of mutated genes wherein said number correlates with the risk of recurrence and / or death.

3. The method of claim 1 or 2 for predicting the survival time of the HCC patient, in particular, for predicting the duration of the overall survival (OS), progression-free survival (PFS) and / or the disease-free survival (DFS) of the HCC patient.

4. The method according to any one of claims 1 to 3 wherein the patient has a tumor that exceeds 20mm.

5. The method according to any one of claims 1 to 4 wherein the patient is at BCLC stage B, Stage C or Stage D.

6. The method according to any one of claims 1 to 5 wherein the sample is a blood sample7. The method according to any one of claims 1 to 5 wherein the sample is a ctDNA sample.

8. The method according to any one of claims 1 to 7 wherein the risk of recurrence and / or death increases with the detection of mutated genes.

9. The method of claim 8 wherein the more genes are mutated, the higher is the risk of recurrence and / or death and similarly and the more genes are mutated, the shorter will be the survival time of the patient.

10. The method according to any one of claims 1 to 9 for discriminating responder from non-responder.

11. The method according to claim 10 for determining whether the patient achieves a response after a preoperative therapy, a locoregional therapy and / or a postoperative therapy.

12. The method of claim 11 wherein the increase of mutated genes or the appearance of new mutated genes along the therapy indicates that the patient does not achieve a response to the therapy, and conversely a decrease of the number of mutated genes or even more the disappearance of mutated gene indicates that the patient achieves a response to the therapy.

13. The method according to any one of claims 10 to 12 that comprises the steps of a) detecting the number of mutated genes in a first sample obtained from the patient before the therapy, b) detecting the number of mutated genes in a second sample obtained from the patient after the therapy and c) comparing the numbers of mutated genes between the two samples wherein when the number of mutated genes in the second sample is equal or higher than the number of mutated genes in the first sample, it is concluded that the patient does not achieve a response to the therapy and whereas when the number of mutated genes in the second sample is lower than the number of mutated genes in the first sample of becomes null, it is concluded that the patient achieves a response to the therapy.

14. The method according to any one of claims 1 to 13 wherein the patient is administered with a TKI selected from sorafenib, tepotinib, cabozantinib, tivantinib, and lenvatinib.

15. The method according to any one of claims 1 to 13 wherein the patient is administered with a combination of an anti-VEGF agent with an anti-PD-Ll antibody more particularly a combination of bevacizumab and atezolizumab.

Images