Machine learning classifiers for diagnosing and / or predicting heart allograft acute cellular rejection

Machine learning classifiers using gene expression profiling of FFPE endomyocardial biopsies address the limitations of current diagnostic methods by providing precise and reliable identification and prediction of acute cellular rejection, improving clinical management of heart allografts.

WO2026238952A1PCT designated stage Publication Date: 2026-11-19INST NAT DE LA SANTE & DE LA RECHERCHE MEDICALE (INSERM) +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/EP2026/064188
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-16
Filing Date
2026-05-15
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

Current diagnostic methods for heart allograft acute cellular rejection, such as endomyocardial biopsies, suffer from inter-observer variability, sample bias, and inadequate capture of disease severity, limiting precise and reliable diagnosis.

Method used

Development of machine learning classifiers using gene expression profiling of formalin-fixed, paraffin-embedded endomyocardial biopsies, leveraging the Banff Human Organ Transplant Panel, to accurately identify and predict acute cellular rejection through quantification of specific gene expressions and implementation of supervised molecular classifiers.

Benefits of technology

The classifiers provide accurate identification and prediction of acute cellular rejection with high ROC-AUC scores, enhancing clinical applicability and refining diagnosis, enabling personalized medicine approaches.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure IMGF000033_0001
    Figure IMGF000033_0001
  • Figure IMGF000034_0001_TABLE
    Figure IMGF000034_0001_TABLE
  • Figure IMGF000034_0002_TABLE
    Figure IMGF000034_0002_TABLE
Patent Text Reader

Abstract

Tissular gene expression profiling applicable to formalin-fixed, paraffin-embedded (FFPE) endomyocardial biopsies (EMBs) may refine the diagnosis of cardiac rejection while being easily implemented in clinical practice. The inventors aimed to develop and validate the first FFPE-based molecular diagnostic system dedicated to heart transplant rejection. An international study was conducted (NCT06436027), establishing a deeply phenotyped cohort of heart transplant recipients. EMBs were graded according to international classifications, and gene expression was analyzed on FFPE-EMBs using the Banff Human Organ Transplant Panel. Molecular classifiers for acute cellular rejection (ACR) were developed, with discrimination and calibration assessed in internal and external validation cohorts. A total of 671 biopsies were included, with 591 in the main cohort (ACR: n=289, matched non-rejection: n=114); this was split into a derivation set (n=475) and an internal validation set (n=116). The external validation cohort comprised 80 biopsies (ACR: n=32, non-rejection: n=28). ACR was characterized by transcripts related to T-cell receptor signaling, CD3 receptor activation, and CD28 signaling. Molecular ACR models accurately identified rejection in the validation cohorts (internal: ROC- AUC: ACR=0.849; external: ROC-AUC: ACR=0.815) and were strongly associated with pathologic severity. Calibration was adequate. An automated report was developed to enhance the clinical applicability of these classifiers. In conclusion, a novel FFPE-based molecular diagnostic system accurately identified cardiac allograft rejection. This tool is readily applicable in clinical practice as a companion to pathology and has the potential to refine the diagnosis of rejection.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] MACHINE LEARNING CLASSIFIERS FOR DIAGNOSING AND / OR PREDICTING HEART ALLOGRAFT ACUTE CELLULAR REJECTION

[0002] FIELD OF THE INVENTION:

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

[0004] BACKGROUND OF THE INVENTION:

[0005] Heart transplantation remains the most effective therapeutic option for patients with end-stage heart failure that is refractory to optimal medical therapy.1,2Globally, more than 6,000 heart transplants are performed annually, with a median post-transplant survival now exceeding 12.5 years.3Despite significant advancements in immunosuppression and transplant care, acute and chronic rejection-induced allograft injuries continue to be major causes of morbidity and mortality, ultimately limiting recipients’ life expectancy.4 6Improving the overall management of rejection remains an unmet medical need.7 9A crucial first step in addressing this challenge is achieving a precise diagnosis of rejection, which is essential for guiding patient care and optimizing management.

[0006] Currently, the diagnosis of cardiac rejection still relies on the pathological assessment of endomyocardial biopsies (EMBs), which involves identifying and grading cellular infiltrates and myocardial damage as assessed by international working formulations.10,11While significant advances have been made in standardizing rejection diagnosis, pathology remains an imperfect gold standard because of certain key limitations, including inter-observer variability,12sample bias,13and the use of qualitative or semi -quantitative scales that oversimplify complex phenotypes.14Additionally, crucial information such as disease severity, the degree of myocardial injury, and the stage of progression is inadequately captured by current diagnostic frameworks. These limitations pose major barriers to achieving a precise and reliable diagnosis of rejection. In this context, gene expression profiling of fresh myocardial samples using a whole-transcriptome approach has emerged as a potential objective companion tool to pathology.15 17However, several critical drawbacks have hindered their routine clinical application. These include extra-core sampling (with the inherent related complications and bias), the absence of pathologist oversight to exclude nonspecific lesions, low reproducibility, technical and analytical complexity, high costs, turnaround times that are poorly suited to clinical settings, and the lack of access to archival samples for model development and

[0007] validation (previous molecular classifiers for cardiac allograft rejection have been developed using kidney biopsies rather than EMBs).9,16

[0008] Recent technologies, such as the Nanostring nCounter® system, offer a promising solution to this limitation by enabling analysis of the same tissue used for histology.18,19Targeted molecular profiling of formalin-fixed, paraffin-embedded (FFPE) EMBs could facilitate the integration of molecular diagnostics into routine clinical practice. Recently, we demonstrated that the Banff Human Organ Transplant (B-HOT) Panel, a consortium-approved panel comprising 770 genes and developed by the Banff Molecular Diagnostics Working Group,20effectively captures key molecular patterns of antibody-mediated rejection (ACR) in heart allograft biopsies. This panel may serve as a viable alternative to whole-transcriptome analysis.21

[0009] SUMMARY OF THE INVENTION:

[0010] The present invention is defined by the claims. In particular, the present invention relates to machine learning classifiers for diagnosing and / or predicting heart allograft acute cellular rejection.

[0011] DETAILED DESCRIPTION OF THE INVENTION:

[0012] Tissular gene expression profiling applicable to formalin-fixed, paraffin-embedded (FFPE) endomyocardial biopsies (EMBs) may refine the diagnosis of cardiac rejection while being easily implemented in clinical practice. The inventors aimed to develop and validate the first FFPE-based molecular diagnostic system dedicated to heart transplant rejection. An international study was conducted (NCT06436027), establishing a deeply phenotyped cohort of heart transplant recipients. EMBs were graded according to international classifications, and gene expression was analyzed on FFPE-EMBs using the Banff Human Organ Transplant Panel. Molecular classifiers for acute cellular rejection (ACR) were developed, with discrimination and calibration assessed in internal and external validation cohorts. A total of 671 biopsies were included, with 591 in the main cohort (ACR: n=289, matched non-rejection: n=114); this was split into a derivation set (n=475) and an internal validation set (n=116). The external validation cohort comprised 80 biopsies (ACR: n=32, non-rejection: n=28). ACR was characterized by transcripts related to T-cell receptor signaling, CD3 receptor activation, and CD28 signaling. Molecular ACR models accurately identified rejection in the validation cohorts (internal: ROC-AUC: ACR=0.849; external: ROC-AUC: ACR=0.815) and were strongly associated with

[0013] pathologic severity. Calibration was adequate. An automated report was developed to enhance the clinical applicability of these classifiers. In conclusion, a novel FFPE-based molecular diagnostic system accurately identified cardiac allograft rejection. This tool is readily applicable in clinical practice as a companion to pathology and has the potential to refine the diagnosis of rejection.

[0014] Main definitions:

[0015] As used herein, the term “heart transplantation” refers to the process of taking a heart (i.e., “graft”) from one individual and placing it into a different individual. The individual who provides the heart for the transplant is called the “donor” and the individual who receives the heart is called the “recipient.” As used herein, the term “allograft” refers to a tissue or organ that is transplanted from one individual to another within the same species.

[0016] As used herein, the term "acute cellular rejection” or “ACR” has its general meaning in the art and refers to the pathological process characterized by infiltration of the interstitium by T cells and macrophages, intense ZFNy and TGFp effects, and epithelial deterioration. The term is also known as “T cell-mediated rejection” or “TCMR”. The three primary sites of acute ACR in transplanted kidneys are indeed the tubular epithelial cells, interstitium, and the vascular endothelial cells. ACR can be acute ACR or chronic acitve ACR. Banff classification recognizes three gradings: tubulitis (the t score), interstitial inflammation (the i score) and endarteritis (the v score).

[0017] As used herein, the term “risk” in the context of the present invention, relates to the probability that an event will occur over a specific time period, as in the conversion to allograft loss, and can mean a subject's “absolute” risk or “relative” risk. Absolute risk can be measured with reference to either actual observation post-measurement for the relevant time cohort, or with reference to index values developed from statistically valid historical cohorts that have been followed for the relevant time period. Relative risk refers to the ratio of absolute risks of a subject compared either to the absolute risks of low risk cohorts or an average population risk, which can vary by how clinical risk factors are assessed. Odds ratios, the proportion of positive events to negative events for a given test result, are also commonly used (odds are according to the formula p / (l p) where p is the probability of event and (1— p) is the probability of no event) to no conversion. Accordingly, the expression “predicting whether a heart transplant recipient is at risk of ACR” in the context of the present invention encompasses making a

[0018] prediction of the probability, odds, or likelihood that allograft loss mediated by ACR may occur. Accordingly, the expression “predicting whether a heart transplant recipient has ACR” or “diagnosing whether a heart transplant recipient is afflicted by ACR” in the context of the present invention encompasses determining the presence of ACR, and thus the likelihood of allograft loss mediated by ACR. The methods of the present invention may be used to make continuous or categorical measurements of the risk of conversion to ACR. According to the present invention, the prediction is a dynamic prediction. As used herein, the term "dynamic prediction" refers to providing an assessment of probability or likelihood for a subject to develop ACR, which may change over time. The method of the present invention is particularly suitable to predict the risk of ACR at 3, 5 and / or 7 years from the date of prediction. As used herein, a “high risk of ACR” means that allograft loss mediated by ACR may occur more probably than not. As used herein, a “low risk of ACR” means that it is more probable that allograft loss mediated by ACR may not occur.

[0019] 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”.

[0020] As used herein, the term "machine learning classifier" refers to a computational model that analyzes gene expression data to predict biological states, such as disease presence or immune responses. These classifiers leverage complex algorithms and statistical methods to identify patterns within the data that correlate with specific biological outcomes. The use of machine learning classifiers facilitates the development of personalized medicine approaches, enabling more accurate predictions of disease progression and responses to therapy.

[0021] As used herein, the term “supervised molecular classifier” has its general meaning in the art and refers to a type of machine learning algorithm that is trained with labeled data. In this context, it involves using pre-determined expression levels of genes as input variables and applying them to a multivariate logistic regression model to predict the risk of acute cellular rejection (ACR) in heart allograft recipients. This classifier is designed to optimize its performance based on these labeled examples, thereby enhancing its predictive accuracy and reliability.

[0022] As used herein, the term “gene” has its general meaning in the art and refers to a sequence of nucleotides in DNA or RNA that encodes the synthesis of a gene product, either RNA or protein, which performs a specific function in the organism. Genes are the fundamental units of heredity and are transferred from parent to offspring during reproduction. They are responsible for the inherited characteristics and functions of living organisms, and their expression levels can be quantitatively assessed to provide insights into various biological processes and disease states. In the present specification, the name of each of the genes of interest refers to the internationally recognised name of the corresponding gene, as found in internationally recognised gene sequences and protein sequences databases, in particular in the database from the HUGO Gene Nomenclature Committee, that is available notably at the following Internet address: http: / / www.gene.ucl.ac.uk / nomenclature / index.html. In the present specification, the name of each of the various biological markers of interest may also refer to the internationally recognised name of the corresponding gene, as found in the internationally recognised gene sequences and protein sequences databases ENTRE ID, Genbank, TrEMBL or ENSEMBL. Through these internationally recognised sequence databases, the nucleic acid sequences corresponding to each of the gene of interest described herein may be retrieved by the one skilled in the art.

[0023] As used herein, the term “expression level” refers to the quantity of mRNA produced from a gene, which serves as an indicator of gene activity within a sample.

[0024] As used herein, the term “sample” refers to any biological sample obtained for evaluation in vitro. In the context of transplantation, this typically includes tissue samples. The term “tissue sample” encompasses sections of tissues such as biopsy or autopsy samples and frozen sections taken for histological analysis.

[0025] Methods of the present invention:

[0026] The first object of the present invention relates to a method of predicting (“prediction method”) whether a heart transplant recipient has or is at risk of having acute cellular rejection (ACR) comprising:

[0027] a) quantifying the expression levels of one or more genes selected from the group consisting of CCL19, LTB, PTPN7, BTLA, GZMK, JAK3, PHEX; CD28, CCL18, LEF1, TCF7, CD5, CD8A, CD27, PECAM1, LCK, MMP9, CD3E, LAG3, and MS4A1 in a sample obtained from the recipient;

[0028] b) implementing an algorithm on data comprising the quantified plurality of gene expression levels so as to obtain an algorithm output; and

[0029] c) determining the probability of ACR from the algorithm output of step b).

[0030] In other words, the invention also relates to a method for diagnosing (especially early diagnosis) whether a heart transplant recipient is afflicted by acute cellular rejection (ACR) comprising:

[0031] quantifying the expression levels of one or more genes selected from the group consisting of CCL19, LTB, PTPN7, BTLA, GZMK, JAK3, PHEX; CD28, CCL18, LEF1, TCF7, CD5, CD8A, CD27, PECAM1, LCK, MMP9, CD3E, LAG3, and MS4A1 in a sample obtained from the recipient;

[0032] implementing an algorithm on data comprising the quantified plurality of gene expression levels so as to obtain an algorithm output; and

[0033] determining the presence of ACR from the algorithm output of step b).

[0034] In some embodiments, regarding all the methods of the invention, the expression level(s) of 1, 2, 3, 4, 5, 6, 7, 8, 9 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 genes among CCL19, LTB, PTPN7, BTLA, GZMK, JAK3, PHEX; CD28, CCL18, LEF1, TCF7, CD5, CD8A, CD27, PECAM1, LCK, MMP9, CD3E, LAG3, and MS4A1 is / are determined in the sample.

[0035] In some embodiments, regarding all the methods of the invention, the expression levels of CCL19, LTB, PTPN7, BTLA, GZMK, JAK3, PHEX; CD28, CCL18, LEF1, TCF7, CD5, CD8A, CD27, PECAM1, LCK, MMP9, CD3E, LAG3, and MS4A1 are determined in the sample.

[0036] In some embodiments, regarding all the methods of the invention, the expression level of further one or more genes selected from the group consisting of ARG1, B2M, BATF, C1QA, CAV1, CCL3 / L1, CCL5, CD163, CD3D, CD45R0, CD45RB, CD48, CD7, CD8B, CD96, CDH5, CLEC4C, COL1A1, COL3A1, CTLA4, CX3CL1, CX3CR1, CXCL10, CXCL11, EZH2, FAS, FCGR3A / B, FN1, FOXP3, GATA3, GZMA, HLA-DRA, HMOX1, HYAL2, ICAM1, IFI27, IFI44, IFITM3, IGHG1, IGHG2, IGHG3, IGHG4, IGKC, IGLC1, IKZF1, IL1R2, IL2RG, IL6, IL7R, IRF1, ISG15, ITGB2, KLF4, LILRB1, LOX, MX1, MYB, NFKB1, NOS3, OASL, PDPN, PLA1A, PLAT, PTGER4, PTPRC, PTX3, ROBO4, RORA, S100A8, S100A9, SELE, SIRPG, SLAMF7, SMAD3, SOCS3, SPP1, STAT1, TM4SF18, TRAT1, and ZAP70 is determined in the sample.

[0037] In some embodiments, regarding all the methods of the invention, the expression levels of ARG1, B2M, BATF, BTLA, C1QA, CAV1, CCL18, CCL19, CCL3 / L1, CCL5, CD163, CD27, CD28, CD3D, CD3E, CD45R0, CD45RB, CD48, CD5, CD7, CD8A, CD8B, CD96, CDH5, CLEC4C, COL1A1, COL3A1, CTLA4, CX3CL1, CX3CR1, CXCL10, CXCL11, EZH2, FAS, FCGR3A / B, FN1, FOXP3, GATA3, GZMA, GZMK, HLA-DRA, HMOX1, HYAL2, ICAM1, IFI27, IFI44, IFITM3, IGHG1, IGHG2, IGHG3, IGHG4, IGKC, IGLC1, IKZF1, IL1R2, IL2RG, IL6, IL7R, IRF1, ISG15, ITGB2, JAK3, KLF4, LAG3, LCK, LEF1, LILRB1, LOX, LTB, MMP9, MS4A1, MX1, MYB, NFKB1, NOS3, OASL, PDPN, PECAM1, PHEX, PLA1A, PLAT, PTGER4, PTPN7, PTPRC, PTX3, ROBO4, RORA, S100A8, S100A9, SELE, SIRPG, SLAMF7, SMAD3, SOCS3, SPP1, STAT1, TCF7, TM4SF18, TRAT1, and ZAP70 are determined in the sample.

[0038] Typically, the methods (prediction and diagnosis) of the invention is performed after the recipient has been transplanted.

[0039] According to the present invention the sample is previously obtained from the transplanted organ, preferably the transplanted heart. According to the present invention the sample has been obtained from the transplanted organ, preferably the transplanted heart. Thus, in some embodiments, the tissue sample may resulted from a biopsy performed on the transplanted organ to assess its health and function. In preferred embodiment, regarding all the methods of the invention, the transplanted organ is a transplanted heart.

[0040] Quantifying the expression levels:

[0041] Methods for determining (quantifying) the expression level of a gene product such as nucleic acid (e.g. RNA) are also well known in the art. Conventional methods typically involve polymerase chain reaction (PCR). For instance, U. S. Pat. Nos. 4,683,202, 4,683,195, 4,800,159, and 4,965,188 disclose conventional PCR techniques. PCR typically employs two oligonucleotide primers that bind to a selected target nucleic acid sequence. Primers useful in the present invention include oligonucleotides capable of acting as a point of initiation of nucleic acid synthesis within the target nucleic acid sequence. A primer can be purified from a restriction digest by conventional methods, or it can be produced synthetically. PCR involves use of a thermostable polymerase. The term “thermostable polymerase” refers to a polymerase enzyme that is heat stable, i.e., the enzyme catalyzes the formation of primer extension productscomplementary to a template and does not irreversibly denature when subjected to the elevated temperatures for the time necessary to effect denaturation of double-stranded template nucleic acids. Thermostable polymerases have been isolated from Thermus fiavus, T. ruber, T. thermophilus, T. aquaticus, T. lacteus, T. rubens, Bacillus stear other mophilus, and Methanothermus fervidus. Nonetheless, polymerases that are not thermostable also can be employed in PCR assays provided the enzyme is replenished. Typically, the polymerase is a Taq polymerase (i.e. Thermus aquaticus polymerase). Quantitative PCR is typically carried out in a thermal cycler with the capacity to illuminate each sample with a beam of light of a specified wavelength and detect the fluorescence emitted by the excited fluorophore. The thermal cycler is also able to rapidly heat and chill samples, thereby taking advantage of the physicochemical properties of the nucleic acids and thermal polymerase. In order to detect and measure the amount of amplicon (i.e. amplified target nucleic acid sequence) in the sample, a measurable signal has to be generated, which is proportional to the amount of amplified product. All current detection systems use fluorescent technologies. Some of them are non-specific techniques, and consequently only allow the detection of one target at a time. Alternatively, specific detection chemistries can distinguish between non- specific amplification and target amplification. These specific techniques can be used to multiplex the assay, i.e. detecting several different targets in the same assay. For example, SYBR® Green I probes, High Resolution Melting probes, TaqMan® probes, LNA® probes and Molecular Beacon probes can be suitable. TaqMan® probes are the most widely used type of probes. They were developed by Roche (Basel, Switzerland) and ABI (Foster City, USA) from an assay that originally used a radio-labelled probe (Holland et al. 1991), which consisted of a single-stranded probe sequence that was complementary to one of the strands of the amplicon. A fluorophore is attached to the 5’ end of the probe and a quencher to the 3’ end. The fluorophore is excited by the machine and passes its energy, via FRET (Fluorescence Resonance Energy Transfer) to the quencher. Traditionally, the FRET pair has been conjugated to FAM as the fluorophore and TACRA as the quencher. In a well-designed probe, FAM does not fluoresce as it passes its energy onto TACRA. As TACRA fluorescence is detected at a different wavelength to FAM, the background level of FAM is low. The probe binds to the amplicon during each annealing step of the PCR. When the Taq polymerase extends from the primer which is bound to the amplicon, it displaces the 5’ end of the probe, which is then degraded by the 5 ’-3’ exonuclease activity of the Taq polymerase. Cleavage continues until the remaining probe melts off the amplicon. This process releases the fluorophore and quencher into solution, spatially separatingthem (compared to when they were held together by the probe). This leads to an irreversible increase in fluorescence from the FAM and a decrease in the TACRA.

[0042] In some embodiments, the expression level is determined by RNA-seq. As used, the term " RNA-Seq" or "transcriptome sequencing" refers to sequencing performed on RNA (or cDNA) instead of DNA, where typically, the primary goal is to measure expression levels, detect fusion transcripts, alternative splicing, and other genomic alterations that can be better assessed from RNA. RNA-Seq typically includes whole transcriptome sequencing. As used herein, the term “whole transcriptome sequencing” refers to the use of high throughput sequencing technologies to sequence the entire transcriptome in order to get information about a sample's RNA content. Whole transcriptome sequencing can be done with a variety of platforms for example, the Genome Analyzer (Illumina, Inc., San Diego, Calif.) and the SOLiD™ Sequencing System (Life Technologies, Carlsbad, Calif.), However, any platform useful for whole transcriptome sequencing may be used. Typically, the RNA is extracted, and ribosomal RNA may be deleted as described in U. S. Pub, No. 2011 / 0111409. cDNA sequencing libraries may be prepared that are directional and single or paired-end using commercially available kits such as the ScriptSeq™ M mRNA-Seq Library Preparation Kit (Epicenter Biotechnologies, Madison, Wis.). The libraries may also be barcoded for multiplex sequencing using commercially available barcode primers such as the RNA-Seq Barcode Primers from Epicenter Biotechnologies (Madison, Wis.). PCR is then carried out to generate the second strand of cDNA to incorporate the barcodes and to amplify the libraries. After the libraries are quantified, the sequencing libraries may be sequenced. Nucleic acid sequencing technologies are suitable methods for expression analysis. The principle underlying these methods is that the number of times a (DNA sequence is detected in a sample is directly related to the relative RNA levels corresponding to that sequence. These methods are sometimes referred to by the term Digital Gene Expression (DOE) to reflect the discrete numeric property of the resulting data. Early methods applying this principle were Serial Analysis of Gene Expression (SAGE) and Massively Parallel Signature Sequencing (MPSS). See, e.g., S. Brenner, et al., Nature Biotechnology 18(6):630-634 (2000). Typically RNA-seq uses Next Generation Sequencing or NGS. As used herein, the term " Next Generation Sequencing" (NGS) refers to a relatively new sequencing technique as compared to the traditional Sanger sequencing technique. For review, see Shendure et al., Nature Biotech., 26(10): 1135-45 (2008), which is hereby incorporated by reference into this disclosure. For purpose of this disclosure, NGS may include cyclic array sequencing, microelectrophoretic sequencing, sequencing by hybridization, among others. Byway of example, in a typical NGS using cyclic-array methods, genomic DNA or cDNA library is first prepared, and common adaptors may then be ligated to the fragmented genomic DNA or cDNA. Different protocols may be used to generate jumping libraries of mate-paired tags with controllable distance distribution. An array of millions of spatially immobilized PCR colonies or "polonies" is generated with each polonies consisting of many copies of a single shotgun library fragment. Because the polonies are tethered to a planar array, a single microliter- scale reagent volume can be applied to manipulate the array features in parallel, for example, for primer hybridization or for enzymatic extension reactions. Imaging-based detection of fluorescent labels incorporated with each extension may be used to acquire sequencing data on all features in parallel. Successive iterations of enzymatic interrogation and imaging may also be used to build up a contiguous sequencing read for each array feature.

[0043] In some embodiments, the nCounter® Analysis system is used to detect intrinsic gene expression. The basis of the nCounter® Analysis system is the unique code assigned to each nucleic acid target to be assayed (International Patent Application Publication No. WO 08 / 124847, U. S. Patent No. 8,415,102 and Geiss etal. Nature Biotechnology. 2008. 26(3): 317-325; the contents of which are each incorporated herein by reference in their entireties). The code is composed of an ordered series of colored fluorescent spots which create a unique barcode for each target to be assayed. A pair of probes is designed for each DNA or RNA target, a biotinylated capture probe and a reporter probe carrying the fluorescent barcode. This system is also referred to, herein, as the nanoreporter code system. Specific reporter and capture probes are synthesized for each target. The reporter probe can comprise at a least a first label attachment region to which are attached one or more label monomers that emit light constituting a first signal; at least a second label attachment region, which is non-over-lapping with the first label attachment region, to which are attached one or more label monomers that emit light constituting a second signal; and a first target- specific sequence. Preferably, each sequence specific reporter probe comprises a target specific sequence capable of hybridizing to no more than one gene and optionally comprises at least three, or at least four label attachment regions, said attachment regions comprising one or more label monomers that emit light, constituting at least a third signal, or at least a fourth signal, respectively. The capture probe can comprise a second target-specific sequence; and a first affinity tag. In some embodiments, the capture probe can also comprise one or more label attachment regions. Preferably, the first targetspecific sequence of the reporter probe and the second target- specific sequence of the capture probe hybridize to different regions of the same gene to be detected. Reporter and captureprobes are all pooled into a single hybridization mixture, the "probe library". The relative abundance of each target is measured in a single multiplexed hybridization reaction. The method comprises contacting the tissue sample with a probe library, such that the presence of the target in the sample creates a probe pair - target complex. The complex is then purified. More specifically, the sample is combined with the probe library, and hybridization occurs in solution. After hybridization, the tripartite hybridized complexes (probe pairs and target) are purified in a two-step procedure using magnetic beads linked to oligonucleotides complementary to universal sequences present on the capture and reporter probes. This dual purification process allows the hybridization reaction to be driven to completion with a large excess of target-specific probes, as they are ultimately removed, and, thus, do not interfere with binding and imaging of the sample. All post hybridization steps are handled robotically on a custom liquid-handling robot (Prep Station, NanoString Technologies). Purified reactions are typically deposited by the Prep Station into individual flow cells of a sample cartridge, bound to a streptavidin-coated surface via the capture probe, electrophoresed to elongate the reporter probes, and immobilized. After processing, the sample cartridge is transferred to a fully automated imaging and data collection device (Digital Analyzer, NanoString Technologies). The level of a target is measured by imaging each sample and counting the number of times the code for that target is detected. For each sample, typically 600 fields-of-view (FOV) are imaged (1376 X 1024 pixels) representing approximately 10 mm2of the binding surface. Typical imaging density is 100- 1200 counted reporters per field of view depending on the degree of multiplexing, the amount of sample input, and overall target abundance. Data is output in simple spreadsheet format listing the number of counts per target, per sample. This system can be used along with nanoreporters. Additional disclosure regarding nanoreporters can be found in International Publication No. WO 07 / 076129 and W007 / 076132, and US Patent Publication No. 2010 / 0015607 and 2010 / 0261026, the contents of which are incorporated herein in their entireties. Further, the term nucleic acid probes and nanoreporters can include the rationally designed (e.g. synthetic sequences) described in International Publication No. WO 2010 / 019826 and US Patent Publication No.2010 / 0047924, incorporated herein by reference in its entirety.

[0044] Typically, the expression level of a gene may be expressed as absolute level or normalized level. For instance, levels are normalized by correcting the absolute level of a gene by comparing its expression to the expression of a gene that is not a relevant for determining the risk. This normalization allows the comparison of the level in one sample, e.g., a subject sample,to another sample, or between samples from different sources. Suitable genes for normalization include housekeeping genes selected from the group consisting of SDHA, STK11IP, TBC1D10B, TBP, UBB, OAZ1, PPIA, POLR2A, NRDE2, GUSB, G6PD, and ABCF1. Preferably the house keeping genes are selected from the group consisting of SDHA, STK11IP, TBC1D10B, TBP, UBB, and OAZ1. This normalization allows the comparison of the level in one sample, e.g., a subject sample, to another sample, or between samples from different sources.

[0045] Machine learning classifier:

[0046] The algorithm described in the present invention involves applying the determined expression levels of genes to a machine learning classifier designed to predict the risk of acute cellular rejection (ACR).

[0047] Typically, the algorithm produces a score. The term “score” refers to a numerical value derived by combining one or more parameters using a mathematical algorithm or formula. The combination of parameters can be achieved, for example, by multiplying each expression level with a defined coefficient and summing the resulting products to yield a score.

[0048] According to the present invention, the machine learning classifier is a supervised molecular classifier.

[0049] The generation of the machine learning classifier intended for diagnosing or predicting the risk of acute cellular rejection (ACR) in heart allograft recipients encompasses a series of meticulously outlined steps as described in the EXAMPLE:

[0050] i) initially, pathway enrichment analysis was conducted using the Over-Representation Analysis method. This method involves comparing the genes identified as differentially expressed in a dataset to a background set of genes, typically the entire genome or transcriptome, to determine if any biological pathways are statistically over-represented. The analysis helps to identify which pathways, biological processes, or molecular functions are potentially involved in the condition being studied. Over-representation analysis can be performed using various bioinformatics tools and databases such as Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Reactome. The results of this analysis provide insights into the biological significanceof the gene expression changes and guide further investigation into the molecular mechanisms underlying the condition.

[0051] ii) the development and validation of molecular classifiers involved utilizing the top 20 ranked differentially expressed genes (DEGs) for acute cellular rejection (ACR) as input variables for a multivariate logistic regression model. This supervised model underwent rigorous training with 10-fold cross-validation to ensure robustness and generalizability. In particular, the cross-validation helps to mitigate overfitting and ensures that the model performs well on unseen data.

[0052] - the efficacy of the classifiers was rigorously validated using an external validation cohort. The performance of the classifiers was assessed using a comprehensive range of metrics, including the Receiver Operating Characteristic Area Under the Curve (ROCAUC), Precision-Recall Area Under the Curve (PRAUC), accuracy, balanced accuracy, Fl score, negative predictive value (NPV), positive predictive value (PPV), sensitivity, specificity, and Brier Score. These metrics provided a thorough evaluation of the classifiers' ability to accurately diagnose or predict the risk of ACR in heart allograft recipients.

[0053] finally, sensitivity analyses incorporated feature selection methods to optimize gene selection, reduce correlation, and maintain model performance. Random selection of biopsies for patients with ≥2 EMBs was carried out to evaluate model stability, and balancing methods (Synthetic Minority Over-sampling Technique (SMOTE), ROSE (Random OverSampling Examples), Random Over-sampling and Random Undersampling) were employed to minimize the risk of overlooking the minority class (ACR). These methods collectively ensured that the model's performance was robust and that the risk of overlooking ACR cases was minimized.

[0054] The algorithm described in this invention can be executed by one or more programmable processors running computer programs that perform functions by processing input data and generating output. Additionally, the algorithm can be implemented using special-purpose logic circuitry, such as FPGA (field-programmable gate arrays) or ASIC (application-specific integrated circuits). Processors suitable for executing a computer program include both general-purpose and special-purpose microprocessors, capable of operating as part of any digital computer. Generally, a processor obtains instructions and data from read-only memory or random-access memory or both. Essential components of a computer include a processor for executing instructions and one or more memory devices for storing instructions and data.

[0055] Typically, a computer is also equipped with or connected to mass storage devices for data storage, such as magnetic disks, magneto-optical disks, or optical disks. However, it is not mandatory for a computer to possess these devices, and it can be embedded within another device. Suitable computer-readable media for storing program instructions and data encompass all forms of non-volatile memory, including semiconductor memory devices (e.g., EPROM, EEPROM, flash memory), magnetic disks (e.g., hard disks, removable disks), magneto-optical disks, and CD-ROM and DVD-ROM disks. The processor and memory may be supplemented by or incorporated into special-purpose logic circuitry.

[0056] To facilitate user interaction, embodiments of this invention can be implemented on a computer equipped with a display device, such as a CRT (cathode ray tube) or LCD (liquid crystal display) monitor, for presenting information to the user, along with input devices like a keyboard and a pointing device (e.g., mouse, trackball). Interaction with the user can also involve other sensory feedback mechanisms, including visual, auditory, or tactile feedback, and inputs can be received through various means, including acoustic, speech, or tactile input.

[0057] Furthermore, the algorithm can be deployed in a computing system comprising a back-end component (e.g., a data server), middleware component (e.g., an application server), or frontend component (e.g., a client computer with a graphical user interface or Web browser) to enable user interaction. These components can interconnect via any form of digital communication network, such as a LAN (local area network) or WAN (wide area network), including the Internet. Such a computing system may consist of clients and servers, typically remote from each other, interacting through a communication network. The client-server relationship is defined by the computer programs running on the respective computers, establishing a client-server relationship.

[0058] Moreover, the computing system described is capable of generating an automated molecular report. This report automates quality control assessment, normalizes data, and generates predictive molecular scores for any new biopsy. The report includes molecular rejection probability scores for ACR and no rejection. It also provides spatial contextualization of the new EMB within molecular rejection clusters derived from the development set using principal component analysis. Additionally, the report highlights key pathophysiological pathways involved and identifies the k-nearest neighbors corresponding to the 20 most similar pathology diagnoses. Instead of using binary cut-offs, the report presents probabilistic scores to enhancethe interpretability of molecular findings. A probability range, derived from the interquartile range (IQR) of each score, is provided to classify cases as low, medium, or high risk.

[0059] Clinical relevance:

[0060] The prediction methods as disclosed herein is useful for identifying patients (i.e. heart transplant recipients) having or at risk of ACR. Especially the prediction method of the invention is useful to predict patients with a high risk of ACR. Especially the diagnosis method of the invention is useful to diagnose patients with ACR. It is also useful to diagnose or predict the severity and / or grade of ACR.

[0061] The present invention relates to a method for identifying a biomarker for ACR, the biomarker being a diagnosis biomarker of ACR, a susceptibility biomarker of ACR, a prognostic biomarker of ACR or a predictive biomarker in response to the treatment of ACR, the method comprising at least the steps of:

[0062] carrying out the steps of the prediction method according to the invention for a first subject, to obtain a first predicted biological state, wherein the first subject is suffering from ACR,

[0063] carrying out the steps of the prediction method according to the invention for a second subject, to obtain a second predicted biological state, wherein the second subject is not suffering from ACR, and

[0064] selecting a biomarker target based on the comparison of the first and second predicted biological states.

[0065] In particular, the method of the present invention is particularly suitable for selecting a therapeutic regimen or determining if a certain therapeutic regimen is more appropriate for a patient identified as having or at high risk of ACR. Typically, protecting the allograft may be achieved using any suitable medical means known to those skilled in the art. In some embodiments, such reduction and protection comprise a therapeutic intervention with the subject such as administration of an immunosuppressive treatments, a corticosteroid and / or a T cell-depleting treatment. Said therapeutic intervention notably includes anti-thymocyte globulin (ATG), or a corticosteroid such as dexamethasone, prednisone or methylprednisone.. Reciprocally, where the recipient does not have ACR or is predicted at low risk of ACR, the immunosuppressive therapy can be reduced in order to diminish the potential for drug toxicity.

[0066] Thus, the invention also relates to a method for selecting a therapeutic regimen for a heart transplant recipient identified as having or being at high risk of ACR comprising:

[0067] performing the prediction method according to the invention, and

[0068] if it is concluded that the heart transplant recipient has or is at high risk of ACR, then selecting a therapeutic regimen chosen from immunosuppressive treatments, corticosteroids and T cell-depleting treatments, preferably chosen from anti-thymocyte globulin, dexamethasone, prednisone and methylprednisone.

[0069] In some embodiments, the methods of the present invention can be used to identify patients in need of frequent follow-up by a physician or clinician to monitor the therapeutic regimen. In some embodiments, a patient can be monitored using the methods as disclosed herein, and if on a first testing (i.e. initial testing) the patient is identified as having ACR or as being at high risk of ACR, the patient can be administered an appropriate therapeutic regimen, and on a second testing (i.e. follow-up testing), the patient is identified as not having ACR or as being at low risk of ACR, the patient can be administered with a therapeutic regimen at a maintenance dose.

[0070] Thus, the invention also relates to a method for monitoring the efficacy of a therapeutic regimen for a heart transplant recipient having or being at risk of ACR comprising:

[0071] i) performing the prediction method according to the invention with a first (i.e. initial) sample obtained from the recipient as a first testing (i.e. initial testing),

[0072] ii) if it is concluded from step i) that the heart transplant recipient has or is at high risk of ACR, then selecting a therapeutic regimen at a first dose, wherein said therapeutic regimen is chosen from immunosuppressive treatments, corticosteroids and T celldepleting treatments, preferably chosen from anti-thymocyte globulin, dexamethasone, prednisone and methylprednisone, and

[0073] iii) performing the prediction method according to the invention with a second (i.e. follow-up) sample obtained from the recipient as a second testing (i.e. follow-up testing), after a time period after steps i) and ii),

[0074] iv) wherein if it is concluded from step iii) that the heart transplant recipient does not have or is at low risk of ACR, then a maintenance dose of the therapeutic regimen ii) which is lower than the first dose, is selected.

[0075] Of course, the first and second samples obtained from the recipient are different.

[0076] Thus, the method of the present invention is particularly suitable for discriminating responder from non-responder.

[0077] As used herein the term “responder” in the context of the present disclosure refers to a subject that will achieve a response, i.e. the risk of ACR does show a reduction. A non-responder subject includes subjects for whom the risk of ACR does not show any reduction or improvement after the treatment.

[0078] The invention also relates to a method for discriminating a responder heart transplant recipient from a non-responder heart transplant recipient having or being at risk of ACR comprising: i) performing the prediction method according to the invention with a first sample obtained from the recipient before any treatment, as a first testing which results in a first probability of ACR,

[0079] ii) performing the prediction method according to the invention with a second sample obtained from the recipient after a treatment, as a second testing which results in a second probability of ACR,

[0080] iii) comparing the first and second probabilities of ACR,

[0081] iv) wherein if it is concluded from step iii) that the recipient does not have or is at low risk of ACR, then concluding that said recipient is a responder,

[0082] - v) and wherein if it is concluded from step iii) that the recipient has or is at high risk of ACR, then concluding that said recipient is a non-responder.

[0083] Typically if the second probability is significantly lower than the first probability, then it can be concluded from step iii) that the recipient does not have or is at low risk of ACR (and is a responder). Typically if the second probability is not significantly different from the first probability, or is higher than the first probability, then it can be concluded from step iii) that the recipient has or is at high risk of ACR (and is a non-responder).

[0084] In some embodiments, screening patients for identifying patients having or being at high risk of ACR using the prediction method as disclosed herein is also useful to identify patients most suitable or amenable to be enrolled in clinical trial for assessing a therapy for management of allograft, which will permit more effective subgroup analyses and follow-up studies. Furthermore, the prediction method as disclosed herein can be suitable for monitoring 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. Accordingly, the output of the algorithm (e.g. thescore) represents a suitable surrogate marker for use in a clinical trial for assessing the efficiency of a particular therapy.

[0085] In some embodiments, screening patients for identifying patients having or being at high risk of ACR using the prediction method as disclosed herein is also useful to give biological information about said patients. It can also inform about the biological pathways that are involved for a given patient, and / or inform about cell types and / or immunological pathways that are involved.

[0086] In some embodiments, screening patients for identifying patients having or being at high risk of ACR using the prediction method as disclosed herein is also useful to identify novel ACR treatments (different from graft) and / or for drug repurposing.

[0087] 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.

[0088] FIGURES:

[0089] Figure 1. Differential gene expression analysis of AMR and ACR in heart allograft biopsies. Panel A. Volcano plot of differentially expressed genes associated with AMR. The top differentially expressed genes were related to natural killer cells, endothelial cells, the IFN-inducible pathway, and monocyte-macrophage recruitment. Panel B. Volcano plot of differentially expressed genes associated with ACR, highlighting the most significant genes involved in T-cell recruitment, activation of innate and adaptive immune responses, CD3 receptor activation, and CD28 signaling. Each dot represents an individual transcript. Light blue points indicate genes with a significant p-value adjusted for a false discovery rate of <0.05. The x-axis represents the log2 fold change, while the y-axis shows the false discovery rate (logic transformed) for the association of each transcript with the rejection profile compared with all other cases. ACR, acute cellular rejection; AMR, antibody -mediated rejection.

[0090] Figure 2. Association between molecular scores and pathologic severity of rejection. Panel A. Derivation set. Panel B. Internal validation set. A stepwise increase in AMR and ACR molecular scores was observed with increasing disease severity, as assessed by conventional histology (ISHLT working formulations), in both the derivation and validation cohorts. ACR, acute cellular rejection; AMR, antibody -mediated rejection.

[0091] Figure 3. Evaluation of model performance: discrimination and calibration curves of molecular models for detecting AMR and ACR. Panels A and B. AMR model performance in the derivation and validation sets, respectively. Panels C and D. ACR model performance in the derivation and validation sets, respectively. Left-hand side: ROC-AUC curves illustrating the performance of molecular-based models. The discrimination of both the AMR and ACR models was excellent in both the derivation and validation sets (all AUC > 0.80).

[0092] Right-hand side: Calibration curves demonstrating adequate calibration for both AMR and ACR models across the derivation and validation sets, reflecting the reliability of probabilistic predictions made by the molecular diagnostic models. AUC, area under the curve; ACR, acute cellular rejection; AMR, antibody -mediated rejection; ROC, receiver operating characteristic.

[0093] EXAMPLE:

[0094] Methods:

[0095] Study design and patient and EMB selection

[0096] We conducted an international multicenter study, assembling a deeply phenotyped cohort of heart transplant recipients from four referral centers: Pitie-Salpetriere and Georges Pompidou hospitals in Paris, France; University General Hospital of Padova, Italy; and Cedars-Sinai Medical Center in Los Angeles, USA (not shown). FFPE-EMBs were collected as part of routine clinical care between 2011 and 2021 and graded according to International Society for Heart and Lung Transplantation (ISHLT) working formulations.10 11Biopsies were selected based on predefined criteria to ensure representation across a large spectrum of cardiac allograft rejection. Ambiguous cases — including biopsy -negative rejection, mixed rejection, and Quilty lesions — were excluded. Control biopsies were matched with rejection cases based on transplant year, time post-transplant, and recipient sex. Biopsies processed before 30 September 2023 were included in the main cohort (derivation and internal validation cohort), while those processed after 1 January 2024 were designated as the external validation cohort and analyzed independently from the main cohort.

[0097] All patients provided written informed consent under a protocol approved by the local institutional review boards: Comite de Protection des Personnes Ile-de-France II (research project 20141226, Institutional Review Board registration 00001072), Ethics Committee of Azienda-Hospital University Padua (protocol number 0062556 / 2022), and the Institutional Review Board of Cedars-Sinai Medical Center. All procedures adhered to the Declaration of Helsinki and the ethical standards of both institutional and national research committees. Thestudy methodology and results were reported in accordance with the 2015 Standards for Reporting of Diagnostic Accuracy Studies guidelines.

[0098] Data collection

[0099] Comprehensive patient- and biopsy -related data were collected at all participating centers and compiled into a centralized, anonymized database to ensure harmonization across study sites. The collected data included (i) recipient and donor baseline characteristics (recipient age, sex, and cause of heart failure; donor age and sex), (ii) transplant procedure parameters (ischemic time and presence of severe primary graft dysfunction), (iii) immunosuppressive therapy (induction therapy and maintenance immunosuppressive regimen), (iv) clinical parameters (indication for EMB, time post-transplant, and presence of symptoms or signs of heart failure), (v) immunological parameters (presence of circulating anti-HLA donor-specific antibodies [DS As]), (vi) echocardiographic parameters (left ventricular ejection fraction), and (vii) histological parameters (ACR and AMR grades according to ISHLT working formulations).

[0100] Procedures and clinical protocols

[0101] Histology and immunohistochemistry

[0102] EMBs were performed for routine clinical purposes, either as protocol or for-cause biopsies, via jugular or femoral access. Samples were formalin-fixed, paraffin-embedded, and routinely stained with hematoxylin and eosin for histopathologic evaluation, then processed and assessed according to international standards.22EMBs were graded based on the latest international working formulations.10’11Immunostaining was performed on tissue sections using specific antibodies, as previously described.23C4d staining was conducted on all biopsies, with grading based on pattern and intensity scored semi-quantitatively on a scale from 0 to 3, following the ISHLT working formulation.11Only C4d-positive EMBs were considered for the diagnosis of pAMRl(I+) cases. Mixed rejection was defined by the presence of both ACR > 2R and AMR > pAMR2. All EMBs were assessed by highly trained cardiovascular pathologists (PB, PR, MF, AA).

[0103] DSAs

[0104] Anti-HLA DSAs were assessed using Luminex technology (Luminex LAB Screen Single Antigen; One-Lambda Inc., West Hills, CA, USA) as previously described.17A mean fluorescence intensity value of 1000 was considered positive.24The immunodominant DS A was defined as the DSA with the highest mean fluorescence intensity, whether class I or II.

[0105] Only tests performed at the time of the EMB or within the previous 3 months for EMBs conducted during the first year post-transplant, or within the previous 6 months for EMBs performed beyond the first year post-transplant, were considered.

[0106] Outcomes measure

[0107] The reference diagnosis of rejection used in model development was based on the histopathological assessment of cardiac allograft biopsies. Diagnostic categories included AMR, ACR, and no evidence of rejection. For all molecular analyses, ACR was defined as ACR> IR and AMR as AMR > pAMRl.10 11

[0108] The primary outcome was the evaluation of the prediction performance of the molecular classifiers (ACR and AMR rejection scores) in the internal and external validation cohorts. A multiparametric approach was used to predict biopsy-proven rejection, defined as ACR ≥ 1R and AMR ≥ pAMR1 for the ACR and AMR molecular models, respectively.

[0109] Sample size calculation

[0110] To ensure the robustness and reliability of our findings, a comprehensive sample size calculation was performed. The minimum sample size was determined based on an acceptable difference of 0.05 between the adjusted and apparent R-squared values, reflecting the anticipated level of precision in our statistical model. A margin of error of 0.05 was deemed acceptable to balance the risk of Type I and Type II errors. Additionally, a desired statistical power of at least 0.8 was set, indicating an 80% probability of detecting a true rejection event. This power threshold mitigates the risk of Type II errors, increasing the likelihood that the study findings will be both valid and replicable. Incorporating these parameters, the required minimum sample size was determined to be 400 biopsies in the derivation set. This substantial sample size was necessary to ensure the statistical robustness of our molecular approach, allowing for reliable detection of differences in immune cell infiltration and gene expression signatures.

[0111] Gene expression profiling

[0112] RNA isolation from FFPE biopsies

[0113] To obtain the 100 ng of RNA required for transcriptome sequencing, three 20-µm-thick sections were cut from each paraffin block. The RNA isolation process involved three key steps: deparaffinization, removal of contaminants (both organic and DNA-related), and final serialcentrifugation to remove salt contaminants and concentrate the isolated RNA. Total RNA was isolated using the RNeasy FFPE Kit (Qiagen, Hilden, Germany). Quantity and quality assessments of RNA were performed before downstream processing using the NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA, USA).

[0114] Transcriptome sequencing

[0115] Gene expression analysis was performed using nCounter technology (Nanostring) with the targeted B-HOT Panel.20Designed by the Banff Molecular Diagnostics Working Group, this panel includes 770 genes (758 endogenous and 12 housekeeping genes) representing key functional pathways involved in solid organ transplantation rejection. The nCounter technology enables digital detection and direct molecular barcoding of targeted mRNA through the design of two specific probes (capture and reporter probes) for each mRNA.25This direct assessment eliminates the need for amplification steps, thereby reducing technical variation and probe instability.

[0116] Bioinformatic and statistical analysis

[0117] Quality control and normalization

[0118] Gene expression raw count data were analyzed using nSolver Analysis Software (version 4.0.70) and subjected to extensive quality control metrics, assessing imaging, binding density, linearity of the positive controls, and limit of detection. Samples that failed to meet at least one quality control metric were flagged and excluded from further analysis. Pre-normalization steps included the assessment of housekeeping gene stability and data visualization to detect potential batch effects. Raw gene expression data were normalized using the Reduced Unwanted Variation method.26,27Quality control, housekeeping gene stability assessment, and normalization steps ensured correction for technical variations in sequencing data. All cohorts included in the study, from the training set to the internal and external validation sets, were independently normalized and corrected for potential batch effects. In the main cohort, 610 EMBs from 461 patients were analyzed, including 192 AMR, 296 ACR, and 122 non-rejection biopsies. Among these, 19 (3.1%) biopsies did not meet quality control criteria and were excluded from further analysis. Biopsies that passed all pre-analytical steps were included in the final molecular reference set. The main cohort ultimately consisted of 591 biopsies from 455 patients, comprising 188 AMR, 289 ACR, and 114 non-rejection cases. These were then divided into a training set and an internal validation set in an 80:20 ratio using the dataPartition function of the caret package, ensuring homogeneous stratification of rejection events acrossboth sets (training set: 152 AMR, 232 ACR, and 92 non-rejection cases; validation set: 37 AMR, 57 ACR, and 22 non-rejection cases). The external validation set consisted of 80 EMBs from 80 patients, including 20 AMR, 32 ACR, and 28 non-rejection cases.

[0119] Differential expression and pathway analysis

[0120] Differential expression analysis was conducted using the DESeq function (DESeq2 package),28with application of the Wald test to compare gene expression between the rejection condition of interest and all other conditions. Differentially expressed genes (DEGs) were identified through linear fitting after calculating the fold change for each comparison of interest. All p-values were adjusted for multiple comparisons using the Benjamini-Hochberg method to control the false discovery rate (FDR), with significant genes filtered based on an FDR p-value threshold of <0.05. A shrinkage procedure was applied to further control the FDR. To annotate DEGs, we used UniProt, the GeneCards database, and the Banff Molecular Report reference.20,29,30Pathway enrichment analysis was performed to characterize the pathophysiological relationships among the most DEGs using the Over-Representation Analysis method through the enrichplot package.31A hypergeometric model with the Reactome repository was applied, with pathway identifiers and general membership classifications retrieved from the repository. All statistical and bioinformatic analyses were conducted using R version 4.0.5 (R Foundation for Statistical Computing).

[0121] Development and validation of molecular classifiers of rejection

[0122] Supervised molecular classifiers were developed in the training set for AMR, ACR, and nonrejection and evaluated in an equally partitioned validation set. The top 20 ranked DEGs for AMR and ACR were used as input variables for a multivariate logistic regression model (glm function from the caret package). The supervised model was trained using 10-fold cross-validation as a resampling procedure.

[0123] To transition from molecular scores to interpretable diagnoses, we applied a threshold derived from the Youden index of the continuous predicted probabilities, classifying each case as AMR, ACR, or non-rejection. Rejection-related classifiers were then evaluated for model performance using a multiparametric approach, including precision-recall area under the curve (PR-AUC), diagnostic accuracy, positive predictive value, true positive rate, true negative rate, receiver operating characteristic (ROC)-AUC, Brier score, and Fl score. A confusion matrix was usedto assess classification accuracy by comparing the agreement between histopathological evaluation and the predictive molecular models.

[0124] Finally, the stepwise increase in rejection-related gene expression with rejection severity was tested for both AMR and ACR scores. A molecular map of cardiac allograft rejection was generated using principal component analysis based on individual molecular scores, reflecting the molecular heterogeneity of rejection.

[0125] Sensitivity analyses

[0126] Various sensitivity analyses were conducted to assess the robustness of the molecular rejection diagnostic models. First, feature selection methods, including maximum relevance-minimum redundancy and random forest models, were applied to optimize gene selection by reducing correlation and collinearity while maintaining similar model performance. Second, for patients with ≥2 EMBs included in the reference set, biopsies were randomly selected to evaluate model stability when including only one EMB per patient, as well as the impact of time post-transplant. Third, to minimize the risk of overlooking the minority class (AMR), four balancing methods were applied: Upsampling, Downsampling, SMOTE, and ROSE.

[0127] Results:

[0128] Characteristics of patients and EMBs

[0129] Overall, the main cohort consisted of 591 biopsies from 455 patients. The baseline characteristics of the recipients, donors, and transplant procedures are detailed in Table 1. The mean recipient age was 51.1 ± 13.1 years. Most patients were male (n=323, 71.0%), and dilated cardiomyopathy was the most common cause of heart failure (n=203, 45.5%). Routine immunosuppression primarily consisted of a combination of calcineurin inhibitors, mycophenolate mofetil, and corticosteroids (n = 389, 85.4%).

[0130] Biopsies were representative of a large spectrum of heart allograft rejection pathology and included 188 (31.8%) AMR cases (pAMR1(I+): n = 51, pAMR1(H+): n = 58, pAMR2-3: n=79), 289 (48.9%) ACR cases (ACR 1R: n = 174, ACR 2-3R: n = 115), and 114 (19.3%) nonrejection cases (ACR 0, pAMRO). The median time between transplant and EMB was 238 days (IQR: 88-844), with 60.1% of EMBs performed within the first year post-transplant. Most biopsies (86%) were protocol EMBs. At the time of biopsy, anti-HLA DSAs were detected in 193 cases (32.7%). Acute allograft dysfunction, defined as a left ventricular ejection fraction of<0.45 and / or an absolute drop of >0.15, was present in 43 cases (7.3%). Evidence of CAV grade 1 on coronary angiography was observed in 17.5% of patients (Table 2).

[0131] Molecular patterns of heart allograft rejection

[0132] AMR

[0133] Analysis of the DEGs comparing AMR cases to all other cases (ACR and non-rejection) is illustrated in Figure 1A. The top 20 ranked DEGs revealed an increased expression of transcripts associated with the adaptive immune response TM4SF18, natural killer cells (FCGR3A / B), endothelial cells (PECAM1, NOS3, KLF4, CAV1, ROBO4), DS A response PLAT), IFN-gamma inducible pathway (CXCL11, CX3CL1, PLA1A, CCL19), cytokine signaling (BAFT, IL1R2, LTB), neutrophil degranulation (CLEC4C), Th2 dedifferentiation (GATA3), and monocyte-macrophage recruitment (LILRB1) (Table 6). The most significant AMR-associated pathways derived from DEG analysis were related to interferon-gamma (q = 1.97E-05), interleukin signaling (q = 1.63E-07), toll-like receptor signaling (q = 7.83E-12), costimulation by CD28 and B-cell activation (q = 2.23E-03), antigen processing and crosspresentation (q = 3.76E-04), lymphoid and non-lymphoid immune interactions (q = 7.17E-06), chemokine binding (q = 7.83E-12), and the NF-kB pathway (q = 3.27E-04) (Table 7).

[0134] ACR

[0135] The top DEGs identified when comparing ACR cases to all other cases were associated with T-cell receptor signaling (CD8A), T-cell checkpoint signaling (CD27, LAG3), natural killer cell and innate lymphoid cell development (TCF7), T-cell mediated rejection (PTPN7), B-cell activity (MS4A1), IFN-gamma inducible response (CCL19), quantitative constitutive macrophage markers (CCL18), TNF family signaling LTB, MMP9), CD3 receptor activation (CD3E), chemokine signaling (JAK3), cytotoxicity (GZMK, LCK), and CD28 co-stimulation signaling (BTLA, CD28) (Figure IB, Table 7).

[0136] The most significant ACR-associated pathways were related to the TCR signaling cascade (q = 5.14E-06), IL-2 signaling (q = 5.14E-06), phosphorylation of the CD3 subunit (q = 2.55E-07), cell surface interactions (q = 8.14E-09), and chemokine receptor activity (q = 8.47E-13) (data not shown).

[0137] Development of a gene expression-based heart allograft rejection classifier

[0138] The derivation set included 475 randomly selected EMBs, distributed as follows: 151 AMR cases (pAMR1(I+): n = 36, pAMR1(H+): n = 46, pAMR2-3: n = 69), 232 ACR cases (ACR 1R: n = 143, ACR 2-3R: n = 89), and 92 non-rejection cases. The top 20 DEGs in ACR and AMR were used as input variables for multivariate logistic regression. Using a 10-fold cross-validation approach, supervised molecular classifiers were developed, and predicted probabilities for each AMR and ACR model were generated.

[0139] A stepwise increase in molecular scores was observed with increasing rejection severity and activity. Both AMR and ACR molecular scores were strongly associated with histopathological assessments of rejection severity according to international working formulations (ACR: p for trend = 4.54E-38, AMR: p for trend = 7.55E-41) (Figure 2). The molecular AMR score showed no association with ACR severity, nor did the ACR molecular score with AMR severity.

[0140] Principal component analysis derived from molecular rejection scores was used to visualize the molecular heterogeneity of cardiac allograft rejection. Rejection probabilities based on molecular scores effectively discriminated between the two main rejection profiles as well as non-rejection biopsies (principal component 1: variance 72%) (data not shown).

[0141] In the training set, the statistical performance of the molecular model for detecting AMR demonstrated a ROC-AUC of 0.83, a PR-AUC of 0.76, with a Brier score of 0.14 and an accuracy of 77.68 (Table 3, Figure 3A). The molecular model for detecting ACR showed comparable performance metrics, with a ROC-AUC of 0.85, PR-AUC of 0.85, Brier score of 0.16, and an accuracy of 77.47 (Table 4, Figure 3C).

[0142] Internal and external validation of the heart allograft rejection classifiers

[0143] To evaluate the generalizability and reproducibility of the gene expression-based rejection system, the AMR and ACR classifiers were tested in both an internal and an external validation set. Quality control and normalization were conducted independently from the training set and applied to raw data. The AMR and ACR gene expression-based models were then applied to the normalized data of the validation set.

[0144] Internal validation

[0145] The internal validation set included 116 EMBs, distributed as follows: 37 AMR cases (pAMR1(I+): n = 15, pAMR1(H+): n = 12, pAMR2-3: n = 10), 57 ACR cases (ACR 1R: n = 31, ACR 2-3R: n = 26), and 22 non-rejection cases.

[0146] A significant stepwise increase in molecular rejection scores was observed with increasing rejection severity, as assessed by ISHLT working formulations for both ACR and AMR (ACR: p for trend = 1.82E-13, AMR: p for trend = 5.17E-10) (Figure 2).

[0147] The molecular model demonstrated excellent performance in detecting AMR, with a diagnostic accuracy of 81.89%, ROC-AUC of 0.844, PR-AUC of 0.742, Brier score of 0.143, and Fl score of 0.70 (Table 3, Figure 3B). The model’s performance in detecting ACR was similarly strong, with an accuracy of 77.58%, ROC-AUC of 0.812, PR-AUC of 0.827, Brier score of 0.176, and Fl score of 0.76 (Table 4, Figure 3D).

[0148] External validation

[0149] The external validation cohort included 80 EMBs from 80 patients, distributed as follows: 20 AMR cases (pAMR1(H+): n = 3, pAMR2-3: n = 17), 32 ACR cases (ACR 1R: n = 7, ACR 2-3R: n = 25), and 28 non-rejection cases. The median time between heart transplantation and EMB was 377 days (IQR: 119-762), with 42.5% of EMBs performed within the first year posttransplant. The mean recipient age was 51.5 ± 12.1 years, and most patients were male (n = 58 / 80, 72.5%).

[0150] In the external validation, we confirmed a significant stepwise increase in molecular rejection scores with increasing rejection severity, as assessed by ISHLT working formulations for both ACR and AMR (ACR: p for trend = 5.58E-08, AMR: p for trend = 1.38E-09).

[0151] The molecular model demonstrated strong performance in detecting AMR, achieving a diagnostic accuracy of 82.5%, ROC-AUC of 0.82, PR-AUC of 0.70, Brier score of 0.14 and an Fl score of 0.69. The performance of the molecular model for detecting ACR was similarly robust, with an accuracy of 75%, ROC-AUC of 0.81, PR-AUC of 0.75, Brier score of 0.18 and Fl score of 0.73 comparable to the development and internal validation cohorts (Table 5).

[0152] Molecular rejection classifiers identify low-risk AMR and ACR subgroups within molecular-negative rejection cases

[0153] Discrepancies between pathologic and molecular diagnoses were assessed by comparing observed and predicted rejection classifications. While overall agreement was excellent, few discrepancies were observed, particularly among molecular-negative rejection cases (data not shown). Among pathology ACR cases, molecular-negative cases were more frequently classified as low-grade ACR 1R rejection than were molecular-positive ACR cases (molecularnegative: 1R = 72.8%, >2R = 27.2%; molecular-positive: 1R = 32.5%, >2R = 67.5; p < 0.001). Notably, among pathology AMR cases, molecular-negative cases were significantly less likely to have circulating anti-HLA DSA (8.91% vs. 72.95%, p < 0.001) or graft dysfunction (2.57% vs. 15.57%, p < 0.005) and were more frequently classified as low-grade pAMRl rejection than were molecular-positive AMR cases (molecular-negative: pAMRl = 91.1%, >pAMR2 = 8.9%; molecular-positive: pAMRl = 54.7, >pAMR2 = 45.3%; p < 0.001). Conversely, control biopsies classified as molecular AMR more frequently exhibited circulating DSAs (14.45% vs.

[0154] 8.69%, p < 0.05) and graft dysfunction (3.47% vs. 1.27%, p < 0.05) than did molecular-negative control biopsies. Taken together, these findings suggest that molecular analysis may capture sub-histological critical information, identifying molecular-negative rejection cases as lower risk and molecular-positive control cases as higher risk.

[0155] Sensitivity analyses

[0156] Various sensitivity analyses were performed to assess the robustness of our findings across different clinical scenarios and subpopulations.

[0157] Learning curves of the models

[0158] Model learning efficiency was achieved at a sample size of approximately 400, as indicated by the convergence of learning curves for both the training and validation sets (data not shown).

[0159] These findings confirm that the statistical performance of the molecular models was not limited by the number of samples.

[0160] Influence of time post-transplant and biopsy indication

[0161] In this study, 79 EMBs (13.5%) were for-cause biopsies, and 355 EMBs (60.1%) were performed within the first year post-transplant. The statistical performance of both AMR and ACR models remained stable and adequate regardless of the clinical indication for EMB or time post-transplant.

[0162] Gene selection method

[0163] Alternative gene selection methods, including maximum relevance-minimum redundancy and random forest-derived approaches for AMR and ACR, did not yield significant improvements in diagnostic performance. However, these methods consistently reduced the number of input variables compared with the linear model (Table 3).

[0164] Effect of multiple biopsies

[0165] Approximately 30% of patients had multiple biopsies, with a maximum of three EMBs analyzed per patient. When randomly selecting one EMB per patient, both the top DEGs and pathway analyses remained consistent, with no noteworthy changes observed.

[0166] Class imbalance

[0167] In both the derivation and validation sets, rejection class events occurred at a rate of approximately 31.8%, with moderate imbalance, particularly in the AMR rejection class. Four balancing methods — Upsampling, Downsampling, SMOTE, and ROSE — were tested, but none significantly improved the prediction performance of the AMR models (data not shown).

[0168] Automated reporting of gene expression-based classifier for AMR and ACR

[0169] To provide clinicians and pathologists with a practical companion tool, we developed a dedicated molecular report that automates quality control assessment, normalizes data, and generates predictive molecular scores for any new biopsy (data not shown). The report includes molecular rejection probability scores for AMR, ACR, and no rejection, along with spatial contextualization of the new EMB within molecular rejection clusters derived from the development set using principal component analysis. Additionally, the report highlights key pathophysiological pathways involved and identifies the k-nearest neighbors corresponding to the 20 most similar pathology diagnoses. Rather than using binary cut-offs, the report presents probabilistic scores to enhance the interpretability of molecular findings. A probability range, derived from the IQR of each score, is provided to classify cases as low, medium, or high risk.

[0170] Discussion:

[0171] In this study, we developed and validated the first FFPE tissue-based molecular diagnostic system for heart transplant rejection using a highly annotated cohort. This system integrates a consensus gene set with a transcriptome technology applicable to FFPE-EMBs. Molecular classifiers demonstrated strong performance, consistency, and validity across both internal and external validation cohorts. Key advantages of this approach — including reproducibility, cost-effectiveness, FFPE compatibility, and the absence of a need for an extra-core biopsy — make it well-suited for routine clinical application. The automated molecular report provides valuable, objective insights into graft status and may serve as a novel companion tool to refine the pathological diagnosis of cardiac allograft rejection.

[0172] While the number of EMBs is declining with the clinical adoption of non-invasive biomarkers for rejection,32 34there remains a need for deeper tissue-based analysis to enhance the diagnostic yield of EMBs. A more nuanced and comprehensive understanding of the underlying mechanisms could lead to more accurate diagnoses.9,33,35The application of kidney rejection molecular classifiers to heart transplantation was the first attempt to use gene expression profiling to refine the diagnosis of cardiac rejection.16,17,36However, significant drawbacks have limited the routine clinical use of fresh or frozen tissue-based molecular approaches. In response, the Banff Heart Group has recommended the use of FFPE-based technologies to assess the clinical applicability of molecular diagnostics in heart transplantation.9,37

[0173] The B-HOT Panel is a commercially available gene expression panel that operates on the NanoString platform, enabling absolute quantification of transcripts from FFPE samples.19,38Our group recently demonstrated in silico that B-HOT Panel-based gene expression analysis effectively captures biologically and clinically relevant genes, pathways, and networks that reflect the key pathophysiological mechanisms of AMR in heart transplantation.21This proof-of-concept study supported the clinical use of the B-HOT Panel as a reliable proxy for whole-transcriptome analysis in gene expression profiling of heart allograft biopsies. The present study advances this work by demonstrating the reliability and validity of targeted molecular diagnostic classifiers. These classifiers may help address the limitations of whole-transcriptome approaches by reducing discrepancies between histology and molecular diagnosis and minimizing the overdiagnosis of AMR — two critical challenges observed in previous molecular strategies.39,40

[0174] In addition to aligning predictions with expected histological diagnoses, evaluating models built on an uncertain ground truth requires a thorough reassessment of discordant cases between histology and gene expression-based probabilities. A cohort with extensive clinical, demographic, and histological data is crucial for investigating these discrepancies because it provides the granularity needed to address the complexities of allograft rejection. In our study, the analysis of discrepancies suggested that classifiers may identify a low-risk subgroup of

[0175] molecular-negative rejection cases, particularly for AMR, which was often associated with subclinical low-grade rejection without DSA. However, some discrepancies may remain inconclusive and could be attributed to early-stage rejection, ischemia–reperfusion injury in early post-transplant biopsies, incomplete histological data, sampling error, or variations in mRNA assay quality.

[0176] We developed an automated diagnostic system that generates a report detailing sample quality and molecular scores for the probability of AMR and ACR. The cohort used for model development represents a clinically relevant population of heart transplant biopsies, which is essential to mitigate challenge bias, where a simplified “sick versus well” design can limit generalizability in real-world clinical settings. Multi-disciplinary expertise was incorporated at every stage of model development and validation to address potential risks and benefits for patients while ensuring interpretability for clinicians.

[0177] Our study represents the first effort to develop and validate B-HOT gene panel -based predictive models for cardiac allograft rejection. The introduction of an automated reporting system may streamline the assessment of the clinical utility of a gene expression-based decision support tool in transplant patient care. However, its potential long-term impact on patient management and graft survival necessitates the development of multicenter prospective studies. Discordant cases underscore the limitations of both gene expression-based and histological diagnostic methods. Given that the true disease state is unknown and graft loss or treatment response are unreliable endpoints, evaluating the added value of molecular diagnostics in cardiac transplantation remains challenging. The successful integration of gene expression analysis into clinical practice requires the adoption of standardized protocols, reproducible technologies, and careful interpretation of results in the context of patient-specific, immunological, histological, and diagnostic information. By incorporating all relevant clinical parameters into multivariate gene expression models, precision diagnostics in cardiac transplant care can be further optimized.

[0178] Conclusion:

[0179] Using targeted FFPE-based transcriptome technology with a gene panel designed to detect key pathophysiological pathways in transplant rejection, we developed and validated a molecular diagnostic system with enhanced reproducibility and reliability, specifically tailored for cardiac allograft rejection. This approach is readily applicable in clinical practice as a companion tool to pathology and has the potential to refine rejection diagnosis.

[0180] TABLES:

[0181] Table 1. Baseline characteristics of patients (n=455) - main cohort

[0182] N Reference set (n=455)

[0183] Recipient characteristics

[0184] Recipient age, years (mean ±SD) 455 51.1 (± 13.05)

[0185] Recipient gender 455

[0186] Female, n (%) 132 (29.01)

[0187] Male, n (%) 323 (70.99)

[0188] Primary heart disease 446

[0189] Dilated cardiomyopathy, n (%) 203 (45.51)

[0190] Ischemic cardiomyopathy, n (%) 130 (29.15)

[0191] Congenital cardiomyopathy, n (%) 21 (4.71)

[0192] Retransplantation, n (%) 13 (2.92)

[0193] Valvular cardimyopathy, n (%) 3 (0.67)

[0194] Hypertrophic cardiomyopathy, n (%) 3 (0.67)

[0195] Other, n (%) 73 (16.37)

[0196] Transplant baseline characteristics

[0197] Cold ischemia time, min (mean ± SD) 441 195 (± 55.93)

[0198] Primary graft dysfunction, n (%) 348 104 (29.88)

[0199] Induction therapy 444

[0200] ATG, n (%) 382 (86.1)

[0201] Anti IL-2R, n (%) 43 (9.70)

[0202] Other, n (%) 13 (3)

[0203] None, n (%) 5 (1.2)

[0204] Maintenance immunosuppression

[0205] Steiroids, n (%) 401 (88.13)

[0206] Tacrolimus / cyclosporine, n (%) 426 (93.66)

[0207] Mycofenolate mofetil or

[0208] mycophenolic acid / azathioprine, n 388 (85.27)

[0209] (%)

[0210] Donor characteristics

[0211] Donor age, years (mean ±SD) 434 46.01 (±14.52)

[0212] Donor gender 416

[0213] Female, n (%) 154 (37.02)

[0214] Male, n (%) 262 (62.98)

[0215] Abbreviations: ATG (anti-thymocyte globulin); SD (Standard Deviation).

[0216] N refers to the total number of variables available.

[0217] Table 2. Endomyocardial biopsy characteristics (N=591) - main cohort

[0218] „ Biopsies included

[0219] W(n=591)

[0220] Characteristics of endomyocardial biopsies

[0221] EMB before 1-year post-transplant, n (%) 591 355 (60.1)

[0222] Time between HTx and EMB, median (25th-75th percentile)

[0223]

[0224] Type of EMB 585

[0225] For Cause, n (%) 79 (13.50)

[0226] Protocol, n (%) 506 (86.50)

[0227] Cardiac function and events

[0228] ISHLT 2013 pAMR grade 591

[0229] pAMR0, n (%) 403 (68.19) pAMR 1 (H+), n (%) 58 (9.81) pAMR 1 (I+), n (%) 51 (8.63)

[0230] > pAMR 2, n (%) 79 (13.36)

[0231] ISHLT 2004 ACR grade 591

[0232] 0R, n (%) 302 (51.10)

[0233] 1R 1A, n (%) 76 (12.86)

[0234] 1R 1B, n (%) 73 (12.35)

[0235] > 2 R, n (%) 140 (23.69)

[0236] LVEF at time of EMB, median (25th-75th percentile) 540 63 (58 - 66)

[0237] Graft dysfunction, % 540 43 (7.3)

[0238] Angiography at time of EMB 486

[0239] CAV grade 0, n (%) 401 (82.51)

[0240] CAV grade 1, n (%) 85 (17.49)

[0241] Anti-HLA DSA at the time of index EMB, n (%) 385

[0242] DSA positive, n (%) 201 (52.20)

[0243] DSA class (I), n (%) 64 (32)

[0244] DSA class (II), n (%) 136 (68)

[0245] Abbreviations: HTx (heart transplantation); AMR (antibody-mediated rejection); ACR (acute cellular rejection); CAV (cardiac allograft vasculopathy); DSA (donor-specific antibody); EMB (endomyocardial biopsy); ISHLT (International Society for Heart and Lung Transplantation); LVEF (left ventricular ejection fraction).

[0246] N refers to the total number of variables available.

[0247] Table 3. Molecular diagnostic model metrics for antibody-mediated rejection in heart.

[0248] AMR model performance metrics are assessed for both training and internal validation sets. The precision recall represents the proportion of positive biopsies in each cohort.

[0249]

[0250] Table 4. Molecular diagnostic model metrics for acute cellular rejection in heart. ACR model performance metrics are assessed for both training and internal validation sets. The precision recall represents the proportion of positive biopsies in each cohort.

[0251]

[0252] Table 5. Molecular diagnostic model metrics for rejection profiles in heart transplantation in external validation set. The precision recall represents the proportion of positive biopsies in each cohort.

[0253]

[0254] Table 6. Top 20 significant genes associated with antibody-mediated rejection in heart allografts.

[0255] Abbreviations: FDR p-value (p-value adjusted for false discovery rate); FC (fold-change)

[0256] Gene Gene name log2FC p-value FDR p-value

[0257] PLA1A phospholipase A1 member A 0.96417 2.89E-36 2.10E-33

[0258] CXCL11 C-X-C motif chemokine ligand 11 1.09031 1.83E-27 1.33E-24

[0259] PECAM1 platelet and endothelial cell adhesion molecule 1 0.40679 1.48E-25 1.08E-22

[0260] CCL19 C-C motif chemokine ligand 19 -1.00396 1.93E-19 1.40E-16

[0261] PLAT plasminogen activator, tissue type 0.47318 1.19E-18 8.64E-16

[0262] TM4SF18 transmembrane 4 L six family member 18 0.42004 5.62E-18 4.09E-15

[0263] PHEX phosphate regulating endopeptidase homolog X-linked -0.76565 6.29E-17 4.58E-14

[0264] LILRB1 leukocyte immunoglobulin like receptor B1 0.42743 2.83E-16 2.06E-13

[0265] FCGR3AIB Fc fragment of IgG receptor lllb 0.62000 4.14E-16 3.01E-13

[0266] CX3CL1 C-X3-C motifchemokine ligand 1 0.48819 5.11E-16 3.72E-13

[0267] NOS3 nitric oxide synthase 3 0.27418 1.10E-14 7.99E-12

[0268] CLEC4C C-type lectin domain family 4 member C -0.58463 1.47E-14 1.07E-11

[0269] LTB lymphotoxin beta -0.60692 5.47E-14 3.98E-11

[0270] IL1R2 interleukin 1 receptor type 2 0.71641 1.45E-13 1.06E-10

[0271] ROB04 roundabout guidance receptor 4 0.33836 4.62E-13 3.37E-10

[0272] HYAL2 hyaluronidase 2 0.30476 4.77E-13 3.47E-10

[0273] BATF basic leucine zipper ATF-like transcription factor -0.34916 7.95E-13 5.79E-10

[0274] GATA3 GATA binding protein 3 -0.29516 1.02E-12 7.43E-10

[0275] KLF4 Kruppel like factor 4 0.24884 1.80E-12 1.31E-09

[0276] CAV1 caveolin 1 0.22343 1.55E-11 1.13E-08

[0277] Table 7. Top 20 significant genes associated with acute cellular rejection in heart allografts.

[0278] Abbreviations: FDR p-value (p-value adjusted for false discovery rate); FC (fold-change).

[0279] Gene Gene name log2FC p-value FDR p-value

[0280] CCL19 C-C motif chemokine ligand 19 1.38319 3.57E-43 2.71E-40

[0281] LTB lym photoxin beta 0.95729 1.23E-40 9.29E-38

[0282] PTPN7 protein tyrosine phosphatase non-receptor type 7 0.59275 1.24E-26 9.39E-24

[0283] BTLA B and T lymphocyte associated 0.86698 1.24E-25 9.43E-23

[0284] GZMK granzyme K 0.67593 9.19E-25 6.96E-22

[0285] JAK3 Janus kinase 3 0.44338 1.18E-24 8.98E-22

[0286] PHEX phosphate regulating endopeptidase 0.85465 1.67E-23 1.27E-20

[0287] CD28 CD28 molecule 0.61047 6.02E-23 4.57E-20

[0288] CCL18 C-C motif chemokine ligand 18 1.23056 6.10E-23 4.63E-20

[0289] LEF1 lymphoid enhancer binding factor 1 0.73146 8.28E-22 6.27E-19

[0290] TCF7 transcription factor 7 0.69722 3.31E-20 2.51E-17

[0291] B3GAT1 beta-1,3-glucuronyltransferase 1 0.54635 3.38E-20 2.56E-17

[0292] CD8A CD8a molecule 0.56766 5.28E-20 4.00E-17

[0293] CD27 CD27 molecule 0.82868 1.43E-19 1.09E-16

[0294] PECAM1 platelet and endothelial cell adhesion molecule 1 -0.34395 3.83E-19 2.91E-16

[0295] LCK lymphocyte-specific protein tyrosine kinase 0.51589 4.84E-19 3.67E-16

[0296] MMP9 matrix metallopeptidase 9 0.97905 6.51E-19 4.93E-16

[0297] CD3E CD3e molecule 0.55634 1.46E-18 1.11E-15

[0298] LAG3 lymphocyte activating 3 0.61392 1.79E-18 1.36E-15

[0299] MS4A1 membrane spanning 4-domains A1 1.00950 2.26E-18 1.71E-15

[0300] REFERENCES:

[0301] 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.

[0302] 1. McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Bohm M, Burri H, Butler J, Celutkiene J, Chioncel O, Cleland JGF, Crespo-Leiro MG, Farmakis D, Gilard M, Heymans S, Hoes AW, Jaarsma T, Jankowska EA, Lainscak M, Lam CSP, Lyon AR, McMurray JJV, Mebazaa A, Mindham R, Muneretto C, Francesco Piepoli M, Price S, Rosano GMC, Ruschitzka F, Skibelund AK, ESC Scientific Document Group. 2023 Focused Update of the 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: Developed by the task force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC) With the special contribution of the Heart Failure Association (HF A) of the ESC. European Heart Journal 2023;44:3627-3639.

[0303] 2. Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, Deswal A, Drazner MH, Dunlay SM, Evers LR, Fang JC, Fedson SE, Fonarow GC, Hayek SS, Hernandez AF, Khazanie P, Kittleson MM, Lee CS, Link MS, Milano CA, Nnacheta LC, Sandhu AT, Stevenson LW, Vardeny O, Vest AR, Yancy CW. 2022 AHA / ACC / HFSA Guideline for the Management of Heart Failure: Executive Summary: A Report of the American College of Cardiology / American Heart Association Joint Committee on Clinical Practice Guidelines. Journal of the American College of Cardiology 2022;79:1757-1780.

[0304] 3. Hsieh E, Singh TP, Cherikh WS, Harhay MO, Hayes D, Perch M, Potena L, Sadavarte A, Lindblad K, Zuckermann A, Stehlik J. The International thoracic organ transplant registry of the international society for heart and lung transplantation: Thirty-ninth adult heart transplantation report — 2022; focus on transplant for restrictive heart disease. The Journal of Heart and Lung Transplantation 2022;41:1366-1375.

[0305] 4. Coutance G, Ouldamar S, Rouvier P, Saheb S, Suberbielle C, Brechot N, Hariri S, Lebreton G, Leprince P, Varnous S. Late antibody-mediated rejection after heart transplantation: Mortality, graft function, and fulminant cardiac allograft vasculopathy. J Heart Lung Transplant 2015;34:1050-1057.

[0306] 5. Loupy A, Coutance G, Bonnet G, Van Keer J, Raynaud M, Aubert O, Bories M-C, Racape M, Yoo D, Duong Van Huyen J-P, Bruneval P, Taupin J-L, Lefaucheur C, Varnous S, Leprince P, Guillemain R, Empana J-P, Levine R, Naesens M, Patel JK, Jouven X, Kobashigawa J. Identification and Characterization of Trajectories of Cardiac Allograft

[0307] Vasculopathy After Heart Transplantation: A Population-Based Study. Circulation 2020;141:1954-1967.

[0308] 6. Khush KK, Cherikh WS, Chambers DC, Goldfarb S, Hayes D, Kucheryavaya AY, Levvey BJ, Meiser B, Rossano JW, Stehlik J, International Society for Heart and Lung Transplantation. The International Thoracic Organ Transplant Registry of the International Society for Heart and Lung Transplantation: Thirty-fifth Adult Heart Transplantation Report-2018; Focus Theme: Multi organ Transplantation. J Heart Lung Transplant 2018;37: 1155-1168.

[0309] 7. Kobashigawa J, Colvin M, Potena L, Dragun D, Crespo-Leiro MG, Delgado JF, Olymbios M, Parameshwar J, Patel J, Reed E, Reinsmoen N, Rodriguez ER, Ross H, Starling RC, Tyan D, Urschel S, Zuckermann A. The management of antibodies in heart transplantation: An ISHLT consensus document. J Heart Lung Transplant 2018;37:537-547.

[0310] 8. Colvin MM, Cook JL, Chang P, Francis G, Hsu DT, Kiernan MS, Kobashigawa JA, Lindenfeld J, Masri SC, Miller D, O’Connell J, Rodriguez ER, Rosengard B, Self S, White-Williams C, Zeevi A, American Heart Association Heart Failure and Transplantation Committee of the Council on Clinical Cardiology, American Heart Association Heart Failure and Transplantation Committee of the Council on Cardiopulmonary Critical Care, Perioperative and Resuscitation, American Heart Association Heart Failure and Transplantation Committee of the Council on Cardiovascular Disease in the Young, American Heart Association Heart Failure and Transplantation Committee of the Council on Clinical Cardiology, Council on Cardiovascular and Stroke Nursing, American Heart Association Heart Failure and Transplantation Committee of the Council on Cardiovascular Radiology and Intervention, American Heart Association Heart Failure and Transplantation Committee of the Council on Cardiovascular Surgery and Anesthesia. Antibody -mediated rejection in cardiac transplantation: emerging knowledge in diagnosis and management: a scientific statement from the American Heart Association. Circulation 2015;131:1608-1639.

[0311] 9. Huyen J-PDV, Fedrigo M, Fishbein GA, Leone O, Neil D, Marboe C, Peyster E, Thusen J von der, Loupy A, Mengel M, Revelo MP, Adam B, Bruneval P, Angelini A, Miller DV, Berry GJ. The XVth Banff Conference on Allograft Pathology the Banff Workshop Heart Report: Improving the diagnostic yield from endomyocardial biopsies and Quilty effect revisited. American Journal of Transplantation 2020;20:3308-3318.

[0312] 10. Stewart S, Winters GL, Fishbein MC, Tazelaar HD, Kobashigawa J, Abrams J, Andersen CB, Angelini A, Berry GJ, Burke MM, Demetris AJ, Hammond E, Itescu S, Marboe CC, McManus B, Reed EF, Reinsmoen NL, Rodriguez ER, Rose AG, Rose M, Suciu-Focia N,

[0313] Zeevi A, Billingham ME. Revision of the 1990 working formulation for the standardization of nomenclature in the diagnosis of heart rejection. J Heart Lung Transplant 2005;24: 1710-1720.

[0314] 11. Berry GJ, Burke MM, Andersen C, Bruneval P, Fedrigo M, Fishbein MC, Goddard M, Hammond EH, Leone O, Marboe C, Miller D, Neil D, Rassl D, Revelo MP, Rice A, Rene Rodriguez E, Stewart S, Tan CD, Winters GL, West L, Mehra MR, Angelini A. The 2013 International Society for Heart and Lung Transplantation Working Formulation for the standardization of nomenclature in the pathologic diagnosis of antibody-mediated rejection in heart transplantation. J Heart Lung Transplant 2013;32: 1147-1162.

[0315] 12. Crespo-Leiro MG, Zuckermann A, Bara C, Mohacsi P, Schulz U, Boyle A, Ross HJ, Parameshwar J, Zakliczynski M, Fiocchi R, Stypmann J, Hoefer D, Lehmkuhl H, Deng MC, Leprince P, Berry G, Marboe CC, Stewart S, Tazelaar HD, Baron HM, Coleman LC, Vanhaecke J. Concordance among pathologists in the second Cardiac Allograft Rejection Gene Expression Observational Study (CARGO II). Transplantation 2012;94:1172-1177.

[0316] 13. Fishbein MC, Kobashigawa J. Biopsy -negative cardiac transplant rejection: etiology, diagnosis, and therapy. Curr Opin Cardiol 2004;19:166-169.

[0317] 14. Coutance G, Zouhry I, Racape M, Drieux F, Viailly P-J, Rouvier P, Francois A, Chenard M-P, Toquet C, Rabant M, Berry GJ, Angelini A, Bruneval P, Duong Van Huyen J-P. Correlation Between Microvascular Inflammation in Endomyocardial Biopsies and Rejection Transcripts, Donor-specific Antibodies, and Graft Dysfunction in Antibody-mediated Rejection. Transplantation 2022;106:1455-1464.

[0318] 15. Halloran PF, Venner JM, Madill - Thomsen KS, Einecke G, Parkes MD, Hidalgo LG, Famulski KS. Review: The transcripts associated with organ allograft rejection. American Journal of Transplantation 2018;18:785-795.

[0319] 16. Halloran PF, Potena L, Van Huyen J-PD, Bruneval P, Leone O, Kim DH, Jouven X, Reeve J, Loupy A. Building a tissue-based molecular diagnostic system in heart transplant rejection: The heart Molecular Microscope Diagnostic (MMDx) System. J Heart Lung Transplant 2017; 36(11): 1192-1200.

[0320] 17. Loupy A, Duong Van Huyen JP, Hidalgo L, Reeve J, Racape M, Aubert O, Venner JM, Falmuski K, Bories MC, Beuscart T, Guillemain R, Francois A, Pattier S, Toquet C, Gay A, Rouvier P, Varnous S, Leprince P, Empana JP, Lefaucheur C, Bruneval P, Jouven X, Halloran PF. Gene Expression Profiling for the Identification and Classification of Antibody -Mediated Heart Rejection. Circulation 2017;135:917-935.

[0321] 18. Veldman-Jones MH, Brant R, Rooney C, Geh C, Emery H, Harbron CG, Wappett M, Sharpe A, Dymond M, Barrett JC, Harrington EA, Marshall G. Evaluating Robustness and Sensitivity of the NanoString Technologies nCounter Platform to Enable Multiplexed Gene Expression Analysis of Clinical Samples. Cancer Res 2015;75:2587-2593.

[0322] 19. Afzali B, Chapman E, Racape M, Adam B, Bruneval P, Gil F, Kim D, Hidalgo L, Campbell P, Sis B, Duong Van Huyen JP, Mengel M. Molecular Assessment of Microcirculation Injury in Formalin-Fixed Human Cardiac Allograft Biopsies With Antibody-Mediated Rejection. Am J Transplant 2016;17(2):496-505.

[0323] 20. Mengel M, Loupy A, Haas M, Roufosse C, Naesens M, Akalin E, Clahsen-van Groningen MC, Dagobert J, Demetris AJ, Duong van Huyen J-P, Gueguen J, Issa F, Robin B, Rosales I, Von der Thusen JH, Sanchez-Fueyo A, Smith RN, Wood K, Adam B, Colvin RB. Banff 2019 Meeting Report: Molecular diagnostics in solid organ transplantati on-Consensus for the Banff Human Organ Transplant (B-HOT) gene panel and open source multicenter validation. Am J Transplant 2020;20:2305-2317.

[0324] 21. Giarraputo A, Coutance G, Aubert O, Fedrigo M, Mezine F, Zielinski D, Mengel M, Bruneval P, Duong Van Huyen J-P, Angelini A, Loupy A. Banff Human Organ Transplant Consensus Gene Panel for the Detection of Antibody Mediated Rejection in Heart Allograft Biopsies. Transpl Int 2023;36:11710.

[0325] 22. Velleca A, Shullo MA, Dhital K, Azeka E, Colvin M, DePasquale E, Farrero M, Garcia-Guereta L, Jamero G, Khush K, Lavee J, Pouch S, Patel J, Michaud CJ, Shullo MA, Schubert S, Angelini A, Carlos L, Mirabet S, Patel J, Pham M, Urschel S, Kim K-H, Miyamoto S, Chih S, Daly K, Grossi P, Jennings DL, Kim I-C, Lim HS, Miller T, Potena L, Velleca A, Eisen H, Bellumkonda L, Danziger-Isakov L, Dobbels F, Harkess M, Kim D, Lyster H, Peled Y, Reinhardt Z. The International Society for Heart and Lung Transplantation (ISHLT) guidelines for the care of heart transplant recipients. J Heart Lung Transplant 2023;42:el-el41.

[0326] 23. Fedrigo M, Leone O, Burke MM, Rice A, Toquet C, Vernerey D, Frigo A-C, Guillemain R, Pattier S, Smith J, Lota A, Potena L, Bontadini A, Ceccarelli C, Poli F, Feltrin G, Gerosa G, Manzan E, Thiene G, Bruneval P, Angelini A, Duong Van Huyen J-P. Inflammatory cell burden and phenotype in endomyocardial biopsies with antibody-mediated rejection (ACR): a multicenter pilot study from the AECVP. Am J Transplant 2015;15:526-534.

[0327] 24. Reed EF, Rao P, Zhang Z, Gebel H, Bray RA, Guleria I, Lunz J, Mohanakumar T, Nickerson P, Tambur AR, Zeevi A, Heeger PS, Gjertson D. Comprehensive assessment and standardization of solid phase multiplex-bead arrays for the detection of antibodies to HLA. Am J Transplant 2013;13:1859-1870.

[0328] 25. Geiss GK, Bumgamer RE, Birditt B, Dahl T, Dowidar N, Dunaway DL, Fell HP, Ferree S, George RD, Grogan T, James JJ, Maysuria M, Mitton JD, Oliveri P, Osborn JL, Peng T, Ratcliffe AL, Webster PJ, Davidson EH, Hood L, Dimitrov K. Direct multiplexed measurement of gene expression with color-coded probe pairs. Nat Biotechnol 2008;26:317-325.

[0329] 26. Waggott D, Chu K, Yin S, Wouters BG, Liu F-F, Boutros PC. NanoStringNorm: an extensible R package for the pre-processing of NanoString mRNA and miRNA data. Bioinformatics 2012;28:1546-1548.

[0330] 27. Bhattacharya A, Hamilton AM, Furberg H, Pietzak E, Purdue MP, Troester MA, Hoadley KA, Love MI. An approach for normalization and quality control for NanoString RNA expression data. Brief Bioinform 2021;22:bbaa163.

[0331] 28. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014;15:550.

[0332] 29. Stelzer G, Rosen N, Plaschkes I, Zimmerman S, Twik M, Fishilevich S, Stein TI, Nudel R, Lieder I, Mazor Y, Kaplan S, Dahary D, Warshawsky D, Guan-Golan Y, Kohn A, Rappaport N, Safiran M, Lancet D. The GeneCards Suite: From Gene Data Mining to Disease Genome Sequence Analyses. Curr Protoc Bioinformatics 2016;54: 1.30.1-1.30.33.

[0333] 30. UniProt Consortium T. UniProt: the universal protein knowledgebase. Nucleic Acids Res 2018;46:2699.

[0334] 31. Boyle El, Weng S, Gollub J, Jin H, Botstein D, Cherry JM, Sherlock G. GO::TermFinder-- open source software for accessing Gene Ontology information and finding significantly enriched Gene Ontology terms associated with a list of genes. Bioinformatics 2004;20:3710-3715.

[0335] 32. Agbor-Enoh S, Shah P, Tunc I, Hsu S, Russell S, Feller E, Shah K, Rodrigo ME, Najjar SS, Kong H, Pirooznia M, Fideli U, Bikineyeva A, Marishta A, Bhatti K, Yang Y, Mutebi C, Yu K, Kyoo Jang M, Marboe C, Berry GJ, Valantine HA, For the GRAfT Investigators. Cell-Free DNA to Detect Heart Allograft Acute Rejection. Circulation 2021; 143:1184-1197.

[0336] 33. Kobashigawa J, Hall S, Shah P, Fine B, Halloran P, Jackson AM, Khush KK, Margulies KB, Sani MM, Patel JK, Patel N, Peyster E, Kobashigawa J, Hall S, Agbor-Enoh S, Bray R, Cadeiras M, Chang P, Chih S, Eisen H, Farr M, Fine B, Ghafourian K, Halloran P, Jackson AM, Kao A, Khush KK, Madabhushi A, Margulies KB, Sani MM, Patel JK, Patel N, Peyster E, Parwani A, Pinney S, Schlendorf K, Shah K, Shah P, Stehlik J, Tambur A, Zeevi A. The evolving use of biomarkers in heart transplantation: consensus of an expert panel. American Journal of Transplantation 2023:23(6)727-735.

[0337] 34. Khush KK, Patel J, Pinney S, Kao A, Alharethi R, DePasquale E, Ewald G, Berman P, Kanwar M, Hiller D, Yee JP, Woodward RN, Hall S, Kobashigawa J. Noninvasive detection of graft injury after heart transplant using donor-derived cell-free DNA: A prospective multicenter study. Am J Transplant 2019;19:2889-2899.

[0338] 35. Halloran PF, Madill-Thomsen K, Aliabadi-Zuckermann AZ, Cadeiras M, Crespo-Leiro MG, Depasquale EC, Deng M, Gökler J, Kim DH, Kobashigawa J, Macdonald P, Potena L, Shah K, Stehlik J, Zuckermann A. Many heart transplant biopsies currently diagnosed as no rejection have mild molecular antibody -mediated rejection-related changes. J Heart Lung Transplant 2022:41(3):334-344.

[0339] 36. Halloran PF, Famulski KS, Reeve J. Molecular assessment of disease states in kidney transplant biopsy samples. Nat Rev Nephrol 2016;12:534-548.

[0340] 37. Fedrigo M, Berry GJ, Coutance G, Reed EF, Lin C-Y, Giarraputo A, Kransdorf E, Thaunat O, Goddard M, Angelini A, Neil DAH, Bruneval P, Duong Van Huyen J-P, Loupy A, Miller DV. Report of the 2022 Banff Heart Concurrent: Focus on non-human leukocyte antigen antibodies in rejection and the pathology of ‘mixed’ rejection. Am J Transplant 2024;24:533-541.

[0341] 38. Adam B, Afzali B, Dominy KM, Chapman E, Gill R, Hidalgo LG, Roufosse C, Sis B, Mengel M. Multiplexed color-coded probe-based gene expression assessment for clinical molecular diagnostics in formalin-fixed paraffin-embedded human renal allograft tissue. Clin Transplant 2016;30:295-305.

[0342] 39. Randhawa P. The MMDx® diagnostic system: A critical re-appraisal of its knowledge gaps and a call for rigorous validation studies. Clinical Transplantation 2022:36(1 l):el4747.

[0343] 40. Randhawa PS. Caveats in Interpretation of Molecular Diagnostics in Heart Allografts. Transplantation 2024;108:1472-1475.

Claims

CLAIMS:

1. A method of predicting whether a heart transplant recipient has or is at risk of having acute cellular rejection (ACR) comprising:a) quantifying the expression levels of one or more genes selected from the group consisting of CCL19, LTB, PTPN7, BTLA, GZMK, JAK3, PHEX; CD28, CCL18, LEF1, TCF7, CD5, CD8A, CD27, PECAM1, LCK, MMP9, CD3E, LAG3, and MS4A1 in a sample obtained from the recipient;b) implementing an algorithm on data comprising the quantified plurality of gene expression levels so as to obtain an algorithm output; andc) determining the probability of ACR from the algorithm output of step b).

2. The method according to claim 1 wherein the expression level(s) of 1, 2, 3, 4, 5, 6, 7, 8, 9 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 genes among CCL19, LTB, PTPN7, BTLA, GZMK, JAK3, PHEX; CD28, CCL18, LEF1, TCF7, CD5, CD8A, CD27, PECAM1, LCK, MMP9, CD3E, LAG3, and MS4A1 is / are determined in the sample.

3. The method according to claim 2 wherein the expression levels of CCL19, LTB, PTPN7, BTLA, GZMK, JAK3, PHEX; CD28, CCL18, LEF1, TCF7, CD5, CD8A, CD27, PECAM1, LCK, MMP9, CD3E, LAG3, and MS4A1 are determined in the sample.

4. The method according to any one of claims 1 to 2 wherein the expression level of further one or more genes selected from the group consisting of ARG1, B2M, BATF, C1QA, CAV1, CCL3 / L1, CCL5, CD163, CD3D, CD45R0, CD45RB, CD48, CD7, CD8B, CD96, CDH5, CLEC4C, COL1A1, COL3A1, CTLA4, CX3CL1, CX3CR1, CXCL10, CXCL11, EZH2, FAS, FCGR3A / B, FN1, FOXP3, GATA3, GZMA, HLA-DRA, HMOX1, HYAL2, ICAM1, IFI27, IFI44, IFITM3, IGHG1, IGHG2, IGHG3, IGHG4, IGKC, IGLC1, IKZF1, IL1R2, IL2RG, IL6, IL7R, IRF1, ISG15, ITGB2, KLF4, LILRB1, LOX, MX1, MYB, NFKB1, NOS3, OASL, PDPN, PLA1A, PLAT, PTGER4, PTPRC, PTX3, ROBO4, RORA, S100A8, S100A9, SELE, SIRPG, SLAMF7, SMAD3, SOCS3, SPP1, STAT1, TM4SF18, TRAT1, and ZAP70 is determined in the sample.

5. The method according to any one of claims 1 to 4 wherein the expression levels of ARG1, B2M, BATF, BTLA, C1QA, CAV1, CCL18, CCL19, CCL3 / L1, CCL5, CD163, CD27, CD28, CD3D, CD3E, CD45R0, CD45RB, CD48, CD5, CD7, CD8A, CD8B, CD96, CDH5, CLEC4C, COL1A1, COL3A1, CTLA4, CX3CL1, CX3CR1, CXCL10, CXCL11, EZH2, FAS, FCGR3A / B, FN1, FOXP3, GATA3, GZMA, GZMK, HLA- DRA, HMOX1, HYAL2, ICAM1, IFI27, IFI44, IFITM3, IGHG1, IGHG2, IGHG3, IGHG4, IGKC, IGLC1, IKZF1, IL1R2, IL2RG, IL6, IL7R, IRF1, ISG15, ITGB2, JAK3, KLF4, LAG3, LCK, LEF1, LILRB1, LOX, LTB, MMP9, MS4A1, MX1, MYB, NFKB1, NOS3, OASL, PDPN, PECAM1, PHEX, PLA1A, PLAT, PTGER4, PTPN7, PTPRC, PTX3, ROBO4, RORA, S100A8, S100A9, SELE, SIRPG, SLAMF7, SMAD3, SOCS3, SPP1, STAT1, TCF7, TM4SF18, TRAT1, and ZAP70 are determined in the sample.

6. The method according to any one of claims 1 to 5 wherein the sample has been obtained from the transplanted organ, preferably the transplanted heart.

7. The method according to claim 6 wherein the sample resulted from a biopsy performed on the transplanted organ to assess its health and function.

8. The method according to any one of claims 1 to 7 wherein expression levels are determined by using uniquely coded fluorescent probes to quantify gene expression levels in a sample through hybridization and automated imaging.

9. The method according to any one of claims 1 to 8 wherein the algorithm involves applying the determined expression levels of genes to a machine learning classifier designed to diagnose and / or predict the risk of acute cellular rejection (ACR).

10. The method according to claim 9 wherein the machine learning classifier is a supervised molecular classifier.IL A method for selecting a therapeutic regimen for a heart transplant recipient identified as having or being at high risk of ACR comprising:performing the method according to any one of claims 1 to 10, andif it is concluded that the heart transplant recipient has or is at high risk of ACR, then selecting a therapeutic regimen chosen from immunosuppressive treatments,corticosteroids and T cell-depleting treatments, preferably chosen from anti-thymocyte globulin, dexamethasone, prednisone and methylprednisone.

12. A method for monitoring the efficacy of a therapeutic regimen for a heart transplant recipient having or being at risk of ACR comprising:i) performing the method according to any one of claims 1 to 10 with a first sample obtained from the recipient as a first testing,ii) if it is concluded from step i) that the heart transplant recipient has or is at high risk of ACR, then selecting a therapeutic regimen at a first dose, wherein said therapeutic regimen is chosen from immunosuppressive treatments, corticosteroids and T celldepleting treatments, preferably chosen from anti-thymocyte globulin, dexamethasone, prednisone and methylprednisone, andiii) performing the method according to any one of claims 1 to 10 with a second sample obtained from the recipient as a second testing, after a time period after steps i) and ii),wherein if it is concluded from step iii) that the heart transplant recipient does not have or is at low risk of ACR, then a maintenance dose of the therapeutic regimen ii) which is lower than the first dose, is selected.

13. A method for discriminating a responder heart transplant recipient from a non-responder heart transplant recipient having or being at risk of ACR comprising:i) performing the method according to any one of claims 1 to 10 with a first sample obtained from the recipient before any treatment, as a first testing which results in a first probability of ACR,ii) performing the method according to any one of claims 1 to 10 with a second sample obtained from the recipient after a treatment, as a second testing which results in a second probability of ACR,iii) comparing the first and second probabilities of ACR,wherein if it is concluded from step iii) that the recipient does not have or is at low risk of ACR, then concluding that said recipient is a responder,and wherein if it is concluded from step iii) that the recipient has or is at high risk of ACR, then concluding that said recipient is a non-responder.