AI-RITMIA · AI-assisted peRsonalized heart-on-chIp as advanced diagnostic Tool for clinical significance assessMent of Inherited cArdiomyopathies

Inherited cardiomyopathies (iCMPs) are a major causeInherited cardiomyopathies (iCMPs) are a major cause of heart disease and account for a significant percentage of deaths in young patients. While recent advances in next-generation sequencing have enabled testing of a large number of genes related to iCMPs, the rate of identification of pathogenic variants is still below 50%. On the other hand, this unprecedented sequencing power has led to an exponential increase in detecting novel unclassified genetic variants (VUS), for which pathogenicity has not been confirmed. VUS uncertainty of diagnosis poses important problems in clinical evaluations, therapeutic decisions and risk-assessment both for patient and family members.The AI-RITMIA Consortium proposes a breakthrough innovation to establish a conclusive link between specific VUSs and CMPclinical phenotypes, yielding to the development of an unprecedented technological platform: a radically new AI-guided patient-specific heart-on-chip diagnostic tool (AI-RITMIA) will be able to interface for the first time three currently diverging sectors: patient specific in vitro modelling, in silico and artificial intelligence (AI) based data handling and clinical evaluation. AI-RITMIA will be the first clinical grade personalized in vitro-in silico platform enabling faster and unbiased correlations of clinical outcomes with functional in vitro profiling for VUS-carrying patients, returning a pathogenicity score directly to the clinicians to unlock informed diagnosis and therapeutic decisions. The achievement of such innovation will leverage on enabling technologies and resources uniquely available in the Consortium, which brings together excellences from the three interfacing worlds: 1) in vitro models based on OoC and microsensors @POLIMI, @BIOMIMX, @NMI and patient-specific iPSCs @ICH; 2) the largest EU cohort of well characterized CMP patients @COCHIN, and 3) clinical grade AI-based data handling interfaces @TUM, @ADVICE.