https://www.idival.org/wp-content/uploads/2022/05/cab-presentacion.jpg

Researchers from IDIVAL and the University of Cantabria develop an AI that predicts cardiac recovery after valve surgery with 90% accuracy

The project Transcriptomics and artificial intelligence for predicting adverse cardiac remodeling in patients undergoing surgery for aortic valve stenosis, funded by the INNVAL 2021 call from the Marqués de Valdecilla Research Institute (IDIVAL), has focused on studying omics data from left ventricle biopsies of patients with severe aortic stenosis who underwent valve replacement surgery at the Marqués de Valdecilla University Hospital.

The team led by Dr. J. Francisco Nistal and Dr. Raquel García from the IDIVAL research group on Cytokines and Pathological Tissue Plasticity Phenomena, in collaboration with Dr. Joaquín Bedia from the Department of Applied Mathematics and Computer Science at the University of Cantabria, has designed a predictive model that, based on the patient’s sex, distinguishes between those who will normalize left ventricular mass one year after valve replacement and those who will retain residual hypertrophy. To achieve this, they applied machine learning methods—a branch of artificial intelligence with extensive applications in medical diagnosis and prognostic estimation. The predictive models developed using Bayesian networks, which combined miRNome sequencing data and clinical data, made it possible to identify a sex-specific profile of deregulated microRNAs that enables the estimation of residual hypertrophy one year after surgery with 90% accuracy.

This project deepens the identification of genetic markers, alongside clinical markers, with diagnostic, prognostic, and therapeutic stratification value from a gender perspective. This contributes to improving both the indication for and the timing of surgery in patients with aortic stenosis. Stratifying and predicting which patients will normalize their left ventricular mass after surgery is a significant clinical challenge, as the persistence of hypertrophy after surgery negatively impacts both short- and long-term survival and functional recovery.

The project, titled Transcriptomics and artificial intelligence for predicting adverse cardiac remodeling in patients undergoing surgery for aortic valve stenosis, is part of the INN-VAL innovation project support program (Ref. INNVAL21/24) of the Marqués de Valdecilla Research Institute (IDIVAL).