Novel Machine Learning Model For Sutureless Aortic Valve Sizing
Rafik Margaryan, Nicola Martini, Giacomo Bianchi, Tommaso Gasbarri, Andrea Gori, Marco Solinas.
Ospedale Del Cuore Fondazione 'G. Monasterio', Massa, Italy.
BACKGROUND: Sutureless valves require intraoperative sizing, with devices provided by the manufacturer company with certain recommendations. We aimed to build a computer tomography based sizing model taking into account its hemodynamics after implantation.
METHODS: 145 patients who underwent Corcym Perceval implantation according to valve sizing recommendations and institutional experience (over 1000 implantations).All the patients had aortic valve annulus measurements in MPR mode as described by our group previously. All measurements and hemodynamic measurements were used to build a machine learning model with multiple cross-validation. Then the model used to predict valve size on 'old style sizing' and 'modern style sizing' datasets.
RESULTS: Model was performing with highest accuracy on the same database that was created with. On cross validation on the data which was never seen was performing with 62 % of accuracy, while maintaining area under curve around 0.91. Prediction on the old database demonstrated that all the valves were oversized almost on all the sizes. Based on this a test web application was created for scientific purposes and can be used http://sizeyourvalve.herokuapp.com/ as a free software without any guarantee.
CONCLUSIONS: Current style of sizing is reproducible and can be used as an actual sizing in clinical settings. Modern sizing gives more accurate and better outcomes vs the old way of measuring.
Back to 2023 Abstracts