During the Saturday, Jan. 31, 7:50 a.m. presentation of the James S. Tweddell Memorial Paper for Congenital Heart Surgery, Elaine Griffeth, MD, of Mayo Clinic will present new research during the “Research in Focus: Distinguished Abstracts” session at the 2026 Society of Thoracic Surgeons (STS) Annual Meeting. Her talk, Extended Validation of an Institutional Machine Learning Model for Postoperative Morbidity and Mortality Risk in Adult Congenital Heart Disease Patients Undergoing Cardiac Reoperation, explores how advanced risk modeling can better inform surgical decision-making for adults with congenital heart disease (CHD).
Adults with CHD represent a growing and medically complex population. Most were born with structural heart defects and underwent surgery early in life, yet many require additional cardiac operations as adults. Prior surgeries, evolving anatomy, and long-term health challenges make it difficult to accurately estimate operative risk using existing tools designed for the broader adult cardiac surgery population, highlighting the need for a CHD-specific national risk assessment model.
The study analyzed cases from the STS Adult Cardiac Surgery Database spanning several years, building on prior Mayo Clinic work using machine learning and logistic regression. Seven factors were strongly associated with postoperative morbidity and mortality: sex, age, single-ventricle physiology, surgical urgency, kidney function, ejection fraction, and prior heart operations.
“This is a work in progress,” says Dr. Griffeth. “We want to have high reliability in the surgeries we are offering, and we are trying to tailor this model with data from past patients. The more informed patients are about their risks for surgery, the better.”