Back to 2024 Abstracts
Chss Avsd Cohort: Investigating Surgical Repair Strategies Utilizing Baseline Echocardiographic Data As Surrogate Measures
Justin Robinson, MD1, Eugene Blackstone, MD
1, Hani Najm, MD, MSc
1, Xiawen Li, MS
1, David Overman, MD
2, Luc Mertens, MD
3, Tara Karamlou, MD, MSc
1.
1Cleveland Clinic, Cleveland, OH, USA,
2Children's Hospitals and Clinics of Minnesota, Minneapolis, MN, USA,
3The Hospital for Sick Children in Toronto, Toronto, ON, Canada.
Objective(s):Evaluate whether echocardiographic measurements can serve as reliable markers for predicting repair strategies in the diverse spectrum of atrioventricular septal defects (AVSD).
Methods:From 2012, to 2022, 889 babies with complete AVSD were enrolled at one of 31 participating Congenital Heart Surgeon Society hospitals. Each baby had pre-intervention echocardiogram that was quantitively analyzed. Of these, 885 had an institutional diagnosis of unbalanced vs. balanced AVSD. Newer random forests unsupervised machine learning variable selection methods of cardiac morphologic features were utilized.
Results:Among 885 institutionally diagnosed infants, variables demonstrating highest predictive power for determining repair strategy included atrioventricular valve index (AVVI), minor index, ventricular index, angle between right ventricular inflow and left ventricular inflow, atrial septal defect (ASD) size, and common atrioventricular (AV) valve diameter. Pulmonary artery banding was observed across a wide range of values for all pre-intervention echocardiographic measurements, with the exception of the common AV valve diameter.
Conclusions:
Surgical decision-making is multifactorial and must be considered within the broader context of individual patient factors. Previous studies have reported that an AVVI < 0.67 is considered a factor for biventricular repair in right dominant complete AVSD. However, our study found significant heterogeneity in surgical strategy with an AVVI value < 0.67. Our study demonstrates that surgical strategies may vary significantly across institutions.
Back to 2024 Abstracts
