Borders as opposed to so-called geography: which should be used to classify isolated ventricular septal defects?

Ventricular septal defects can be classified according to their borders or according to the fashion in which they open to the right ventricle, so-called geography. As yet, there is no consensus as to how they should be classified. In an attempt to achieve agreement, the International Society for Nomenclature of Congenital and Paediatric Heart Disease, in 2018, proposed a system incorporating both approaches. We have assessed the subjectivity of their suggested terms hoping to determine their suitability in the desired universal system for classification.

We examined 212 specimens held in the archive of Birmingham Women's and Children's Hospital. Each defect was described by 3 independent examiners on the basis of borders and their relationship to the landmarks of the right ventricle. The interobserver agreement was then calculated using Fleiss' method.

Calculations to assess interobserver agreement showed that the examiners were more likely to agree on the borders of the defects than their so-called geography (κ = 0.804 vs κ = 0.518). The landmarks of the right ventricle proved to be highly variable such that the application of 'geographic' terms to hearts with perimembranous defects proved particularly challenging.

Interobserver agreement is lower when using terms based on 'geography' as opposed to borders. Whilst providing important morphological detail, the terms based on right ventricular landmarks are highly subjective. They should not be prioritized in a universal system of classification. Instead, the defects can be classified simply by using 'perimembranous', 'muscular', or 'doubly committed and juxta-arterial' as first-order terms.