Pericardial segmentation: a proposed model to quantify disease burden

Cardiac MR (CMR) provides a comprehensive assessment of pericardial structure, tissue characteristics and constriction-related haemodynamics, making it an excellent tool for assessing the complex mechanisms that underlie constrictive pericarditis. To facilitate the reproducible quantification of pericardial disease burden, this article introduces the first standardised pericardial segmentation model validated against anatomical specimens, obtainable using standard CMR sectioning planes, and designed with equal weighting of individual pericardial segments for ease of use.

The model was created by assessing 100 morphologically normal forensic cardiac specimens with an equal gender distribution and a broad age range. Direct measurements of left ventricular (LV) and right ventricular (RV) surface areas were obtained on standard cardiac short-axis forensic dissection slices. To assess variability in respective LV and RV segment sizes, data from the 100 specimens were compared with an idealised model developed to have identical respective LV and RV segment sizes.

On average, the LV and RV contributed similar areas (49.9% and 50.1%, respectively) of the total ventricular surface area. The LV pericardial area was well represented by those 11 segments of the 17-segment American Heart Association model that borders pericardium (measuring 4.51%±0.2% of the total pericardial area, per segment). The RV surface area was best represented by nine novel segments (measuring 5.54%±0.3% of the total pericardial area, per segment). A correlation between the measured and idealised models showed a mean difference of only 0.04% and 0.02% per segment for the LV and RV, respectively.

This pericardial segmentation model, validated against anatomical specimens and obtainable using only standard CMR views, incorporates equal-sized LV and RV pericardial segments to ensure clinical usability. By enabling quantification of disease distribution and burden across both ventricles, the model has the potential to improve clinical decision-making and enhance precision in pericardial research.