Noninvasive estimation of heart organ mass by postmortem magnetic resonance imaging and computed tomography compared to conventional autopsy in fetuses: a pilot study

Abstract

Objectives

The purpose of this pilot study is to investigate the accuracy of mass estimation of the heart using postmortem MRI and CT in this population.

Methods

The mass of the heart was estimated using three different methods: sequential segmentation of the right and left ventricles by MRI in 30 patients; sequential segmentation of the left ventricle chamber using CT and an ellipsoid method of the left ventricle (LV) using CT in 6 patients. The estimated mass by each method was compared to conventional autopsy mass using linear regression.

Results

The correlation between estimated and measured cardiac mass using postmortem MRI measurements of the right and left ventricles, CT derived volumes of the left ventricle, or by the LV ellipsoid CT method are as follows: y = 0.87744x - 0.74813 (R² = 0.958); y = 0.6004x - 2.16 (R² = 0.8681); y = 0.6019 - 2.1615 (R² = 0.8599), respectively. When separating the fetuses evaluated by pmMRI into a ≤20 weeks subset or >20 weeks subset, the correlation are as follows: y = 0.5176x + 0.2428 (R² = 0.645) and y = 0.91237x - 1.53647 (R² = 0.9475), respectively.

Conclusion

Mass of the heart at autopsy modestly correlates with MRI derived volumes measuring the right and left ventricles. The small sample in the CT derived volume of the left ventricle alone or by the simplified estimated method using the ellipsoid measurements of the left ventricle by CT compared to autopsy weights make it difficult to conclude CT's utility using these methods.