Assessing the Reliability of MRI and CT Scans for Liver Volume Estimation in Patients Undergoing Liver Transplantation.
Authors
Affiliations (4)
Affiliations (4)
- Department of Radiology, Weill Cornell Medicine.
- Department of Surgery, NYU Langone Health.
- Department of Pathology, Weill Cornell Medicine.
- Department of Radiology, Columbia Vagelos College of Physicians and Surgeons, New York, NY.
Abstract
Magnetic resonance imaging (MRI) and computed tomography (CT) are commonly used to measure organ volumes, but accuracy has not been methodically confirmed. This research seeks to fill this gap by comparing preoperative MRI and CT for liver volumetric characterization to surgical specimen weights as a reference standard. In this cross-sectional study, we retrospectively enrolled subjects who underwent liver explant and had MRI or CT scans within 1 month preoperatively. Organs' labels were corrected by physicians after deep learning model inferences using ITK-SNAP software. Liver surgical weights were collected from pathology reports. Liver volume was estimated as liver weight (g)×a conversion factor (1.2 mL/g), taking into account density, blood loss, and post-explant changes. We excluded subjects who had gallbladder weight reported together with liver weight. For assessing the accuracy of MRI and CT scan volume measurement, we reported the average volume of all available axial, coronal, and sagittal sequences. For MRI, there was no significant difference in the liver volume between sequences; the average volume over all sequences was used as the MRI measurement. There was good agreement between liver explant volume measurements on MRI [23 subjects, 26 exams, 1730 mL (1195, 2108) compared with pathology 1760 mL (1206, 2337); P=0.71] as well as between CT measurements [16 subjects, 23 exams, 1754 mL (1175, 2558) and pathology 1776 mL (1211, 2820); P=0.37]. In 5 subjects who had both MRI and CT (MRI: 1657±533 mL, CT: 1745±591 mL, surgical reference: 1750±679 mL; P=0.96), MRI and CT scans have similar high accuracy. There was a trend toward contrast-enhanced CT scans having greater volume accuracy compared with CT without contrast, with a mean error=6%±4% versus 11%±11%; P=0.29, respectively. Our study shows that MRI and CT scans both have high accuracy for liver volume measurements compared with surgical specimen weights. CT and MRI both accurately measured liver volume before liver explant evaluation.