Artificial intelligence (AI)-driven deep learning has fundamentally transformed hepatosplenic volumetry, elevating it from a labor-intensive research procedure to a scalable, clinically deployable biomarker. This narrative review synthesizes the current evidence on automated liver and spleen volumetry derived from abdominal CT and MRI. The discussion encompasses technical developments, normative reference intervals, and clinical applications—including hepatic fibrosis staging, portal hypertension assessment, prediction of liver-related outcomes, and preoperative risk stratification for post-hepatectomy liver failure. We have further examined the critical limitations, particularly the restricted inter-vendor and cross-institutional generalizability and absence of prospective validation, and delineate possible future directions in segmentation refinement, multimodal data integration, and opportunistic screening pipelines.
Heo et al. (Tue,) studied this question.