Predicting Hepatocellular Neoplasm - Not Otherwise Specified (HCN-NOS)

Abstract:

Hepatocellular neoplasm-not otherwise specified (HCN-NOS) is increasingly recognized as a subset of hepatoblastoma (HB) with poor prognosis, characterized by pleomorphism and/or hepatocellular carcinoma-like features. Identifying morphologic features of HCN-NOS is a challenge and subject to inter-observer variability. We benchmark deep learning (DL) model to predict HCN-NOS using hematoxylin-eosin-stained (H&E) whole-slide images (WSIs) of pre-therapy tumor samples (biopsies or upfront resections). This work lays the foundation for the construction of a multi-modal risk stratification algorithm for pediatric liver tumors, including artificial intelligence-based features

References:

S. Kharangate, C. Woods, A. Schirmer, B. Huey, S. Prasath, S. Ranganathan, K. R. Patel. A deep learning model to predict hepatocellular neoplasm-not otherwise specified (HCN-NOS). Society for Pediatric Pathology Fall Meeting, Columbus, OH, USA, October 2026.

S. Kharangate, C. Woods, A. Schirmer, B. Huey, S. Prasath, S. Ranganathan, K. R. Patel. Comparative analysis of artificial intelligence models for predicting hepatocellular neoplasm-not otherwise specified (HCN-NOS). United States and Canadian Academy of Pathology's 116th Annual Meeting (USCAP), Vancouver, BC, Canada, March 2027.

Back to Histopathology Projects Main Page. Back to Research Page.