What question did this study set out to answer?

To develop deep learning models for segmenting hepatoblastoma in pediatric CT and MRI scans.

February 22, 2026Radiology Imaging Cancer1 citationsOpen Access

Pediatric Personalized Deep Learning Models for Segmentation of Hepatoblastoma at CT and MRI

GMGourav ModanwalCardiac Imaging SKSaurabh KumarThe Wallace H. Coulter Department of Biomedical Engineering VVVidya Sankar ViswanathanEmory University

Key Points

To develop deep learning models for segmenting hepatoblastoma in pediatric CT and MRI scans.
Trained deep learning models using pediatric CT and MRI scans.
Compared model performance with human annotators and their consensus.
Evaluated a pediatric contrast-enhanced CT model against an adult CT-trained model.
The pediatric models showed good agreement with human annotators.
The pediatric CT model outperformed the adult CT-trained model in segmentation tasks.

Abstract

Deep learning–based hepatoblastoma segmentation models trained on pediatric CT and MRI scans demonstrated good agreement with human annotators and their consensus. Pediatric contrast-enhanced CT model outperformed the adult CT-trained model.

AI에게 질문

Bookmark

View Full Paper