Multimodal model for the diagnosis of biliary atresia based on sonographic images and clinical parameters

It is still challenging to diagnose biliary atresia (BA) in current clinical practice. The study aimed to develop a multimodal model incorporated with uncertainty estimation by integrating sonographic images and clinical information to help diagnose BA. Multiple models were trained on 384 infants and validated externally on 156 infants. The model fused with sonographic images and clinical information yielded best performance, with an area under the curve (AUC) of 0.941 (95% CI: 0.891-0.972) on the external dataset. Moreover, the model based on sonographic video still yielded AUC of 0.930 (0.876-0.966). By excluding 39 cases with high uncertainty (>0.95), accuracy of the model improved from 84.6% to 91.5%. In addition, six radiologists with different experiences showed improved diagnostic performance (mean AUC increase: 0.066) when aided by the model. This fusion model incorporated with uncertainty estimation could potentially help radiologists identify BA more accurately and efficiently in real clinical practice.

Authors

Institutions

Publication Details

Journal
npj Digital Medicine
Published
2025-06-17
DOI
https://doi.org/10.1038/s41746-025-01694-z
Citations
3
Primary Topic
Pediatric Hepatobiliary Diseases and Treatments
Type
article
Field-Weighted Citation Impact
6.02

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Multimodal model for the diagnosis of biliary atresia based on sonographic images and clinical parameters

Hualin Yan, Juxian Liu, Yuanhang Zheng, Wenying Zhou et al.
3 citations
npj Digital Medicine
Pediatric Hepatobiliary Diseases and Treatments
6.02
article

Multimodal model for the diagnosis of biliary atresia based on sonographic images and clinical parameters

Hualin Yan, Juxian Liu, Yuanhang Zheng, Wenying Zhou, Luyao Zhou, Zijian Tang, Wen Ling, Run Lin, Shan Wang, Ruixuan Wang, Cheng Yu, Bin Xu, Zongjie Weng, Guangliang Huang
article en
3 citations

Abstract

It is still challenging to diagnose biliary atresia (BA) in current clinical practice. The study aimed to develop a multimodal model incorporated with uncertainty estimation by integrating sonographic images and clinical information to help diagnose BA. Multiple models were trained on 384 infants and validated externally on 156 infants. The model fused with sonographic images and clinical information yielded best performance, with an area under the curve (AUC) of 0.941 (95% CI: 0.891-0.972) on the external dataset. Moreover, the model based on sonographic video still yielded AUC of 0.930 (0.876-0.966). By excluding 39 cases with high uncertainty (>0.95), accuracy of the model improved from 84.6% to 91.5%. In addition, six radiologists with different experiences showed improved diagnostic performance (mean AUC increase: 0.066) when aided by the model. This fusion model incorporated with uncertainty estimation could potentially help radiologists identify BA more accurately and efficiently in real clinical practice.

npj Digital MedicineVol. 8(1)
Fujian Medical University (CN), Sun Yat-sen University (CN), Sichuan University (CN), West China Medical Center of Sichuan University (CN), West China Hospital of Sichuan University (CN), Fuzhou Maternity and Child Health Care Hospital (CN), The First Affiliated Hospital, Sun Yat-sen University (CN), Wuhan Union Hospital (CN), Union Hospital (CN)
National Science Foundation, National Natural Science Foundation of China, China Postdoctoral Science Foundation, Major Research Plan
Openalex Percentile: Top 4%
Pediatric Hepatobiliary Diseases and Treatments
6.02
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.