Robust multicentre detection and classification of colorectal liver metastases on CT: application of foundation models

Colorectal liver metastases (CRLM) remain a major cause of cancer-related mortality. Early, reliable detection on CT imaging is essential for curative treatment planning, yet the diagnostic performance of AI models often declines across scanners and institutions, limiting clinical generalizability. In this study, we developed and evaluated a multi-center foundation-model-based AI pipeline for patient-level classification and lesion-level detection of CRLM on contrast-enhanced CT. Using data from the EuCanImage consortium (n = 2437) and TCIA_CRLM (n = 197, all CRLM), we benchmarked several pretrained foundation models and identified UMedPT as the optimal encoder. The final model achieved an AUC of 0.89 and a sensitivity of 0.82 on the EuCanImage test set, with a sensitivity of 0.85 on the external TCIA cohort. Excluding the most uncertain 20% of cases improved AUC to 0.90 and balanced accuracy to 0.85. Decision-curve analysis indicated superior net benefit over "treat-all" and "treat-none" strategies for threshold probabilities between 0.35 and 0.75. The lesion-level detector identified 69.1% of lesions overall, increasing substantially with lesion size. Grad-CAM maps showed strong correspondence between attention regions and metastases in high-confidence predictions. These findings demonstrate that foundation-model-based pipelines enable robust, generalizable, and interpretable solutions for CRLM detection and classification in multi-center CT imaging.

Authors

Institutions

Publication Details

Journal
npj Precision Oncology
Published
2026-05-27
DOI
https://doi.org/10.1038/s41698-026-01482-2
Primary Topic
Radiomics and Machine Learning in Medical Imaging
Type
article
Field-Weighted Citation Impact
0.00

Funders

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

Robust multicentre detection and classification of colorectal liver metastases on CT: application of foundation models

Henry C. Woodruff, Melanie Goisauf, Maciej Bobowicz, Zohaib Salahuddin et al.
npj Precision Oncology
Radiomics and Machine Learning in Medical Imaging
article

Robust multicentre detection and classification of colorectal liver metastases on CT: application of foundation models

Henry C. Woodruff, Melanie Goisauf, Maciej Bobowicz, Zohaib Salahuddin, Samuele Ghezzo, André Aichert, Lidia Garrucho, Jordi Rimola, Razvan L. Miclea, Emanuele Neri, Roberto Francischello, Lorenzo Faggioni, Davide Zaccagnini, Shruti Atul Mali, Karim Lekadir, Josep Lluís Gelpí, Jeroen Cant, J. Smith, Pilar Nicolás, Pär Kragsterman, Jordi Rambla, Salvador Capella-Gutierrez, Katrine Riklund, Fred Prior, Yumeng Zhang, Philippe Lambin, Xian Zhong, Henry Woodruff, Martijn Starmans, Peter Gordebeke, Juozas Kupčinskas, Javier del Riego, On behalf of EUCanImage working group, Xavier Bargalló, Tobias Heimann
article en

Abstract

Colorectal liver metastases (CRLM) remain a major cause of cancer-related mortality. Early, reliable detection on CT imaging is essential for curative treatment planning, yet the diagnostic performance of AI models often declines across scanners and institutions, limiting clinical generalizability. In this study, we developed and evaluated a multi-center foundation-model-based AI pipeline for patient-level classification and lesion-level detection of CRLM on contrast-enhanced CT. Using data from the EuCanImage consortium (n = 2437) and TCIA_CRLM (n = 197, all CRLM), we benchmarked several pretrained foundation models and identified UMedPT as the optimal encoder. The final model achieved an AUC of 0.89 and a sensitivity of 0.82 on the EuCanImage test set, with a sensitivity of 0.85 on the external TCIA cohort. Excluding the most uncertain 20% of cases improved AUC to 0.90 and balanced accuracy to 0.85. Decision-curve analysis indicated superior net benefit over "treat-all" and "treat-none" strategies for threshold probabilities between 0.35 and 0.75. The lesion-level detector identified 69.1% of lesions overall, increasing substantially with lesion size. Grad-CAM maps showed strong correspondence between attention regions and metastases in high-confidence predictions. These findings demonstrate that foundation-model-based pipelines enable robust, generalizable, and interpretable solutions for CRLM detection and classification in multi-center CT imaging.

npj Precision Oncology
University of Pisa (IT), Lithuanian University of Health Sciences (LT), University of Nottingham (GB), University of Liège (BE), Sun Yat-sen University (CN), University of the Basque Country (ES), Barcelona Supercomputing Center (ES), Maastricht University Medical Centre (NL), Erasmus MC (NL), European Society of Radiology (AT), Maastricht University (NL), Biobanking and Biomolecular Resources Research Infrastructure Consortium (AT), Hospital Clínic de Barcelona (ES), Lynkeus (Italy) (IT), The First Affiliated Hospital, Sun Yat-sen University (CN), European Institute for Biomedical Imaging Research (AT), Institute of Research and Innovation Parc Tauli (ES), Centre for Genomic Regulation (ES), Camber Collective (United States) (US), Siemens Healthineers (Germany) (DE), Maastro Clinic (NL), Universitat de Barcelona (ES), Gdańsk Medical University (PL), University of Arkansas for Medical Sciences (US), Umeå University (SE), Universitat Politècnica de Catalunya (ES)
European Commission, European Research Council, HORIZON EUROPE European Research Council
Good health and well-being
Openalex Percentile: Top 7%
Radiomics and Machine Learning in Medical Imaging
DOIOpen Access PDFOpenAlex
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.