%0 Journal Article
%A Rabasco Meneghetti, Asier
%A Campani, Claudia
%A Roux, Charles
%A Carrero, Zunamys Itzel
%A Popica, Dan-Adrian
%A Amaddeo, Giuliana
%A Lequoy, Marie
%A Hollande, Clémence
%A Mouri, Sarah
%A Wagner, Mathilde
%A Plaforet, Vincent
%A Sidali, Sabrina
%A Ronot, Maxime
%A Rudler, Marika
%A Luciani, Alain
%A Sutter, Olivier
%A Spitzer, Eleonore
%A Regnault, Hélène
%A El Mouhadi, Sanaâ
%A Ozenne, Violaine
%A Ganne-Carrié, Nathalie
%A Bouattour, Mohamed
%A Nault, Jean Charles
%A Thabut, Dominique
%A Kather, Jakob Nikolas
%A Allaire, Manon
%T Detection of Esophageal Varices and Prediction of Hepatic Decompensation in Unresectable Hepatocellular Carcinoma using AI: AI Detection of Varices and Decompensation.
%J Journal of hepatology
%V nn
%@ 0168-8278
%C Amsterdam [u.a.]
%I Elsevier Science
%M DKFZ-2026-00359
%P nn
%D 2026
%Z #DKTKZFB9# / #NCTZFB9# / epub
%X In hepatocellular carcinoma (HCC) with cirrhosis, portal hypertension (PHT) worsens outcomes. Esophagogastroduodenoscopy (EGD), current screen for esophageal varices (EVs), is invasive and delays therapy. We aimed to develop and externally validate non-invasive models to detect EV and predict hepatic decompensation (bleeding, ascites or hepatic encephalopathy), a major cause of HCC mortality, using routine contrast-enhanced CT (CECT) and clinical data.This multicenter retrospective study included 489 patients with unresectable HCC treated with Atezolizumab-Bevacizumab (AtezoBev) from five French centers, split into a development (n=279) and external validation (n=210) cohorts. Arterial-phase CECTs were processed through a Deep Learning pipeline using a foundation model (HepatoSageCT). Logistic and Cox models generated clinical models and combined models integrating the HepatoSageCT scores with key clinical variables for EVs and hepatic decompensation. Performance was assessed using AUCROC, sensitivity, specificity, concordance index and cause-specific hazard ratio.Portosystemic shunts (PSS) at imaging identified EVs with AUCROC of 0.78, increasing to 0.84 when combined with HepatoSageCT. A decision algorithm incorporating PSS and HepatoSageCT missed 4.2
%K deep learning (Other)
%K hepatic decompensation (Other)
%K hepatocellular carcinoma (Other)
%K portal hypertension (Other)
%K portosystemic shunt (Other)
%K varices (Other)
%F PUB:(DE-HGF)16
%9 Journal Article
%$ pmid:41679555
%R 10.1016/j.jhep.2026.01.021
%U https://inrepo02.dkfz.de/record/309874