Primary liver cancer arises either from hepatocytic or biliary lineage cells, giving rise to hepatocellular carcinoma (HCC) or intrahepatic cholangiocarcinoma (ICCA). Combined hepatocellular- cholangiocarcinomas (cHCC-CCA) exhibit equivocal or mixed features of both, causing diagnostic uncertainty and difficulty in determining proper management. Here, we perform a comprehensive deep learning-based phenotyping of multiple cohorts of patients. We show that deep learning can reproduce the diagnosis of HCC vs. CCA with a high performance. We analyze a series of 405 cHCC-CCA patients and demonstrate that the model can reclassify the tumors as HCC or ICCA, and that the predictions are consistent with clinical outcomes, genetic alterations and in situ spatial gene expression profiling. This type of approach could improve treatment decisions and ultimately clinical outcome for patients with rare and biphenotypic cancers such as cHCC-CCA.

Individuals with alcohol use disorder (AUD) are at risk of liver disease. There is scarce information on the effectiveness of screening for liver fibrosis on alcohol consumption. Thus, we evaluated the efficacy of a screening program for liver fibrosis on alcohol consumption in individuals with AUD. We performed a prospective interventional study in the Hospital Clinic of Barcelona. The screening cohort included individuals with AUD from the addiction unit who underwent screening for liver fibrosis with transient elastography and counselling on lifestyle habits in the liver unit. The control cohort included individuals with similar characteristics who attended the same unit in a previous period but did not undergo screening. Effects on alcohol consumption were evaluated at 6 months, after clinical follow-up, with clinical assessment by addiction specialists and urine ethyl glucuronide monitoring. In the screening cohort, 149/334 (45%) individuals were abstinent at 6 months (68% confirmed with urine ethyl glucuronide). Alcohol abstinence was higher in the screening cohort than in the control cohort (40/137 [29%], p = 0.002). Factors associated with alcohol abstinence in the multivariate analysis of the two combined cohorts (n = 471) were: receiving AUD medications (odds ratio [OR] 1.72, 95% CI 1.11-2.67), absence of illicit drug use (OR 0.50, 95% CI 0.31-0.80) and participating in the screening program (OR 1.77, 95% CI 1.14-2.74). In the screening cohort, 40 (12%) individuals had increased liver stiffness (≥8 kPa), which was associated with obesity (p = 0.03), arterial hypertension (p = 0.03), gamma-glutamyltransferase (p <0.001) and platelet levels (p = 0.001). This study shows that an integrated screening program for liver fibrosis associated with counselling on alcohol consumption in individuals with AUD allows for early diagnosis of alcohol-associated liver disease and is associated with alcohol abstinence. Individuals with high alcohol consumption are at higher risk of liver disease compared to the general population. The potential beneficial effects of screening for liver disease in this population have scarcely been studied. We show that a screening program for liver fibrosis together with a lifestyle counselling intervention favoured alcohol abstinence among individuals with alcohol use disorder attending an addiction unit at 6 months, compared to a matched cohort who did not undergo screening. These findings suggest that screening programs for liver fibrosis have a therapeutic role in individuals with alcohol use disorder, supporting the implementation of these programs in addiction units.