Abstract
Hepatocellular carcinoma (HCC) is the primary form of liver cancer and a major cause of cancer death worldwide. Early detection is key to effective treatment. Yet, early diagnosis is challenging, especially in patients with cirrhosis, who are at high risk of developing HCC. Dysfunction or loss of function of the transforming growth factor β (TGF-β) pathway is associated with HCC. Here, using quantitative immunohistochemistry analysis of samples from a multi-institutional repository, we evaluated if differences in TGF-β receptor abundance were present in tissue from patients with only cirrhosis compared with those with HCC in the context of cirrhosis. We determined that TGFBR2, not TGFBR1, was significantly reduced in HCC tissue compared with cirrhotic tissue. We developed an artificial intelligence (AI)-based process that correctly identified cirrhotic and HCC tissue and confirmed the significant reduction in TGFBR2 in HCC tissue compared with cirrhotic tissue. Thus, we propose that a reduction in TGFBR2 abundance represents a useful biomarker for detecting HCC in the context of cirrhosis and that incorporating this biomarker into an AI-based automated imaging pipeline could reduce variability in diagnosing HCC from biopsy tissue.