Asan Medical Center Identifies New Biomarker for Intractable Liver Cancer...Proposes AI-Based Diagnosis and Treatment Strategy

Development of an AI Pathological Diagnostic Model
Potential for Combination Therapy with Cell Division Inhibitors and PARP Inhibitors

A new biomarker associated with poor prognosis has been identified in patients with liver cancer, a disease that is difficult to treat and has low survival rates. The research team also developed an artificial intelligence (AI) diagnostic model capable of identifying this patient group using only pathological tissue, and proposed a therapeutic strategy that combines cell division inhibitors with PARP inhibitors.

On October 6, Asan Medical Center announced that a joint research team led by Professors Juhyun Shim (Department of Gastroenterology) and Changok Sung (Department of Pathology) at Asan Medical Center, together with University of Macau, Pohang University of Science and Technology, and Hanyang University Guri Hospital, have identified that “RB1 bi-allelic inactivation (RB1-Bi)”—where both alleles of the tumor suppressor gene RB1 are either deleted or inactivated—serves as a novel poor prognosis biomarker for liver cancer.

Researchers Identify New Biomarker for Refractory Liver Cancer. From left: Ju-hyun Shim, Department of Gastroenterology, Asan Medical Center; Chang-ok Sung, Department of Pathology, Asan Medical Center; Joongsub Shim, University of Macau; Sanghyun Park, Pohang University of Science and Technology; Jihyun Ahn, Hanyang University Guri Hospital. Asan Medical Center

The team conducted multi-omics analyses, including whole exome sequencing and RNA sequencing, on a total of 561 liver cancer patients—206 patients from Asan Medical Center and 355 patients from The Cancer Genome Atlas (TCGA), a publicly available cancer research database of the U.S. National Cancer Institute. They then validated their results using genetic and single-cell/spatial transcriptomics data from an independent cohort of 450 additional patients.

The analysis revealed that RB1 bi-allelic inactivation was present in approximately 14.6% of all liver cancer patients. This group exhibited lower tumor cell differentiation and faster tumor progression. Furthermore, their risk of death was 3.32 times higher and risk of recurrence was 3.15 times higher than other patients. Based on these findings, the researchers concluded that RB1 bi-allelic inactivation is an independent poor prognosis factor in liver cancer.

The team also developed FR-MIL, a deep learning-based pathological AI model, to identify RB1 bi-allelic inactivation without the need for costly genomic testing. The model predicts RB1 bi-allelic inactivation status using standard stained pathology slide images. In external validation cohorts, it recorded F1 scores ranging from 84.39% to 91.58%.

A therapeutic strategy was also proposed. The team tested 876 drugs on RB1-deficient liver cancer cells and analyzed the drug responses, finding that these cells showed selective responses to drugs that inhibit cell division or disrupt damaged DNA repair. Notably, a synthetic lethality phenomenon was observed in which RB1-deficient cancer cells were selectively eliminated when treated with PARP inhibitors and related drugs.

In cell and animal experiments, the researchers applied a combination therapy with cell division inhibitors and PARP inhibitors. The results showed that tumor suppression effects increased without noticeable systemic side effects.

Professor Shim stated, "It is encouraging that a new breakthrough tailored to genetic characteristics has been developed for patients with intractable liver cancer, for whom treatment options are currently limited and resistance frequently occurs." She added, "If the newly developed AI diagnostic model and combination therapy strategy are implemented in clinical practice, we expect this could help improve patient survival rates and accelerate the advent of precision medicine." The research findings were recently published in the international journal Signal Transduction and Targeted Therapy.

This content was produced with the assistance of AI translation services.