AI treatment strategy unveiled for intractable liver cancer... Survival gains expected [Health LAB]

[Financial News] A new strategy has been proposed for patients with intractable liver cancer, using artificial intelligence (AI) for rapid pathological diagnosis and personalized targeted treatment.
According to Pohang University of Science and Technology (POSTECH) on the 7th, a joint research team comprising Shim Juhyun of the Department of Gastroenterology and Sung Chang-ok of the Department of Pathology at Asan Medical Center, Professor Shim Jung-seop of the University of Macau, POSTECH Professor Park Sang-hyun, and Ahn Ji-hyun of Hanyang University Guri Hospital identified the complete loss of RB1, a tumor suppressor gene, as a new biomarker for liver cancer. The team also developed an AI-based diagnostic model targeting this biomarker and found that combination therapy using a cell-division inhibitor and a PARP inhibitor has a strong anticancer effect.
The findings were published in the latest issue of Signal Transduction and Targeted Therapy, a leading journal in the field of translational medicine.
The research team said that combination therapies such as atezolizumab and bevacizumab are currently used to treat advanced liver cancer. However, only some patients respond, and resistance eventually develops. Earlier studies of liver cancer patients found mutations in the RB1 gene, which acts as a brake to prevent cells from proliferating excessively and becoming cancerous.
The research team assembled a study group of 561 liver cancer patients: 206 patients from Asan Medical Center and 355 patients in the U.S. National Cancer Institute’s large-scale cancer research database, The Cancer Genome Atlas (TCGA). The team conducted multi-omics profiling—including whole-exome sequencing and RNA sequencing—to integrate and analyze data such as genomic and transcriptomic data. To improve the objectivity and reliability of the study, the team then validated its findings in a new group of 450 patients with varied disease stages and treatment histories, using genomic, single-cell, and spatial transcriptomic data.
The results showed that patients with 'complete RB1 loss (RB1-Bi),' in whom both copies of the RB1 gene were deleted or inactivated, accounted for about 14.6% of all liver cancer cases. This patient group had less differentiated cancer cells and faster tumor progression. Compared with other patients, their risk of death was 3.32 times higher and their risk of recurrence was 3.15 times higher. The findings confirmed that complete RB1 loss is an independent adverse prognostic factor in liver cancer.
The research team developed a deep learning-based pathology AI model (FR-MIL) to rapidly identify this high-risk group in clinical settings without costly, complex genomic analysis. Using only images of routinely stained pathology tissue slides, the model predicts whether a patient has complete RB1 loss. It demonstrated high accuracy in external validation cohorts, achieving F1 scores ranging from 84.39% to 91.58%.
The team also cultured liver cancer cells with damaged RB1 genes (Huh7, PLC/PRF/5, and HepG2) and treated them with 876 different drugs, including epigenetic agents, kinase inhibitors, and highly selective inhibitors. The results showed that RB1-deficient liver cancer cells selectively died when exposed to certain drugs, including PARP inhibitors, that block cell division or the repair of damaged DNA. Normal cells withstood these drugs, but in cancer cells whose RB1 gene was already defective, treatment with the drugs created a second defect alongside the first, causing the cells to die through synthetic lethality.
In experiments using cells and mice, a combination regimen pairing a cell-division inhibitor with a PARP inhibitor produced a powerful anticancer synergy, maximizing tumor suppression without clear systemic side effects such as liver toxicity.
This multidisciplinary, collaborative study is significant because it marks a new turning point, spanning the full process of identifying new targets through omics analysis and developing AI-based pathology diagnosis to providing powerful personalized combination targeted therapies for patients who responded poorly to existing standard anticancer drugs, with a focus on high-risk liver cancer patients whose prognosis had been extremely poor.
Shim Juhyun, a professor in the Department of Gastroenterology at Asan Medical Center, University of Ulsan College of Medicine, said, “It is encouraging that a new breakthrough tailored to the genetic characteristics of patients with intractable liver cancer—who currently have limited treatment options and are prone to developing resistance—has been established. If the AI diagnostic model and combination treatment strategy developed in this study are introduced into clinical practice, we expect them to make a major contribution to dramatically improving patients’ survival rates and advancing precision, personalized medicine.”
[email protected] Yeon Jian Reporter