AI-Driven MRI Radiomics Predicts Liver Cancer Therapy Response
June 1, 2025
AI-powered MRI radiomics significantly improves prediction of treatment response in advanced liver cancer patients.
Key Details
- Multicenter study presented at ASCO 2025 focused on advanced hepatocellular carcinoma (HCC).
- AI-based radiomics model analyzed MRI data to predict response to atezolizumab and bevacizumab therapy.
- Study included 240 patients; training cohort of 161, validation cohort of 79.
- Radiomics model achieved AUC of 0.913 (training) and 0.825 (validation); combined with a key MRI feature, AUC increased to 0.951 and 0.835, respectively.
- Significant correlation found between radiomic and conventional MRI features for intrahepatic lesions.
Why It Matters
This work demonstrates the power of AI-driven radiomics to personalize cancer treatment planning, improve patient selection for liver cancer therapies, and highlights growing integration of advanced imaging analytics in clinical oncology workflows.