Prediction of Micrometastases in Colorectal Cancer Liver Metastasis Using Habitat Analysis Based on Multi-parametric MRI

Motivation: Micrometastases in colorectal cancer liver metastasis are difficult to detect preoperatively, leading to positive margins and poor prognosis. A more precise detection method is urgently needed. Goal(s): To accurately predict micrometastases by comparing habitat analysis based on multimodal MRI with whole-tumor radiomics analysis. Approach: Multi-parametric MRI was used for detailed subregional radiomics analysis in patients, compared with whole-tumor analysis. Results: Habitat analysis showed significant advantages in predicting micrometastases, potentially reducing positive margin risks and improving prognosis. Impact: This study provides a powerful tool for preoperative prediction, optimizing surgical strategies and enhancing patient quality of life.