Identification of Predominant Histopathological Growth Patterns of Colorectal Liver Metastasis by Multi-Habitat and Multi-Sequence Based Radiomics Analysis.

摘要: Purpose: Developing an MRI-based radiomics model to effectively and accurately predict the predominant histopathologic growth patterns (HGPs) of colorectal liver metastases (CRLMs). Materials Methods: In this study, 182 resected histopathological proven CRLMs chemotherapy-naive patients from two institutions, including 123 replacement 59 desmoplastic CRLMs, were retrospectively analyzed. Radiomics analysis was performed on regions interest (ROI), tumor zone tumor-liver interface (TLI) zone. Decision tree (DT) algorithm used for modeling each MR sequence, fused constructed by combining signature sequence. The clinical combination models developed through multivariate logistic regression method. performance assessed receiver operating characteristic (ROC) curves with indicators area under curve (AUC), accuracy, sensitivity, specificity. A nomogram evaluate discrimination, calibration, usefulness. Results: radiomicstumor radiomicsTLI showed better than any single sequence model. addition, exhibited (AUC 0.912 vs. 0.879) in internal validation cohort. good AUC 0.971, 0.909, 0.905 training, validation, external cohorts, respectively. Conclusion: method has high potential predicting HGPs CRLM. Preoperative non-invasive identification could further explore ability as a biomarker treatment strategy, reflecting different biological pathways.