Objective To explore the predictive value of multi-parameter 3D MRI radiomics model for colorectal cancer liver metastases(CCLM)and the analysis of related clinical risk factors.Methods From January 2019 to June 2022,the clinical and imaging data of 198 patients with colorectal cancer were collected from our hosipital who underwent preoperative MRI examination and were confirmed by postoperative pathology.It was divided into training set(n=159)and verification set(n=39)according to the ratio of 8:2.The T2WI,DWI and e-THRIVE+sequences were delineated layer by layer to obtain 3D VOI,and the omics features were extracted and dimensionality reduction was performed to construct the radiomics model.The independent risk factors of CCLM were screened out by logistic regression analysis and the clinical model was constructed.Results The combined T2WI,DWI and e-THRIVE+imaging model was more effective in predicting CCLM than the single sequence model.The AUC values of the training set and validation set were 0.919 and 0.896,respectively.Multivariate logistic regression analysis showed that CEA,CA199 and NLR were the independent risk factors for liver metastasis of colorectal cancer(P＜0.05),and the AUC values of the clinical models constructed by the three were 0.800 and 0.791 in the training set and validation set,respectively.Conclusion Multi-parameter 3D MRI imaging model is effective in predicting liver metastasis of colorectal cancer.The clinical models based on clinical indicators CEA,CA199 and NLR have certain predictive efficacy for liver metastasis of colorectal cancer.
关键词
结直肠癌/影像组学/临床危险因素/结直肠癌肝转移
Key words
Colorectal Cancer/Imaging Omics/Clinical Risk Factors/Liver Metastasis of Colorectal Cancer
维普
万方数据