Comparison of Malignancy-Prediction Efficiency between Contrast and Non-Contract CT Based Radiomics Features in Gastrointestinal Stromal Tumours: A Multicenter Study

Abstract

Background and Aim: To develop and validate radiomics signature from non-enhanced computed tomography (CT, NE-RS) to predict the malignancy potential of gastrointestinal stromal tumours (GIST) preoperatively and compare it with radiomics signature from contrast-enhanced CT. Method: A total of 370 GIST patients from four centres were divided into the training, internal validation, and external validation cohort. Minimum redundancy maximum relevance (mRMR) algorithm, and least absolute shrinkage and selection operator (LASSO) algorithm were used to filter unstable features. (1) NE-RS and CE-RS were built and compared for the prediction of malignancy potential of GIST using the area under the curve (AUC). (2) The radiomics model, with a combination of tumour size and NE-RS, was also developed. Results: The AUC values were comparable between NE-RS and CE-RS in the training (0.965 versus 0.936; P=0.251), internal validation (0.967 versus 0.960; P=0.801) and external validation (0.941 versus 0.899; P=0.173) cohorts in diagnosis of high malignancy potential of GISTs. We next focused on the NE-RS. With 0.185 selected as the cut-off of NE-RS for diagnosis of the malignancy potential of GISTs, accuracy, sensitivity, and specificity for diagnosis high-malignancy potential GIST was 90.0, 88.2, and 92.3%, respectively, in the training cohort; 89.1, 94.9, and 80.0%, respectively, in the internal validation cohort; and 84.6, 76.1, and 91.0%, respectively, in the external cohort. Compared with only NE-RS, the radiomics model increased the sensitivity in the diagnosis of GIST with high-malignancy potential by 5.9 (P=0.025), 2.5 (P=0.317), 10.5% (P=0.008) in the training, internal validation, and external validation cohort, respectively. Conclusions: The NE-RS had comparable prediction efficiency in diagnosis of high-malignant potential of GISTs to CE-RS. The NE-RS and radiomics model both had excellent accuracy in predicting malignancy potential of GISTs. Funding Statement: This work was supported by grants from Shanghai Municipal Key Clinical Specialty (Grant No. shslczdzk05902). Declaration of Interests: The authors declare that they have no competing interests. Ethics Approval Statement: Ethical approval obtained from all four participating hospitals. All patents provided written informed consent.