Objective To explore the value of MRI radiomics in the differential diagnosis of cerebral alveolar echinococcosis(CAE)and brain metastases(BM).Methods The clinical and imaging data of 30 patients with CAE confirmed by postoperative pathology in the First Affiliated Hospital of Xinjiang Medical University and 38 patients with BM in the Affiliated Hospital of Traditional Chi-nese Medicine of Xinjiang Medical University during the same period were retrospectively analyzed.Preoperative T1-enhanced axial images using 3D Slicer software were used to delineate the volume of interest(VOI)of the lesions,and Python was used to extract 18 radiomics features including the 10th and 90th percentile and median of the lesions.The radiomics parameters of the two groups were compared and statistically analyzed,respectively.The efficacy of each parameter in the differential diagnosis of CAE and BM was explored.Results There was no significant difference in age and body mass index(BMI)between CAE and BM.There was sig-nificantly more males in CAE than that in BM(P<0.001).Previous history of primary tumor and life history in epidemic area were the key to the differential diagnosis of CAE and BM.High correlation features were filtered from 18 radiomics,and statistical differ-ences were found in 7 features such as 90th percentile and entropy(P<0.05).The combined analysis of these seven features had the best diagnostic performance,with area under the curve(AUC)of 0.732,sensitivity of 79.5%,specificity of 61.1%,and accuracy of 72.3%,respectively.Logistic regression analysis showed that kurtosis odds ratio(OR)value was 0.10,95%confidence interval was 0.01,0.83(P<0.05).Conclusion For CAE and BM with very similar conventional imaging findings,clinical history combined with radiomics can provide better differential diagnosis basis.

Objective To explore the value of MRI radiomics in the differential diagnosis of cerebral alveolar echinococcosis(CAE)and brain metastases(BM).Methods The clinical and imaging data of 30 patients with CAE confirmed by postoperative pathology in the First Affiliated Hospital of Xinjiang Medical University and 38 patients with BM in the Affiliated Hospital of Traditional Chi-nese Medicine of Xinjiang Medical University during the same period were retrospectively analyzed.Preoperative T1-enhanced axial images using 3D Slicer software were used to delineate the volume of interest(VOI)of the lesions,and Python was used to extract 18 radiomics features including the 10th and 90th percentile and median of the lesions.The radiomics parameters of the two groups were compared and statistically analyzed,respectively.The efficacy of each parameter in the differential diagnosis of CAE and BM was explored.Results There was no significant difference in age and body mass index(BMI)between CAE and BM.There was sig-nificantly more males in CAE than that in BM(P<0.001).Previous history of primary tumor and life history in epidemic area were the key to the differential diagnosis of CAE and BM.High correlation features were filtered from 18 radiomics,and statistical differ-ences were found in 7 features such as 90th percentile and entropy(P<0.05).The combined analysis of these seven features had the best diagnostic performance,with area under the curve(AUC)of 0.732,sensitivity of 79.5%,specificity of 61.1%,and accuracy of 72.3%,respectively.Logistic regression analysis showed that kurtosis odds ratio(OR)value was 0.10,95%confidence interval was 0.01,0.83(P<0.05).Conclusion For CAE and BM with very similar conventional imaging findings,clinical history combined with radiomics can provide better differential diagnosis basis.