Table 4.

Radiomics in uterine masses differentiation.

Authors of the Study	Year	Diagnostic Method	The Number of Patients	Main Findings
Malek et al. [100]	2019	MRI	42 patients with 60 uterine lesions: 10 uterine sarcomas and 50 ULM	They extracted 21 radiomics features achieving the sensitivity and the specificity: 100% and 90%, respectively
Xie et al. [101]	2019	MRI	78 patients with 29 sarcomas and 49 ULM	They analyzed three different volumes of interests (VOIs.): (1) tumor; (2) tumor and small piece of surrounded tissue; and (3) whole uterus. The best diagnostic performance of the VOI radiomic model was one that covered the whole uterus.
Wang et al. [104]	2021	MRI	134 patients including 81 with LM and 53 malignant uterine mesenchymal tumors	They used three clinical parameters: the age, menopausal status, and symptoms and seven radiomic features. The highest AUC value was the highest when the radiomics model was combined with a clinical model to 0.91.
Chiappa et al. [105]	2021	Ultrasound	70 patients with uterine mesenchymal lesions: 20 with sarcoma and 50 ULM	390 radiomic IBSI-compliant features were extracted and 308 radiomic features were found to be stable. The best classification system showed an accuracy of 0.85, sensitivity of 0.80, specificity of 0.87, AUC of 0.86.