OBJECTIVES: We retrospectively compared and quantified magnetic resonance (MR) images to distinguish major histological types of uterine sarcomas and malignant and benign tumors. METHODS: MR images were obtained from patients who underwent preoperative examinations. We compared 25 pathologically confirmed uterine sarcomas (8 leiomyosarcomas, 11 carcinosarcomas, 6 endometrial stromal sarcomas) with 25 uterine leiomyomas. MR findings included tumor size, location, contour, signal intensity (SI), and contrast enhancement. Analysis focused on the contrast ratio (CR) of SI in T2-weighted images for the areas of lowest, highest, and main SI of each tumor as well as the contrast-enhanced ratio (CER) for the main solid part of each tumor in contrast-enhanced T1-weighted images. We evaluated diffusion-weighted (DW) images and apparent diffusion coefficient (ADC) values in 18 tumors (4 sarcomas, 14 leiomyomas). RESULTS: Uterine sarcomas and leiomyomas differed significantly in tumor location, contour, hemorrhaging, necrotic and cystic components, CR for the area of lowest SI (P < 0.05), CR for the area of main SI (P < 0.01), and CER (P < 0.05). Leiomyosarcomas were larger than carcinosarcomas or endometrial stromal sarcomas, and the CR for the area of lowest SI of leiomyosarcomas (P < 0.05) was significantly lower. The CER for endometrial stromal sarcomas (P < 0.05) showed the most homogeneous enhancement. Hemorrhagic or necrotic and cystic components were found more often in larger tumors, although there was no significant difference in their occurrence between sarcoma types. All uterine sarcomas showed high intensity on DW images. The ADC values were lower of uterine sarcomas than leiomyomas, although the difference was not statistically significant. CONCLUSION: Quantitative assessment using the CR or CER was useful for distinguishing benign and malignant uterine tumors as well as major histological types of uterine sarcomas.
OBJECTIVES: We retrospectively compared and quantified magnetic resonance (MR) images to distinguish major histological types of uterine sarcomas and malignant and benign tumors. METHODS: MR images were obtained from patients who underwent preoperative examinations. We compared 25 pathologically confirmed uterine sarcomas (8 leiomyosarcomas, 11 carcinosarcomas, 6 endometrial stromal sarcomas) with 25 uterine leiomyomas. MR findings included tumor size, location, contour, signal intensity (SI), and contrast enhancement. Analysis focused on the contrast ratio (CR) of SI in T2-weighted images for the areas of lowest, highest, and main SI of each tumor as well as the contrast-enhanced ratio (CER) for the main solid part of each tumor in contrast-enhanced T1-weighted images. We evaluated diffusion-weighted (DW) images and apparent diffusion coefficient (ADC) values in 18 tumors (4 sarcomas, 14 leiomyomas). RESULTS: Uterine sarcomas and leiomyomas differed significantly in tumor location, contour, hemorrhaging, necrotic and cystic components, CR for the area of lowest SI (P < 0.05), CR for the area of main SI (P < 0.01), and CER (P < 0.05). Leiomyosarcomas were larger than carcinosarcomas or endometrial stromal sarcomas, and the CR for the area of lowest SI of leiomyosarcomas (P < 0.05) was significantly lower. The CER for endometrial stromal sarcomas (P < 0.05) showed the most homogeneous enhancement. Hemorrhagic or necrotic and cystic components were found more often in larger tumors, although there was no significant difference in their occurrence between sarcoma types. All uterine sarcomas showed high intensity on DW images. The ADC values were lower of uterine sarcomas than leiomyomas, although the difference was not statistically significant. CONCLUSION: Quantitative assessment using the CR or CER was useful for distinguishing benign and malignant uterine tumors as well as major histological types of uterine sarcomas.