Preoperative MRI of Uterine Malignant Mixed Mullerian Tumors versus Endometrial Carcinomas with Emphasis on Dynamic Enhancement Characteristics
Sandra Alheli Garza1, Tara Sagebiel2, Wei Wei3, Jingfei Ma4, and Priya Bhosale2

1Department of Diagnostic and Interventional Imaging, University of Texas Health Science Center at Houston, Houston, TX, United States, 2Diagnostic Radiology Department, MD Anderson Cancer Center, Houston, TX, United States, 3Biostatistics Department, MD Anderson Cancer Center, Houston, TX, United States, 4Department of Imaging Physics, MD Anderson Cancer Center, Houston, TX, United States


Distinguishing uterine malignant mixed mullerian tumors (MMMTs) from endometrial carcinomas preoperatively by pelvic MRI may help with surgical and treatment planning of this highly aggressive tumor. Our study found that prolapse of tumor through the external cervical os, delayed iso- or hyper-enhancement, high mean tumor: myometrium positive enhancement integral (PEI) ratio and low tumor signal enhancement ratio (SER) are more commonly seen in patients with MMMT and may alert the radiologist to the possibility of this diagnosis.


Uterine carcinosarcoma, also known as Malignant Mixed Mullerian tumor (MMMT) makes up less than 5% of uterine tumors. However, it is highly aggressive, has a worse prognosis than endometrial carcinoma (EC), and disproportionately accounts for up to 16% of uterine tumor-related deaths (1). More than half of patients will have extrauterine spread at the time of presentation (1). Preoperative determination of uterine MMMT may alert the surgeons of prognosis and guide surgical approach and adjuvant therapies. The objective of this study was to determine if the dynamic contrast enhancement characteristics by preoperative pelvic MRIs can help distinguish uterine MMMT from endometrial carcinoma.


Preoperative pelvic MRIs were retrospectively reviewed for 37 patients with histologically proven uterine MMMT and 42 patients with histologically proven EC after IRB approval and waiver of informed consent. All cases had sagittal T2-weighted sequences. In the uterine MMMT group, 23 MRIs had fat-suppressed T1-weighted 3D dynamic contrast enhancement (DCE) sequences (21 in sagittal plane, 2 in axial plane), 10 had delayed T1-weighted fat-suppressed post contrast sequences and 4 were without contrast. In the EC group, 41 MRIs had fat-suppressed T1-weighted 3D DCE sequences (all in sagittal plane) and 1 was without contrast. Presence or absence of tumor prolapse through the internal or external cervical os and the length of prolapse were recorded using the sagittal T2-weighted images. Qualitative enhancement characteristics assessed included enhancement of tumor relative to myometrium in the early and delayed phases; these were scored as tumor hypo-enhancement, iso-enhancement or hyper-enhancement relative to the myometrium. For semi-quantitative dynamic contrast-enhancement analysis, functional maps were generated for positive enhancement integral (PEI), maximum slope of increase (MSI) and signal enhancement ratio (SER) parameters on a GE Advantage Window (AW) workstation. Values of these parameters were obtained for tumor and outer myometrium ROIs and tumor: myometrium PEI, MSI and SER ratios were calculated. Comparison between EC and MMMT groups was performed using Fisher’s exact test for qualitative assessments and Wilcoxon rank sum test for quantitative measurements.


MMMTs showed a statistically significant higher incidence of tumor prolapse through the external os (p=0.02). Length of prolapse was longer for MMMTs (1.1 +/- 1.8 cm) compared to ECs (0.19 +/-0.4 cm), findings which trended toward statistical significance (p=0.06). No significant difference was found in the early contrast enhancement of uterine MMMT and EC, with the majority of both tumor groups showing early hypo-enhancement relative to myometrium, 68.2% (15/22) MMMT and 68.4% (26/38) EC (p>0.99). Late contrast enhancement was significantly different between both groups (p=0.02), with up to 15.6% (5/32) of MMMTs showing delayed iso- or hyper-enhancement compared to none in the EC group. MMMTs had a statistically significant higher mean tumor: myometrium PEI ratio of 0.92 (SD 0.39) compared to 0.53 (SD 0.19) for EC (p<0.0001). MMMT had statistically significant lower tumor SER of 92.5 (SD 38.5) compared to 118.5 (SD 36.7) for EC (p=0.006).


Uterine MMMTs have been shown to often be indistinguishable from ECs in several studies in terms of their morphologic appearance and signal characteristics (2). To our knowledge, this is the largest study to date exploring the dynamic enhancement characteristics of these tumors and whether tumor prolapse through the cervical os is a distinguishing factor. Presence of tumor prolapse through the external cervical os is highly suggestive of MMMT. Although there was no difference in the early enhancement pattern in our study, MMMTs showed delayed iso- and hyper-enhancement compared to endometrial carcinomas. Based on our study, semi-quantitative parameters may also help distinguish between MMMTs and ECs, with higher mean tumor: myometrium PEI ratios and lower tumor SER ratios suggesting MMMT. Preoperative determination of MMMT may guide surgical planning and treatment strategy given their aggressive behavior and propensity towards nodal or distant metastatic disease at presentation (1). Our initial results are promising and larger studies are needed to determine the significance of enhancement pattern and semi-quantitative parameters in differentiating MMMT from EC.


MMMTs may demonstrate tumor prolapse through the external cervical os, delayed iso- or hyper-enhancement, high mean tumor: myometrium PEI ratios and low tumor SER ratios. Knowledge of this tumor morphology and enhancement pattern may help in appropriate diagnosis, surgical and treatment planning in these patients.


No acknowledgement found.


1. Cantrell LA, Blank SV, Duska LR. Uterine carcinosarcoma: a review of the literature. Gynecol Oncol 2015; 137: 581-588.

2. Bharwani N, Newland A, Tunariu N, et al. MRI appearances of uterine malignant mixed mullerian Tumors. AJR 2010; 195: 1268-1275.

Proc. Intl. Soc. Mag. Reson. Med. 25 (2017)