Magnetic Resonance Imaging Grading System for Preoperative Diagnosis of Leiomyomas and Uterine Smooth Muscle Tumors.

STUDY OBJECTIVE: To evaluate a new magnetic resonance imaging (MRI) grading system for preoperative differentiation between benign and variant-type uterine leiomyomas including smooth muscle tumors of uncertain malignant potential (STUMPs). DESIGN: Retrospective analysis (Canadian Task Force classification III). SETTING: Teaching hospital (Teine Keijinkai Hospital). PATIENTS: Three-hundred thirteen patient medical records were retrospectively reviewed if treated for uterine myomas and diagnosed with variant type leiomyomas or STUMPs (n = 27) or benign, typical leiomyomas (n = 286) and treated between January 2012 and December 2014. INTERVENTION: Uterine myoma classifications using MRI findings according to a 5-grade system (grades I-V) based on 3 elements. MEASUREMENTS AND MAIN RESULTS: Uterine myoma MRI classifications were based on 3 elements: T2-weighted imaging (high or low), diffusion-weighted imaging (high or low), and apparent diffusion coefficient values (high or low; apparent diffusion coefficient < 1.5 × 10-3 mm2/sec was considered low). Grades I to II were designated as typical or benign leiomyomas, grade III as degenerated leiomyomas, and grades IV to V as variant type leiomyomas or STUMPs. Accuracy levels were 98.9%, 100%, 94.3%, 58.8%, and 41.9% for grades I through V lesions, respectively. The grades were divided into 2 groups to discriminate benign leiomyomas and STUMPs (grades I-III were considered negative and grades IV-V positive). Grades IV to V scored 85.2% for sensitivity, 91.3% for specificity, 47.9% positive predictive value, 98.5% negative predictive value, a 9.745 positive likelihood ratio, and a .162 negative likelihood ratio. CONCLUSION: This novel MRI grading system for uterine myomas may be beneficial in differentiating benign leiomyomas from STUMPs or variant type leiomyomas and could be a future effective presurgical assessment tool.