Purpose: We evaluated the diagnostic effectiveness of magnetic resonance (MR) imaging at 3 tesla to visualize ovarian tumors and problems associated with its use.
Materials and methods: From the records of 423 consecutive women who underwent pelvic MR imaging using a 3T system from April 2009 to June 2010, we analyzed 50 continuous cases of ovarian tumors proved by histopathology. We evaluated visualization of these tumors for image quality and artifacts using 5-point scales. For qualitative assessment, we scored overall image quality (1, poor, to 5, excellent), degree of conviction regarding the diagnosis (1, undiagnosable, to 5, diagnosable with high certainty), and 4 representative artifacts (penetrating, chemical shift, motion, and susceptibility artifact) (1, severe, to 5, little degradation). We also retrospectively reviewed the diagnostic features of the ovarian tumors and preoperative diagnostic accuracy. For quantitative assessment, we determined tumor size and ADC value.
Results: Overall quality score was scored 4.9±0.5, and conviction regarding diagnosis was 4.9±0.3. Artifacts caused little degradation in most cases: penetrating, 4.8±0.5; chemical shift, 4.3±0.5; motion, 4.6±0.6; and susceptibility, 3.8±0.9. Preoperative diagnostic accuracy was 92% (sensitivity 94.7%, specificity 90.3%). Mean tumor diameter was 88.3±61 mm. The mean ADC value was 1.04±0.3 in malignant tumors and 1.15±0.5 (×10⁻⁶ mm²/s) in benign tumors.
Conclusion: The quality of ovarian tumor images obtained with a 3T MR imaging system is adequate for diagnosis, with only slight degradation from penetrating or susceptibility artifacts.