Perfusion fraction of diffusion-weighted MRI for predicting the presence of blood supply in ovarian masses.

Purpose: To evaluate whether perfusion fraction (PF) calculated with diffusion-weighted magnetic resonance imaging (MRI) predicts the presence of blood supply in ovarian masses. Materials and Methods: PFs of 92 ovarian lesions in 53 patients administered gadolinium were retrospectively calculated with diffusion-weighted images at b-values of 0, 500, and 1000 sec/mm(2). PFs were compared between ovarian lesions, except for fat, with (n = 21) or without contrast enhancement (n = 57), using Student's t-test and receiver operating characteristics (ROC) curve analysis. Lesion enhancement rates of contrast-enhanced images at 30 and 180 seconds after gadolinium injection (ER(30sec) and ER(180sec)) and PFs were compared using Pearson's correlation coefficient. Results: PFs of the lesions with contrast enhancement were significantly higher than those without contrast enhancement (0.22 +/- 0.09 and 0.02 +/- 0.08, respectively, P < 0.0001). The ROC curve identified the best cutoff point for PF at 0.135 (95.2% sensitivity and 94.7% specificity) as a predictor of the contrast enhancement effect. The area under the ROC curve was 0.984. PF correlated moderately with ER(30sec) (0.62, y = 0.13x + 0.04, P < 0.0001) and ER(180sec) (0.74, y 0.13x + 0.03, P < 0.0001). Conclusion: PF calculated with diffusion-weighted images can potentially predict blood supply in ovarian masses.