Objectives: Our aim was to establish threshold criteria based on quantitative DCE-MRI data as independent predictors of malignancy in a complex (solid, solid/cystic) ovarian mass.
Methods: The MRI of 26 lesions in 25 patients with a complex ovarian mass (age range, 17-80 years; mean 43 years) was retrospectively reviewed and correlated with histology following resection. Cases with solid tumour components, definitive histology and relevant dynamic imaging were included. These were categorised into two groups, benign (N = 14) and malignant (N = 12). Following dynamic contrast-enhanced imaging, regions of interest were drawn around the solid tumour component. Maximum actual enhancement (SImax), maximum relative enhancement (SIrel), wash-in rate (WIR) and SImax (tumour)/SImax (psoas) ratio were analysed. Threshold criteria for malignancy were established.
Results: There was a significant difference in SImax (p < 0.001), SIrel (p < 0.05), WIR (p < 0.001) and SImax (tumour)/SImax (psoas) between the two groups. Optimal threshold criteria for malignancy were established; SImax > or = 250 or SImax (tumour)/SImax (psoas) > or = 2.35 divided the two groups with 100% sensitivity, specificity and accuracy.
Conclusion: Threshold criteria established in this preliminary study using quantitative DCE-MRI provide an accurate method for the prediction of malignancy, particularly in preoperative indeterminate cases.