Classification of Benign Ovarian Lesions and Epithelial Ovarian Cancer Subtypes Using MRI Cytometry-Derived Microstructural Parameters

Motivation: Preoperative differentiation between benign ovarian lesions and epithelial ovarian cancers (EOCs), and further subclassification type I and II EOCs, is essential for guiding treatment. Goal(s): To evaluate MRI cytometry-derived microstructural parameters for noninvasive differentiating benign ovarian lesions and EOCs, and between type I and II EOCs. Approach: MRI cytometry extracted microstructural parameters, including diameter, intracellular volume fraction, cellularity and extracellular diffusivity. Diagnostic abilities were evaluated using unpaired two-tailed t-tests, linear regression, AUC, sensitivity and specificity. Results: Microstructural parameters showed significant differences and strong diagnostic performance for distinguishing benign ovarian lesions from EOCs and differentiating type I and II EOCs. Impact: For the first time, MRI cytometry demonstrated high feasibility and diagnostic accuracy for non-invasively distinguishing benign ovarian lesions from epithelial ovarian cancers (EOCs) and further differentiating between type I and II EOCs.