Pre-treatment ADC histogram-analysis at whole body diffusion-weighted MRI predicts disease free survival in ovarian cancer

Methods Institutional review board approval and informed consent were obtained for this prospective study. Forty-four women diagnosed with FIGO stage III or IV ovarian carcinoma underwent 3-Tesla WB-DWI/MRI using 2 b-values (b=0-1000 s/mm), T2-weighted and contrast-enhanced T1-weighted sequences prior to treatment. The primary tumour was delineated using semi-automated software and was analysed by using an ADC histogram approach: mean and median ADC, standard deviation (SD), coefficient of variation (CoV, SD/mean), kurtosis and skewness were calculated. Kaplan-Meier with log-rank statistics were used to correlate baseline ADC parameters to disease free survival (DFS). Effects of confounding patientsand tumour-related factors were taken into consideration using Cox proportional hazard model.