Apparent diffusion coefficient (ADC) measurement in ovarian tumor: Effect of region-of-interest methods on ADC values and diagnostic ability.

Abstract

To investigate the difference in apparent diffusion coefficient (ADC) measurements of differently shaped regions of interest (ROIs) in ovarian tumors, and to investigate the diagnostic ability of differently shaped ROIs in differentiating benign from malignant ovarian tumors. Fifty-four patients with ovarian tumors with a solid component were evaluated. The patients underwent magnetic resonance (MR) examinations including diffusion-weighted imaging using a 3.0T MR system. Two readers measured ADCs using four ROI methods: freehand, square, round, and five small round ROIs. The interclass correlation coefficient (ICC) and repeated-measures analysis of variance were used to assess their measurement reliability and to compare ADCs for each ROI method. Receiver operating characteristic curve analysis and unpaired t-test on each ROI were used to differentiate benign and malignant ovarian tumors and assess the diagnostic ability. All ROI methods except the square ROI (0.56) showed good or excellent correlations (0.70-0.91). Minimum and mean ADC values differed significantly between the ROIs (P < 0.05). When using the freehand ROI, the minimum and mean ADC values were the lowest and highest, respectively. The optimal cutoff minimum and mean ADC values of each ROI for differentiating benign and malignant tumors were 0.81-1.06 × 10(-3) mm(2) /s and 1.15-1.52 × 10(-3) mm(2) /s, respectively. The areas under the curve showed no significant differences among the ADCs in the different ROI methods (P > 0.05). Minimum and mean ADCs from all ROIs showed significant differences between benign and malignant tumors (P < 0.05). The ROI shape influences ADC values and the optimal cutoff ADC values for differentiating benign from malignant ovarian tumors.