Preliminary study of histogram analysis of mono-exponential and intravoxel incoherent motion models in epithelial ovarian cancer

Abstract

Objective To evaluate the diagnostic performance of the histogram analysis of mono-exponential and intravoxel incoherent motion(IVIM) models to the dualistic model of epithelial ovarian cancer(EOC). Methods Forty female patients with histopathologically proven epithelial ovarian cancer underwent preoperative MR examination. Scanning sequences included conventional imaging, diffusion-weighted magnetic resonance imaging with 11 b values (0, 30, 50, 100, 150, 200, 400, 600, 800, 1 000, 1 500 s/mm2) and dynamic contrast enhanced MRI (DCE-MRI). Based on the dualistic model of EOC, all patients were divided into two groups: type Ⅰ (low grade, n=16) and type Ⅱ (high grade, n=24). ADC, D, D* and f maps and their corresponding histograms were generated by post-processing software. Based on an entire-tumour measurement, the following histogram parameters were recorded, respectively: (a) Mean; (b) the 10th percentile (10th); (c) the mean of the top 10 percent (MeanL); (d) the 90th percentile (90th); (e) the mean of the bottom 10 percent (MeanR). Two types were compared using independent sample t test or Mann-Whitney U test. And areas under ROC curve between two groups were assessed. Results For ADC , D, and f, all indices(Mean, 10th,MeanL,90th,MeanR) of the histogram were significantly lower in type Ⅱ than in type Ⅰ (P 0.05). D demonstrated a comparable accuracy with ADC in differentiating the grade of EOC (area under curve: Mean, 0.898 vs. 0.893; 10th, 0.880 vs. 0.846; MeanL, 0.878 vs. 0.858; 90th, 0.895 vs. 0.839; MeanR, 0.872 vs. 0.814), and both ADC and D have better performance than f. Conclusion It is feasible to stratify the grade of EOC by mono-exponential and IVIM models with histogram metrics,diagnositic efficiency of ADC and D values are higher. Key words: Ovarian neoplasms; Magnetic resonance imaging; Comparative study