Presentation Authors: Jie Gao*, Qing Zhang, Chengwei Zhang, Hongqian Guo, Nanjing, China, People's Republic of
Introduction: Prostate cancer risk stratification at the time of diagnosis is necessary and cellularity metrics correlate with aggressiveness. We aimed to evaluate the relationship between apparent diffusion coefficients (ADCs) on MRI and cellularity metrics of prostate cancer (PCa) with whole-mount histopathology.
Methods: 225 PCa patients with preoperative 3.0 T MRI and lesions more than 0.5cm on whole-mount specimens were enrolled. The mean, median and 10th percentile ADC values were derived with histogram from pathologic-matching MRI images. With digitized pathology, gland components including epithelium (nuclei and cytoplasm), stroma and glandular lumen were segmented by different settings with ImageScope, and cellularity metrics including nuclear fraction and nuclear-to-cytoplasmic ratio (NCR) were derived. Correlation of ADCs with cellularity metrics and gland components were evaluated with Pearson correlation coefficient r.
Results: Total 256 lesions were identified and a representative pathologic segmentation illustration is shown in Fig 1. ADCs presented moderate correlation with NCR and nuclear fraction in both peripheral zone (PZ) and transition zone (TZ) (r=-0.370 to -0.403 and -0.471 to -0.528 for all lesions, P < 0.001) (Fig 2, Table 1). Besides, of all lesions and PZ lesions, ADCs showed negative correlation with epithelial fraction and positive correlation with luminal fraction. Generally, epithelial fraction correlated most strongly with ADCs.
Conclusions: Quantitative ADC parameters moderately correlated with pathologic cellularity metrics including nuclear fraction and nuclear-to-cytoplasmic ratio of prostate cancer with whole-mount pathology.