Changes in Epithelium, Stroma, and Lumen Space Correlate More Strongly with Gleason Pattern and Are Stronger Predictors of Prostate ADC Changes than Cellularity Metrics

Purpose: To investigate the hypothesis that the clinically observed decrease in apparent diffusion coefficient (ADC) at diffusion-weighted magnetic resonance imaging with increasing prostate cancer Gleason grade can be attributed to an increasing volume of low-diffusivity epithelial cells and corresponding decreasing volumes of higher-diffusivity stroma and lumen space rather than to increased cell density. Materials and Methods: Tissue samples were acquired after institutional ethics review committee approval and informed consent from patients were obtained. Nuclear count, nuclear area, and gland component volumes (epithelium, stroma, lumen space) were measured in tissue from 14 patients. Gland component volumes and cellularity metrics were correlated with Gleason pattern (Spearman rank correlation coefficient) and measured ADC (Pearson correlation coefficient) in six prostates ex vivo. Differences between metrics for cancerous tissue and those for normal tissue were assessed by using a two-tailed two-sample t test. Linear mixed models with a post hoc Fisher least significant difference test were used to assess differences between gland component volumes and cellularity metrics for multiple groups. To adjust for a clustering effect due to repeated measures, the organ mean value of the measured metric for each tissue type was used in the analysis. Results: There were significant differences between Gleason patterns for gland component volumes (P<.05) but not nuclear count (P = .100) or area (P = .141). There was a stronger correlation of Gleason pattern with gland component volumes (n = 553) of epithelium (Spearman rho = 0.898, P < .001), stroma (rho = -0.651, P < .001), and lumen space (rho = -0.912, P = .007) than with the cellularity metrics (n = 288) nuclear area (rho = 0.422, P = .133) or nuclear count (rho = 0.082, P = .780). There was a stronger correlation between measured ADC and lumen volume (r = 0.688, P < .001) and epithelium volume (r = -0.647, P < .001) than between ADC and nuclear count (r = -0.598, P,.001) or nuclear area (r = -0.569, P < .001) (n = 57). Conclusion: Differences in the gland compartment volumes of prostate tissue having distinct diffusivities, rather than changes in the conventionally cited cellularity metrics, are likely to be the major contributor to clinically observed variations of ADC in prostate tissue. (C) RSNA, 2015