Establishing significant correlations between multiparametric magnetic resonance histology (MRH) and cytometric features in prostate cancer

Motivation: We are developing magnetic resonance histology (MRH) for surgical pathology to improve prostate cancer tissue sampling and analysis. Goal(s): Our goal was to measure the quantitative relationships between multi-parametric MRH signal, radiomic features, and cytometric (pathology) features in clinical prostate cancer tissues. Approach: We collected tissues from prostate cancer patients and performed high resolution MRH (40-55μm3). MRH volumes were registered to pathology at the microscopic level. Linear relationships between quantitative MRH and cytometric features were measured across the cohort. Results: We found statistically significant linear relationships between multiple tissue cytometric features and quantitative MRH, including T2 measurements, diffusion tensor scalers, and radiomic features. Impact: This research provides quantitative links between magnetic resonance histology and ground-truth histopathology in prostate cancer. Not only will this improve how we implement MRH for 3D pathology applications, we will also improve our interpretation of clinical prostate MRI.