Biopsy-Integrated 3D Magnetic Resonance Imaging Modeling of Prostate Cancer and Its Application for Gleason Grade and Tumor Laterality Assessment

Abstract

Context: Grade Group assessed using Gleason combined score and tumor extent is a main determinant for risk stratification and therapeutic planning of prostate cancer.

Objective: To develop a 3-dimensional magnetic resonance imaging (MRI) model regarding Grade Group and tumor extent in collaboration with uroradiologists and uropathologists for optimal treatment planning for prostate cancer.

Design: We studied the data from 83 patients with prostate cancer who underwent multiparametric MRI and subsequent MRI-transrectal ultrasound fusion biopsy and radical prostatectomy. A 3-dimensional MRI model was constructed by integrating topographic information of MRI-based segmented lesions, biopsy paths, and histopathologic information of biopsy specimens. The multiparametric MRI-integrated Grade Group and laterality were assessed by using the 3-dimensional MRI model and compared with the radical prostatectomy specimen.

Results: The MRI-defined index tumor was concordant with radical prostatectomy in 94.7% (72 of 76) of cases. The multiparametric MRI-integrated Grade Group revealed the highest agreement (weighted κ, 0.545) and a significantly higher concordance rate (57.9%) than the targeted (47.8%, P = .008) and systematic (39.4%, P = .01) biopsies. The multiparametric MRI-integrated Grade Group showed significantly less downgrading rates than the combined biopsy (P = .001), without significant differences in upgrading rate (P = .06). The 3-dimensional multiparametric MRI model estimated tumor laterality in 66.2% (55 of 83) of cases, and contralateral clinically significant cancer was missed in 9.6% (8 of 83) of cases. The tumor length measured by multiparametric MRI best correlated with radical prostatectomy as compared with the biopsy-defined length.

Conclusions: The 3-dimensional model incorporating MRI and MRI-transrectal ultrasound fusion biopsy information easily recognized the spatial distribution of MRI-visible and MRI-nonvisible cancer and provided better Grade Group correlation with radical prostatectomy specimens but still requires validation.