Use of cylindrical coordinates to localize prostate cancers on MRI and prostatectomy pathology

Purpose

To describe and test a quantitative system for designating prostate tumor location on magnetic resonance imaging (MRI) and prostatectomy. A system for describing tumor location will facilitate research correlating MRI and pathology.

Materials and methods

The prostate cylindrical coordinate (PCC) system was developed for locating prostate tumors using 3 coordinate values. The 3 coordinate values include the angular location centered on the urethra, the radial distance to the periphery and the long axis from apex to base. To evaluate this system, 26 tumors were identified where the prostate cancer was noted by both the radiologist and the pathologist. PCC values were assigned independently to MRI lesions and corresponding tumors. Intraclass correlation coefficient (ICC) was calculated to assess agreement between PCC assigned using MRI and pathology. The coordinates were used to calculate the average distance between the centers of the same lesion measured by MRI and pathology.

Results

Each of the cylindrical coordinates assigned by MRI and pathology were compared and there was no significant difference. The agreement was excellent, and the ICC was 0.70 (P<0.001) for the angular coordinate, 0.81 (P<0.001) for the radial distance, and 0.94 (P<0.001) for the long axis. Compared to pathology, lesions on MRI were significantly larger (1.17 vs. 0.86 cm², P<0.001) but there was strong agreement between the measurements on MRI and pathology (ICC = 0.89, P<0.001). The distance between the centers of the lesions measured on MRI and pathology was small (10.13 mm, s.d. = 8.70).

Conclusions

The PCC system quantitatively characterizes lesions seen on MRI and prostatectomy pathology with good agreement.