Multiparametric MRI does not improve transition zone (TZ) prostate cancer detection or localization accuracy compared with 3T T2-weighted imaging, according to a study published in the January issue of Radiology.

Caroline M. A. Hoeks, MD, of Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands, and colleagues explained that between 25 and 30 percent of prostate cancers in the U.S. are in the TZ, which surrounds the proximal urethra. Studies have shown up to 9 percent of TZ cancers have a Gleason grade 4-5 component with extracapsular extension and positive resection margins. Anterior TZ cancers are particularly hard to detect with standard prostate biopsy schemes.

“Therefore, improvement of TZ cancer detection is an important goal in prostate cancer diagnostics,” wrote the authors.

Since TZ prostate cancers are difficult to differentiate from benign prostatic hyperplasia on T2-weighted MRI, the authors retrospectively analyzed detection rates and localization accuracy of multiparametric MRI, which added low-b-value diffusion-weighted (DW) MR apparent diffusion coefficient (ADC) maps and dynamic contrast material-enhanced (DCE) images.

From a group of 197 patients with radical prostatectomy specimens, 28 patients with TZ cancer were included. Another 35 patients without TZ cancer were randomly selected as a control group. Four radiologists randomly scored T2-weighted and DW ADC images; T2-weighted and DCE MR images; and T2-weighted, DW ADC and DCE MR images.

Results showed that detection accuracy did not significantly differ between T2-weighted and multiparametric MRI for all cancers, with accuracy rates of 68 percent and 66 percent, respectively. The authors also reported no significant differences when TZ cancers were separated into high and low Gleason grades. Multiparametric MRI did not improve T2-weighted imaging localization accuracy.

“Our results imply that there is ample room for improvement of [multiparametric] MR imaging techniques for accurate TZ evaluation,” wrote the authors. “Studies on [multiparametric] MR imaging (including high-b-value DW imaging) with increased spatial resolution are needed to improve detection and localization of TZ cancers. Furthermore, false-positive [multiparametric] MR imaging readings may be reduced by training radiologists in recognition of cancer-specific patterns rather than in quantitative evaluation of the TZ.”

Hoeks and colleagues also speculated that computer-aided diagnosis may one day have an important supportive role in the detection and localization of TZ cancers.