Are Dynamic Contrast-Enhanced Images Necessary for Prostate Cancer Detection on Multiparametric Magnetic Resonance...

To compare the results of magnetic resonance imaging (MRI) with and without contrast-enhanced sequences for the detection of clinically significant prostate cancer. The value of contrast in the MRI protocol for the detection of prostate tumors has been discussed in previous studies.

We performed an institutional review board-approved, retrospective study of patients who had undergone prostate MRI followed by fusion biopsy during a 16-month period. The patients had undergone MRI on a 3-Tesla scanner with a phased-array coil using a routine multiparametric protocol: T2-weighted, diffusion, and dynamic postcontrast enhancement sequences. The examination films were read independently by 2 readers in 2 sessions: first without contrast-enhanced images and second with contrast-enhanced images. The images were graded according to the suspicion of clinically significant prostate cancer (scale, 1-5). The kappa coefficient was used to compare the readings with and without contrast. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for both readers were calculated using the biopsy findings as a reference standard. The level for statistical significance was set at P < .05.

A total of 118 patients were included, and clinically significant prostate cancer was found in 48 patients (40.7%). The MRI findings on both contrast-enhanced and noncontrast-enhanced images had no statistically significant differences for either reader (P > .05 for all levels of suspicion). The sensitivity ranged from 68.3% to 80.7%, specificity from 57.1% to 77.1%, positive predictive value from 29.9% to 38.8%, negative predictive value from 88.0% to 91.3%, and accuracy from 60.6% to 73.0%.

Our results show that contrast-enhanced sequences provide minimal or no increased value for the detection of clinically significant prostate cancer.