ABSTRACT
Objectives: To study high-frequency 29 MHz transrectal side-fire micro-ultrasound (micro-US) for the detection of clinically significant prostate cancer (csPCa) on prostate biopsy, and validate an image interpretation protocol for micro-US imaging of the prostate. Materials and methods: A prospective randomized clinical trial was performed where 1676 men with indications for prostate biopsy and without known prostate cancer were randomized 1:1 to micro-US vs conventional end-fire ultrasound (conv-US) transrectal-guided prostate biopsy across five sites in North America. The trial was split into two phases, before and after training on a micro-US image interpretation protocol that was developed during the trial using data from the pre-training micro-US arm. Investigators received a standardized training program mid-trial, and the post-training micro-US data were used to examine the training effect. Results: Detection of csPCa (the primary outcome) was no better with the first-generation micro-US system than with conv-US in the overall population (34.6% vs 36.6%, respectively, P = .21). Data from the first portion of the trial were, however, used to develop an image interpretation protocol termed PRI-MUS in order to address the lack of understanding of the appearance of cancer under micro-US. Micro-US sensitivity in the post-training group improved to 60.8% from 24.6% (P < .01), while specificity decreased (from 84.2% to 63.2%). Detection of csPCa in the micro-US arm increased by 7% after training (32% to 39%, P < .03), but training instituted mid-trial did not affect the overall results of the comparison between arms. Conclusion: Micro-US provided no clear benefit over conv-US for the detection of csPCa at biopsy. However, it became evident during the trial that training and increasing experience with this novel technology improved the performance of this first-generation system.
