Prostate cancer detection in patients with prior negative biopsy undergoing cognitive-, robotic- or in-bore MRI target biopsy.

ABSTRACT

PURPOSE: To evaluate the detection rate among three different targeted biopsy approaches of robot-assisted MRI/TRUS fusion (RA-TB), mpMRI in-bore (MRGB), cognitive fusion guidance biopsy (COG-TB) for the detection of prostate cancer (PC) and clinically significant PC (csPC). METHODS: Between 2014 and 2016, 156 patients with a lesion on mpMRI, performed in accordance with ESUR guidelines, due to cancer suspicion or on-going cancer suspicion after prior negative prostate biopsy, underwent targeted biopsy with RA-TB, MRGB or COG-TB. All lesions were rated according to PI-RADS v2. We compared detection rates between techniques. Models were constructed to predict the detection of overall PC and csPC and using a 1000 boot-strap sample. RESULTS: In the all cohort, 73, 45 and 38 patients underwent RA-TB, MRGB or COG-TB, respectively. Overall PC was found in 39 (52.42%), 23 (51.11%) and 11 (28.95%) (p = 0.04) patients of RA-TB, MRGB and COG-TB arm, respectively. As concerning the detection of csPC, it was found in 26 (35.62%),18 (40.0%) and 9 (23.68%) patients of RA-TB, MRGB and COG-TB arm (p = 0.27). Model 1 showed that RA-TB [OR 10.08 (95% CI 1.95-51.97); p < 0.01] and MRGB [OR 12.88 (95% CI 2.36-70.25); p < 0.01] were associated with overall PC detection in TB, while only MRGB was associated with csPC at TB (model 2) [OR 5.72; (95% CI 1.40-23.35); p < 0.01]. The c-index for model 1 and model 2 was 0.86 and 0.85, respectively. We did not report significant complications between groups. CONCLUSION: In-bore biopsy and MRI/TRUS fusion-guided biopsy showed greater accuracy in detecting PC compared to cognitive fusion as modeled in a newly established normogram.