How many cores are needed to detect clinically significant prostate cancer on targeted MRI-ultrasound fusion biopsy?

View Poster


The optimal number of MRI-US fusion biopsy cores to adequately sample regions of interest (ROI) remains unknown. To better understand the optimal approach to lesion targeting, we aimed to examine the cancer detection rate based on sequential number of cores obtained.


Of 744 patients undergoing MRI-US fusion biopsy between 2012 and 2016 at our institution, we identified 628 men with targets on multi-parametric MRI (mpMRI) who underwent targeted and systematic fusion biopsy using the Artemis platform for clinical suspicion (n=465) or known history of PCa (n=163). mpMRI studies were reviewed by genitourinary radiologists using a 3-tiered Likert scale and PI-RADS classification schema. Biopsy was performed by two urologists performing a high volume of fusion biopsies. Cores were taken sequentially from each ROI with an even distribution. The primary outcome was the proportion of high-grade (Gleason ≥3+4) cancers missed on a 2-core lesion biopsy.


We biopsied 1,233 ROI with a median of 5 cores (IQR 3-5) from each ROI. A total of 581 ROI (47%) were positive for any Gleason grade PCa, in 380 (61%) patients. On a per-lesion basis, 84% of any Gleason score cancers were detected with a two-core biopsy and 77% of Gleason ≥3+4 tumors. Cancer detection rates improved with increasing number of cores (Table 1). For any Gleason grade PCa, additional cores beyond 5 cores had no significance. For all G≥3+4 cancer, there was no significance with additional sampling from 3 to 4 cores, but improved detection from 3 to 5 cores (P


On a per-lesion basis, sampling two cores of mpMRI-evident lesions at the time of fusion biopsy misses nearly one-quarter of clinically significant PCa that would be detected on additional sampling.

Funding: National Institute of Diabetes And Digestive And Kidney Diseases of the National Institutes of Health