Introduction:

Previous research has demonstrated that cumulative cancer location (CCLO) derived from diagnostic and confirmatory biopsy is associated with grade reclassification during active surveillance (AS) and has higher predictive value than number of positive cores alone (GR; Erickson et al, Eur Urol Onc 2018). However, these data have yet to be validated in additional large cohorts and were based on patients who did not undergo prostate MRI. It is unclear whether CCLO is still predictive of GR when taking MRI and targeted biopsy data into account. Therefore, we aimed to validate prior findings in our institutional AS cohort as well as evaluate whether CCLO is still associated GR when accounting for both traditional biopsy data and contemporary MRI metrics.  

Methods:

We identified patients in our institutional AS program enrolled after diagnostic biopsy with grade group (GG) 1 disease and subsequent confirmatory biopsy with <GG1 disease between 2011 and 2021. Both were transrectal biopsies. CCLO was calculated by summing the total number of involved sextants (in addition to transition zone) positive for cancer in both diagnostic and confirmatory biopsies (score range 1-7). We stratified patients into low, intermediate, and high CCLO scores (1, 2, >3, respectively). We excluded those missing confirmatory biopsy core data. The primary outcome was disease reclassification to GG2 or higher on subsequent biopsies. Kaplan-Meier survival analysis with the log-rank test was used to compare time to GR between CCLO risk groups. Multivariable analysis was performed among patients with pre-confirmatory biopsy MRI to evaluate whether CCLO was associated with GR after controlling for demographics, PSA density, confirmatory biopsy core data, and highest pre-biopsy PI-RADS score.

Results:

A total of 537 patients fit our inclusion criteria. Median follow-up time was 37 months (IQR: 23-60 mo.). Survival analysis demonstrated a progressively shorter time to GR with increasing CCLO risk group (p<0.001, Figure). At 2 years after confirmation biopsy, the proportion of patients experiencing GR was 6.5% for low CCLO, 12.8% for intermediate CCLO, and 57.4% for high CCLO. At 5 years, the proportions experiencing GR were 20.2% for low CCLO, 35.3% for intermediate CCLO, and 57.4% for high CCLO.  Multivariable analysis in the subset of patients with pre-biopsy MRI and targeted biopsies as indicated for PI-RADS > 3 (n=381) demonstrated a significant association between high CCLO risk group and GR when controlling for demographics, biopsy core data, and PI-RADS score (Table). Increasing age, PSA density > 0.15 and Black race were also associated with GR; PI-RADS score was not significantly associated with GR.

Conclusion:

Our study revealed that increasing CCLO risk group was significantly associated with shorter time to GR. When controlling for PSA density, traditional biopsy core metrics, and PI-RADS score, only the highest CCLO risk group was significantly associated with GR. This suggests a need to further calibrate traditional cancer volume measurement by incorporating cancer location within the prostate rather than just the number or percent of positive cores. Doing so may allow for improved accuracy in determining candidacy for continuation on AS.

Funding: N/A