High-volume prostate biopsy core involvement is not associated with an increased risk of cancer recurrence following 5-fraction stereotactic body radiation therapy monotherapy

The local Institutional Review Board (Study # 00001269) approved this single institutional review of patients treated for PCa. Exclusion criteria for the specific investigation of PSA nadir outcomes was as follows: (1) received ADT as a component of treatment, (2) did not have a pathology report available, (3) never achieved a PSA nadir (i.e. PSA < 3 ng/mL), (4) demonstrated disease progression, and (5) less than 2 years of FU. Analysis of core involvement association with oncologic outcomes excluded (3) and (4) above. All patients underwent rigorous surgical pathologic review of the prostate biopsy specimen. Each prostate pathology report was independently reviewed by our team to document the number of cores that were sampled, location, and the Gleason score that was identified. This evaluation included adenocarcinoma core involvement, overall percentage core involvement, and specific GS for a given positive core (PC). All patients were evaluated by a radiation oncologist and deemed appropriate for definitive 5-fraction SBRT. All patients underwent computed tomography (CT)-based radiation treatment planning simulation. A prostate MRI was obtained in the majority of cases at the time of simulation. Patients underwent robotic SBRT with a clinical target volume (CTV) which included the entire prostate and proximal seminal vesicles. A 5 mm isometric expansion of the CTV with a tighter 3 mm posterior margin was used to create the PTV.

Patients were typically followed using serial PSA and clinical examination at 3-month intervals for the first year and subsequently every 6 to 12 months thereafter. PSA nadir was defined as the lowest post-SBRT PSA obtained after at least 2 years of FU. High-volume core involvement was defined in accordance with the NCCN definition of ≥ 50% biopsy core involvement with adenocarcinoma. Descriptive statistics (mean ± standard deviation or median [25th, 75th] percentiles for continuous variables; frequency and percent for categorical variables) were calculated for the overall sample for patient, tumor, and treatment characteristics. A graphical display of PSA nadir was constructed using boxplots for the overall sample as well as stratified by low-, intermediate-, and high-risk groups.

The association between PSA nadir and percentage of PC, continuous percentage of core involvement and age, and PSA and prostate CTV was assessed using Spearman correlation coefficients. Core involvement was then analyzed by giving higher weight to PC involvement of higher-grade grouping. For example, a single core of GS 10 was weighted 5 × a single core involved of GS 6. Grade group weighting was as follows: GS6 – 1x, GS7 – 2x, GS8 – 3x, GS9 – 4x, GS10 – 5×. Analysis of variance was used to assess the association between categorical variables such as PSA, GS, NCCN Risk and continuous percent core involvement. Percent core involvement was dichotomized as < 50% and ≥ 50%. The two groups were compared using the chi-square test or Fisher’s exact test for categorical variables, and the two-sample t-test or Mann–Whitney test for continuous variables.

From 2009 to 2018, 1,590 patients were identified with a median age of 66 years, who underwent robotic SBRT for localized PCa. The distribution of risk grouping was as follows: low (n = 474, 30%), intermediate (n = 1061, 67%), and high (n = 55, 3%). Of note, due to the exclusion of patients who received ADT, there were very few high-risk patients included in this analysis. Median pretreatment PSA for low-, intermediate-, and high-risk PCa was 5.4 ng/mL (IQR: 3.6 to 6.7 ng/mL), 6.6 ng/mL (IQR: 4.7 to 8.7 ng/mL), and 7.7 ng/mL (IQR: 5.5 to 10.1 ng/mL), respectively. The clinical stage distribution was predominantly T1 (n = 1269, 84.49%). All patients were treated with definitive robotic SBRT over five treatment fractions to a total dose of 3500 (n = 1451, 91.26%) or 3625 (n = 131, 8.23%) cGy. Of note, treatment dose information is unknown for eight patients (< 1%) due to the transition from physical to electronic records. Median prostate CTV was 77.61 cc (IQR: 61.3 to 99.2 cc) and was prescribed to a median isodose line of 84% (IQR: 83% to 85%). Detailed patient, tumor, and treatment characteristics are listed in Table 1.

Surgical pathology review demonstrated an overall GS distribution as follows: GS6 (n = 546, 34%), GS7 (n = 993, 62%), GS8 (n = 45, 3%), and GS9 (n = 6, 1%). The median percent biopsy core involvement was 25% (IQR: 16.7% to 41.7%). Median percent core involvement for low-, intermediate-, and high-risk was as follows: 20.7% (IQR: 8.3 to 33.3), 25% (IQR: 16.7 to 41.7), and 33.3% (IQR: 16.7 to 41.7), respectively. Overall, high-volume pretreatment biopsy core involvement was uncommon in this cohort with only 19% of the entire group (n = 303) demonstrating this elevated burden of disease. Biopsy core involvement distribution is illustrated for the overall cohort and stratified by risk group in Fig. 1A-B.

Analysis of pretreatment clinical characteristics was performed to determine if there were any associations with high-volume biopsy core involvement (≥ 50%). When stratified by risk grouping, high-volume core involvement was observed in 14% of low- (66 of 474), 21% of intermediate- (226 of 1061), and 20% of high-risk (11 of 55) patients. Significantly more patients diagnosed with GS7 were found to have high-volume core involvement (73% vs. 60%, p = 0.0001). Similarly, significantly more patients with intermediate-risk disease were found to have high-volume core involvement compared to the rest of the cohort (75% vs. 65%, p = 0.005). In contrast, significantly fewer patients diagnosed with GS6 (25% vs. 38%, p = 0.0001) and high-risk disease (23% vs. 33%, p = 0.0053) were found to have high-volume core involvement. Finally, smaller prostate volume (PV), as defined by the CTV, was associated with high-volume core involvement (70 cc vs. 79 cc, p < 0.0001). There was no significant difference in high-volume core involvement based on patient age or initial PSA, analyzed either as a continuous or discrete variable. Detailed analysis on the association between core involvement and patient characteristics are found within the Additional file 1, 2, 3, 4, 5: Tables 1A, 1B, 1C, 1D, 1E.

Biopsy core involvement was also analyzed as a continuous variable rather than using a discrete 50% breakpoint. Analogously, we observed intermediate-risk disease to display a significantly higher percent core involvement (32.0% ± 19.7%) relative to low- (30.0% ± 17.3%) and high-risk (25.6% ± 17.0%) disease (p < 0.0001). Furthermore, prostate CTV displayed a significant negative correlation with biopsy core involvement (ρ = − 0.14, p < 0.0001). Again, there was no association with patient age (p = 0.95) or pretreatment PSA (p = 0.57) (Supplementary Table 1E).

In the overall cohort, high-volume core involvement was not associated with a higher PSA nadir. In fact, there was a non-significant trend towards higher PSA nadir with low-volume core involvement (0.35 vs 0.34 ng/mL, p = 0.063). Interestingly, when specifically analyzing the intermediate-risk cohort, low-volume biopsy core involvement demonstrated a significantly higher PSA nadir (0.35 vs 0.34 ng/mL, p = 0.004). Moreover, after adjusting for prostate CTV, there was still a statistically significant difference in PSA nadir (0.24 vs. 0.21, p = 0.046). The remaining low- and high-risk cohorts did not demonstrate any significant association with biopsy core involvement and PSA nadir. Box plots of PSA nadir and biopsy core involvement overall and stratified by risk group are displayed in Figs. 2A–D.

When weighted percent core involvement was analyzed as a continuous variable for the entire cohort, there was no significant association with PSA nadir using Spearman correlation coefficients (ρ = − 0.012, p = 0.617). However, when the intermediate-risk cohort was similarly analyzed, a significantly negative correlation was identified between higher weighted core involvement and PSA nadir (ρ = − 0.063, p = 0.049). No similar association was observed in the low- (ρ = 0.054, p = 0.249) or high-risk (ρ = − 0.0721, p = 0.667) cohorts. Figure 3 displays detailed Grade group weighted core involvement analysis.

Core involvement was analyzed to determine its association with biochemical progression free survival (bPFS). With a median FU of 4.3 years (IQR: 3.1 to 6.2 years) for the entire cohort, there was no significant difference in time-to-failure between patients with high- versus low-volume biopsy core involvement (p = 0.234). In contrast, there was a significant association between time-to-failure and weighted total PC (HR = 1.02, 95% CI: 1.00–1.03, p = 0.012). However, after adjusting for prostate CTV this significance vanished (HR = 1.01, 95% CI: 0.99–1.03, p = 0.2562). Each NCCN risk subgroup was then analyzed in similar fashion independently, and there was again no significant association between bPFS and high- versus low-volume core involvement for low (p = 0.168), intermediate (p = 0.345), and high (p = 0.305) risk disease. Correspondingly, there was no significant association between weighted core involvement and each risk group when analyzed independently for low (HR = 0.94, 95% CI: 0.88–1.01, p = 0.115), intermediate (HR = 1.01, 95% CI: 0.99–1.03, p = 0.284), and high (HR = 1.02, 95% CI: 0.99–1.04, p = 0.141) risk disease. Of note, after adjusting for prostate CTV in the high-risk cohort, subjects with low-volume core involvement had a lower risk of progression (HR = 0.29, 95% CI: 0.09–0.99, p = 0.048). Multivariable analysis also did not demonstrate dichotomized PC involvement as associated with PCa failure. However, given the small number of patients with high-risk disease, the clinical significance of this finding is unclear. Figures 4A–D display detailed weighted core involvement analysis stratified by risk group.