Race, tumor location, and disease progression among low‐risk prostate cancer patients

Abstract Background The relationship between race, prostate tumor location, and BCR‐free survival is inconclusive. This study examined the independent and joint roles of patient race and tumor location on biochemical recurrence‐free (BCR) survival. Methods A retrospective cohort study was conducted among men with newly diagnosed, biopsy‐confirmed, NCCN‐defined low risk CaP who underwent radical prostatectomy (RP) at the Walter Reed National Military Medical Center from 1996 to 2008. BCR‐free survival was modeled using Kaplan‐Meier estimation curves and multivariable Cox proportional hazards (PH) analyses. Results There were 539 eligible patients with low‐risk CaP (25% African American, AA; 75% Caucasian American, CA). Median age at CaP diagnosis and post‐RP follow‐up time was 59.2 and 8.1 years, respectively. Kaplan‐Meier analyses showed no significant association between race (P = .52) or predominant tumor location (P = .98) on BCR‐free survival. In Cox PH multivariable analysis, neither race (HR = 1.18; 95% CI = 0.68‐2.02; P = .56) nor predominant tumor location (HR = 1.13; 95% CI = 0.59‐2.15; P = .71) was an independent predictor of BCR‐free survival. Conclusions Neither race nor predominant tumor location was associated with adverse oncologic outcome.

outcomes in low-risk CA and AA men who underwent radical prostatectomy (RP). Two studies that examined biochemical recurrence (BCR)-free survival after RP showed no differences for CA vs AA men, 4,5 while two other studies did find a difference between CA and AA men. 3,6 In general, studies that found no difference in BCR-free survival across race also found few differences in adverse pathology. [3][4][5][6] One anatomical feature of the prostate that has been less explored for short-and long-term CaP outcomes, both independently and jointly with race, is predominant tumor location, specifically, harboring a predominant anterior tumor could lead to poorer oncologic outcomes for CaP patients, if such tumors are more difficult to detect through standard diagnosis procedures. 7 Both Faisal and colleagues 8 and Tiguert and colleagues 9 found that AA men were more likely to harbor anterior tumors than CA men. 8,9 In contrast, prior work conducted in this study setting found no difference in the prevalence of anterior tumors among AA and CA men treated with RP at the Walter Reed National Military Medical Center (WRNMMC). 10 To further understand the social and/or biological underpinnings of CaP progression, a racially diverse, surgically treated cohort of NCCN-defined low-risk men enrolled at WRNMMC, an equal access military health care center, was examined. The aim of this study was to examine the independent and joint roles of self-reported race and predominant tumor location on BCR-free survival, in a surgical cohort for whom detailed anatomical classification of prostate tumor location was possible.

| Study design and participants
A retrospective cohort study was conducted on patients enrolled in the WRNMMC Biospecimen CaP Repository linked to the Center for Prostate Disease Research (CPDR) Multicenter National Database who self-reported as Caucasian (CA) and African American (AA) and who underwent RP for treatment of CaP at the WRNMMC between January 1, 1996 and December 31, 2008. The study cohort was further restricted to those with low-risk CaP, per National Cancer Comprehensive Network (NCCN) guidelines (ie, clinical T stage ≤ pT2a, prostate-specific antigen (PSA) <10 ng/mL, and biopsy Gleason score ≤6) 11 with a life expectancy of more than 10 years. Patients were excluded from the study if they underwent neoadjuvant therapy treatment, or adjuvant treatment (defined as treatment within six months of RP), and one patient who was misassigned primary treatment type and one patient for whom accurate staging could not be accurately assigned were also removed ( Figure S1). Detailed demographic, clinical treatment, pathologic, and outcomes information was collected as part of routine patients follow up on all CPDR enrollees. Further details about the biospecimen repository and database have been reported previously. 12 The repository and database have Institutional Review Boards (IRB) approval at the WRNMMC and the Uniformed Services University of the Health Sciences (USUHS).

| RP Specimen processing and clinicopathologic variables
All RP specimens were processed by whole mount and sectioned at 2.2-mm as previously described. 13 Pathologic parameters were measured based on evaluation by central pathology review (IS) including tumor volume (cc), pathologic T stage (pT2, pT3-pT4), 2014 International Society for Urological Pathology (ISUP) Gleason score (≤6, 3 + 4, 4 + 3, ≥8), 14 surgical margin status (negative, positive), extra-capsular extension (negative, positive), and seminal vesicle invasion (negative, positive). All tumors were regraded based on the ISUP 2014 Gleason grade parameters by a single pathologist (IS). Because only 22% of men had a nodal dissection, nodal status was not examined. Clinical variables included age at CaP diagnosis (years), post-RP follow-up time (years), time from biopsy to RP (months), PSA level (ng/mL) at time of CaP diagnosis, tumor volume (cc), tumor volume (after removal of microscopic tumors) (cc), number of total biopsy cores, number of positive biopsy cores, and percent of positive biopsy cores.

| Independent study variables: Selfreported race and tumor location
Self-reported race categories of interest to this study were CA and AA. Tumor location was assigned in the following manner: the prostate gland was divided into six regions (IS): Anterior, anterior lateral, lateral, posterior lateral, posterior, or peri-urethral ( Figure 1). RP specimens were evaluated and the predominant tumor was assigned to a region of the prostate by determining the anatomical location of the largest portion of the index tumor (the tumor with the highest 2014 ISUP Gleason score and/or the largest volume). Diffuse predominant tumors were those that included involvement with multiple prostate gland regions, spanning anterior and/or anterior lateral, lateral and/or peri-urethral, posterior lateral and/or posterior. Predominant tumors located in either the anterior prostate or anterior lateral prostate were collectively referred to as anterior predominant tumors. 10 Predominant tumors located in either the lateral, posterior lateral, posterior, or peri-urethral prostate, or diffuse predominant tumors were collectively referred to as nonanterior predominant tumors. Following the Epstein et al guidelines for "insignificant tumors", 15 microscopic tumors were defined as those with a volume <0.2 cc, without seminal vesicle invasion and Gleason score <8, in any region of the prostate.

| Dependent study outcome
BCR was defined as two successive post-RP PSA levels ≥0.2 ng/mL or initiation of salvage therapy for a rising PSA. 16 BCR was modeled as a time-dependent study endpoint with three possible outcomes: achieved endpoint, censored at date of last known medical visit or death, or achieved end of study with no event.

| Statistical analysis
Descriptive distributions were examined in the overall cohort, as well as stratified for race (CA vs AA) and stratified for predominant tumor location (anterior vs nonanterior). The chi-square test was used to compare categorical variables and the Mann-Whitney U test was used to compare continuous variables. In contingency tables which had ≥20% of cells <5, the Fisher's exact test was used. Kaplan-Meier unadjusted estimation curves were used to model BCR-free survival stratified by race and by predominant tumor location. Multivariable Cox proportional hazards analysis was used to calculate hazard ratios (HRs) and their 95% confidence intervals (95% CIs) for race and predominant tumor location as independent predictors of BCR-free survival. Models were adjusted for the potential confounders: age at CaP diagnosis, PSA level at diagnosis, pathologic T stage, surgical margin status, and 2014 ISUP Gleason score. All statistical analysis was performed using SAS version 9.4 (North Carolina) and reported p-values are based on two-sided tests (summary alpha = 0.05).

| RESULTS
There were a total of 539 eligible patients of whom 137 (25.4%) were AA and 402 (74.6%) were CA (Table 1). Median age at time of CaP diagnosis and follow-up time were 59 and 8 years, respectively. Few differences in    greater percent positivity in biopsy cores, and greater pT2 disease. In this low risk cohort, the percent of those whose disease was upgraded to ISUP Gleason 4 + 3 or 8-10 at time of RP did not differ across race or tumor location status. However, there was a slightly greater proportion of patients upstaged to pT3-4 at RP across tumor location status, with greater advanced stage observed in the nonanterior tumor patients (20% vs 10%, P < .05). During this study period, 67 (12.4%) patients developed BCR. Unadjusted Kaplan-Meier estimation curve analysis demonstrated no difference in BCR-free survival across race (P = .52) or predominant tumor location (P = .98) (Figure 2A,B). Similarly, in multivariable analysis, neither race nor predominant tumor location was an independent predictor of BCR-free survival, after adjusting for multiple clinicopathologic characteristics (HR = 1.18; 95% CI = 0.68-2.02; P = .56 and HR = 1.13; 95% CI = 0.59-2.15; P = .71, respectively) ( Table 2) (Table S1).
There were 104 (19%) men who had single focal tumors and 435 men (81%) who had multifocal tumors. Of the single focal tumors, there were 12 (11%) men who had anteriorly located tumors and 92 (89%) men who had nonanteriorly located tumors. Of the multifocal tumors, 166 men (38%) had at least one anteriorly located tumor and 269 men (62%) who had no anteriorly located tumors. Clinical features of men with single focal and multifocal tumors were similar. Single focal tumors were smaller and less likely to be anteriorly located than any of the multifocal tumors (P = .003 and P = .006, respectively) and the first multifocal tumor was more likely to be of higher grade than either single focal tumors or the second or third multifocal tumor (P < .001).
To confirm consistency in study results, the analysis was repeated with removal of diffuse or microscopic tumors. When diffuse tumors (N = 38) or microscopic tumors (N = 98) were excluded from the analysis, study results remained unchanged (data available upon request).

| DISCUSSION
In this study, a racially diverse cohort of NCCN-defined lowrisk CaP patients with equal health care access was examined to clarify the independent and joint roles of self-reported race and predominant tumor location on BCR-free survival. This study supports that neither AA race nor anterior tumor location is predictive of BCR-free survival, when examined independently or jointly.
In our previous findings, in the same study setting, Mygatt and colleagues 10 observed no difference between tumor location and BCR-free survival or race. 10 Key differences in this study was exclusive focus on the NCCN-defined low-risk cohort and updated assignment of tumor location, reviewed by multiple pathologists (IS, AB, GW, WG), expanded through 2008, with both race and tumor location examined concurrently in one multivariable model.
Two other studies examined race and tumor location; however, neither study examined race and tumor location individually and jointly with BCR-free survival as an endpoint. 8,9 Tumors in the anterior portion of the prostate are more difficult to detect during standard posteriorly approached biopsy procedures, which may lead to missed or incorrectly staged and graded tumors. 8 In our study cohort, 12% of men with single focal tumors had an anteriorly located tumor and 38% of men with multifocal tumors had at least one anteriorly located tumor. Our finding that single focal tumors were smaller than any of the multifocal tumors could make detection of single focal anteriorly located tumors harder to detect than multifocal anteriorly located tumors. The two other studies that examined race and tumor location did compare the single focal tumors to multifocal tumors. Faisal and colleagues 8 counted strikingly more anterior tumors in both CA and AA  men than we counted, with 29% and 51% (P = .003) of prostate tumors located anterior to the urethra in CA and AA men, respectively. 8 Tiguert and colleagues 9 results were more similar to our findings with 11% and 16% (P = .045) of prostate tumors located anterior to the anterior-posterior diameter in CA and AA men, respectively. 9 Similar to our study, both Faisal and colleagues 8 and Tiguert and colleagues 9 counted only the tumor with the highest Gleason score and/or largest volume. Faisal and colleagues 8 examined men with very lowrisk CaP, enabling these very small tumors to be precisely mapped only to one region in the prostate. When our analysis was restricted to microscopic tumors, there was a slighter larger difference between prevalence of predominant anterior tumors across race; however, the overall percent of anterior tumors was still comparable for both racial groups. Tiguert and colleagues 9 examined clinically localized prostate cancer. Across race, there were few differences in clinicopathologic features such as tumor volume, pathologic T stage, 2014 ISUP Gleason score, surgical margin status, and extra-capsular extension due in part to the equal access to healthcare in our military cohort. Margin positivity and pathology Gleason stage were the major predictors of BCRfree survival in our study, which only slightly differed by race likely due to smaller numbers of AA men (Table S2), while race and predominant tumor location did not predict BCR-free survival. The two previous studies that strictly included low-risk patients with equal access to health care did not present results for margin status or Gleason stage; however, the SEARCH (Shared Equal Access Regional Cancer Hospital) study found no association (HR = 1.11, 95% CI = 0.81-1.50, P = .52), 17 while a study from New York Harbor VA hospitals found an association at 5 years (98% CA vs 82% AA, P = .006) for BCR-free survival, but most likely lacked sufficient CA men for this finding to be replicated. 3,4 Results from the SEARCH study with allrisk patients also showed no association between CA and AA race and CaP metastasis (HR = 1.21, 95% CI = 0.87-1.57, P = .26), CaP specific death (HR = 1.00, 95% CI = 0.61-1.64, P = .99), and overall death (HR = 1.02, 95% CI = 0.90-1.17, P = .76). 17 Although these results from equal access health care centers are in sharp contrast to US National statistics that consistently show considerably worse long-term CaP outcomes for AA men compared to CA men, 18 recent adjusted analysis of National CaP data also show reduced disparity between AA and Ca men with long-term CaP outcomes. 18,19 Our findings lend support to the oncological safety of Active Surveillance in low risk patients, irrespective of patient race or predominant prostate tumor location. While our cohort was not restricted to CaP managed on AS, detailed examination of anatomical location of prostate tumors would not have been possible without examining a cohort whose prostate was surgically removed. However, this study was restricted to include only men who would be candidates for CaP management on AS (ie, NCCN low risk, life expectancy ≥10 years at diagnosis) but who instead underwent RP and donated their prostatectomy specimen.
Each RP specimen was re-graded by a single pathologist (IS) using the updated 2014 ISUP Gleason grading system instead of the pre-2014 grading system. This regrading resulted in additional upgraded tumors, which is consistent with other studies that have examined upgrading pre-and post-2014 Gleason grading system. 20,21 In our study, there were 200 (36.0%) patients who were reclassified from pre-2015 Gleason grade 6 to 2014 ISUP Gleason grade 3 + 4, 4 + 3, or 8-10 disease. Under the new 2014 ISUP Gleason grading system, however, upgrading should be less extensive than previously reported. 14 There are some limitations to consider in interpreting our findings First, the methodology to assign tumor location was one of several methods. 22 Second, the cohort included men who underwent RP during a time period when changes were made to prostate biopsy regimens, the Gleason grading system, and AS eligibility criteria. And third, we were somewhat underpowered to detect a weak to modest association between race or predominant tumor location and BCR-free survival. With our sample size, we had 14%, 52%, 85%, and 97% power to detect an association size of 1.10, 1.25, 1.40, and 1.55, respectively (P = .05, median time to BCR for CA men = 8 years, and follow-up time = 20 years).
In conclusion, our findings show no difference between race or predominant tumor location, both independently and jointly, on BCR-free survival, in a cohort of men who underwent RP at an equal access health care center. This is a single institute study that benefited from detailed anatomical classification of prostate tumor location. Although our findings suggest that active surveillance in low-risk patients may be oncologically safe regardless of race or predominant tumor location, other studies are needed to confirm whether active surveillance is safe for black men.

ACKNOWLEDGMENTS
This work was performed under the Cooperative Research Agreement Funds with Uniformed Services University for the Health Sciences (Award #HU0001-10-2-0002).

CONFLICTS OF INTEREST
None declared.

AUTHOR CONTRIBUTIONS
Justin G. Mygatt contributed to conceptualization and writing the original draft; Jennifer Cullen also contributed to conceptualization, methodology, and writing the original draft; Samantha A. Streicher was involved in formal analysis, methodology, and writing the original draft;