Reducing Prostate Cancer Racial Disparity: Evidence for Aggressive Early Prostate Cancer PSA Testing of African American Men

Abstract

Introduction

There is continuing controversy about prostate cancer (PCa) testing and the recent AUA guidelines. We hypothesize that the reduction and elimination of racial survival disparity among AAM (high risk group)compared to EAM (intermediate risk group) during the PSA testing era compared to the pre-PSA era strongly supports the use of PSA testing in AAM.

Method

We used Surveillance, Epidemiology and End Results (SEER) data to investigate relative survival disparities between AAM and EAM. To evaluate pre-PSA testing era, we selected malignant first primary prostate cancer in AAM and EAM males, all stages, diagnosed 1973–1994. To evaluate relative survival disparities in the current PSA testing era, we selected malignant first primary local, regional and distant stage prostate cancers diagnosed 1998–2005 to calculate five-year relative survival rates.

Results

Age-adjusted five-year relative survival of prostate cancer diagnosed 1973–1994 in the national SEER data revealed significantly shorter survival for AAM compared to EAM (p < 0.0001). The SEER-based survival analysis from 1995 to 2005 indicated no statistical difference in relative survival rates between AAM and EAM by year of diagnosis of local, regional or distant stage PCa.

Conclusion

We conclude that the elimination of PCa racial disparity of local, regional and metastatic PCa relative survival in the current PSA testing era compared to pre-PSA era as an endpoint to test PSA efficacy as a marker for PCa diagnosis is evidence for aggressive testing of AAM.

Impact

Evidence for screening African American men.

Introduction

African American men (AAM) have a greater risk of dying from prostate cancer (PCa) than European American Men (EAM) (1). The PCa age specific mortality rates are 2.4 times greater among AAM compared to EAM (1). AAM also have a higher incidence of PCa than EAM (age adjusted rate: 220.0 (AAM) vs. 138.6 (EAM) (1, 2). The CDC Behavioral Risk Factor Surveillance Study (BRFSS) from 2007 to 2009 reports similar rates of PSA testing among AAM and EAM age 40 and above, and 81% and 88% of AAM and EAM, respectively, have insurance (3). Thus, similar rates of testing appears to have had minimal effect on the substantial PCa mortality disparity. We have reported evidence that strongly suggests that prostate tumors grow faster among AAM than EAM. Autopsy results demonstrated that PCa starts at similar ages among AAM and EAM with similar stage and grade at diagnosis, but subsequent development of distant metastasis occurs at a disproportionate rate of approximately three AAM to one EAM (4). Therefore, we concluded that PCa progresses faster among AAM than EAM. A study of outcomes after treatment from 1991 to 1996 for clinically localized PCa reported more advanced disease and greater PSA recurrence among AAM compared to EAM ages 40 to 69 years (5). This difference in disease severity and recurrence, in addition to disproportionate mortality among young AAM, provides strong evidence that AAM should be tested more aggressively and at earlier ages than EAM.

In May 2012, the United States Preventive Service Task Force (USPSTF) recommended against PSA screening of healthy men (6). This recommendation was based primarily on the PLCO (Prostate, Lung, Colon and Ovarian) prospective randomized clinical trial which reported no decrease in mortality rate in the screened arm and therefore no PSA screening benefit. This recommendation applied to men in the U.S. population that did not have symptoms that were highly suspicious for PCa, regardless of age, race, or family history. Only 4% of the PLCO study population was AAM (7). The PLCO results, however, understated the impact of screening on PCa mortality due to a high contamination rate in the control arm. During the study, 52% of controls underwent opportunistic PSA testing. In addition, Pinsky et al. reported that 85% of controls had a PSA test before or during the study; only 15% of PLCO controls never had a PSA test (8). Furthermore, a seven year median follow-up time was inadequate for estimating an impact of PSA screening on PCa mortality, as the study was designed to report the mortality rate comparison after 13 years (7).

The 2013 AUA (American Urological Association) PCa Guidelines strongly recommended shared decision-making for men age 55 to 69 years, at intermediate risk, who were considering PSA testing (9). The AUA panel did not recommend routine testing in men between ages 40 to 54 years at average risk. The panel further commented that “for men younger than age 55 years at higher risk (e.g. positive family history or African American race), decisions regarding prostate cancer screening should be individualized and discussed with their doctor.” This statement was based on results from the updated European Randomized Study of testing for Prostate Cancer (ERSPC) that demonstrated a 31% reduction in mortality in the PSA testing arm at 9 years median follow-up (10). This trial included European men only, thus it did not directly apply to AAM. The AUA guideline was clearly inadequate and provided no clear direction or guideline for AAM (11). Thus, it is necessary to provide the most available scientific data to evaluate the benefit of PSA testing of AAM younger than 55.

Level 1 scientific evidence may not be achievable in this country because of the potential for a high PSA testing contamination rate, similar to that reported for PLCO, in any future clinical trial of PSA testing. Therefore, other scientific evidence with reasonable and achievable endpoints must be considered. We hypothesize that intensive education and aggressive and early PSA testing of AAM will result in the reduction and/or elimination of racial disparity in PCa survival among AAM (a high risk group) compared to EAM (an intermediate risk group). We will present data from the pre-PSA and current PSA era to demonstrate the reduction and elimination of PCa racial disparity in relative survival that occurred after the introduction of PSA testing. Survival analyses are utilized to test the efficacy of therapies and therefore should be justified to test the utility of clinical markers between high and intermediate risk populations.

Materials and Methods

Study cohort

The cohort for our study analyses was the accumulated prostate cancer cases in the very large national Surveillance, Epidemiology and End Results (SEER) registry. We used the national SEER registry data to investigate relative survival disparities by race in two successive time periods: the pre-PSA testing era; and the current PSA testing era. We also used the national SEER registry data to investigate incidence and mortality disparities by race. The statistical methods used are described as follows for each of those 3 sets of analyses.

1. Relative survival disparities in the pre-PSA testing era: We selected malignant first primary PCa in AAM and EAM males diagnosed 1973–1994. Cases of all stages were combined into one group, since the national SEER registry data for 1973–1994 does not include prostate cancer stage. Age-adjusted five-year relative survival rates and 95% confidence intervals (CIs) were calculated for each race, by year of diagnosis. Z-scores were also generated by year of diagnosis and for the entire timeframe in SEER*Stat, comparing AAM and EAM. The Z-score for the entire timeframe was used to calculate an overall p-value using Graph Pad.

2. Relative survival disparities in the current PSA testing era: We selected malignant first primary local, regional, and distant stage prostate cancers diagnosed during 1998–2005 (with survival follow-up through December 31, 2010). Five-year relative survival rates, 95% CIs and Z-scores were calculated by year of diagnosis and race for each stage, thus eliminating the need to adjust for stage. Z-scores were also calculated for the entire PSA testing era, comparing AAM and EAM within each stage. The overall p-value was calculated and a graph prepared for each stage, comparing AAM and EAM across the entire PSA testing era. We also investigated survival disparities by treatment category in the PSA era. Within each stage, five-year relative survival rates, 95% CIs, Z-scores, overall p-values and graphs by stage were calculated comparing AAM and EAM for cases that received radiation, for cases that received radical prostatectomy and for the combined group of cases that received either definitive PCa treatment.

3. Incidence and mortality disparities in the current PSA testing era: We performed 3 sets of analyses. First, we calculated age-specific incidence rates of distant prostate cancer in AAM and EAM for ages 40–79 (40–49, 50–59, 60–69, 70–79) from 1995–2010. Rate ratios and p-values were calculated comparing AAM to EAM within each age stratum. Second, we generated age-specific prostate cancer mortality rates (no stage restrictions) by race (AAM, EAM) with rate ratios, 95% CIs and p-values for U.S. deaths from 1995–2010. These data were stratified by 5-year age of death categories (40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74). Within each age group, the rate ratio and p-value was used to evaluate whether AAM had higher risk of prostate cancer mortality than their EAM counterparts. Third, we evaluated Gleason score (2–6 vs. 7–10) in men diagnosed with prostate cancer in the most recent part (2004–2010) of the PSA era, who underwent radical prostatectomy. The racial distribution of patients was determined by 10-year age intervals for ages 40–69 (40–49, 50–59, 60–69). Within each age group, the proportions of men with Gleason score 7–10 were compared by race using the Chi-Square test, without continuity correction.

Results

The SEER dataset for the pre-PSA era survival analysis included N=212,719 cases (AAM=23,782, EAM=188,937) diagnosed in 1973–1994. Age-adjusted five-year relative survival of first primary PCa (all stages) in AAM and EAM males by race, in SEER cases diagnosed 1973–1994 revealed statistically significantly shorter survival for AAM compared to EAM (p-value < 0.0001, Figure 1) during the pre-PSA era.

Figure 1. Age-adjusted Five-Year Relative Survival of Malignant First Primary Prostate Cancer (All Stages) in African-American and European-American Males by Race SEER-18, Year of Diagnosis 1973–1994(survival time through 2010).

Source: Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) Seer*Stat Database: Incidence SEER 18 Regs Research Data + Hurricane Katrina Impacted Louisiana Cases, Nov 2012 Sub (1973–2010 varying) Linked To County Attributes Total U.S., 1969–2011 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch, released April 2013, based on the November 2012 submission.

The SEER dataset for survival analysis in the PSA era included N=309,793 PCa cases (AAM=44,934, EAM=264,859), (local stage: AAM=36,688, , EAM=219,765; regional: AAM=5,390, EAM=33,994; distant: AAM=2,856, EAM=11,100) diagnosed 1998 to 2005. Within each stage there was no significant difference in survival rates between AAM and EAM for either local (P=1.0), regional (P=0.1490) or distant stage (P=0.8399) PCa for the PSA era. Analysis of treatment by each stage also revealed no significant racial disparities. For cases that received surgery there were no significant differences in the PSA era (local: P=1.0, regional: P=1.0, distant: P=0.1416). Likewise, for cases that received radiation therapy there were no significant differences by race in the PSA era (local: P=1.0, regional: P=0.2236, distant: P=0.1004). And likewise, for the combined data-set of cases that received either radiation therapy or surgery, there were no significant differences by race in the PSA era (local: P=1.0, regional: P=0.2059, distant: P=0.0903).

The incidence analysis of distant stage cases in the PSA era included N=2,742 AAM and N=9,804 EAM cases from SEER diagnosed 1995–2010. In each age group, AAM with distant disease had significantly higher PCa incidence than EAM (Table 1). Analysis of local/regional disease cases combined yielded similar results (data not shown). While PSA testing has apparently eliminated racial disparity in the ability to survive for at least 5 years following a PCa diagnosis, there still appear to be disparities in the ratios of new prostate cancer diagnoses and deaths. AAM diagnosed with distant stage PCa had a three-fold increased risk of being diagnosed with a PCa, compared to their same age cohort EAM counterpart (Table 1). AAM also had about a three-fold increased risk of dying from PCa, compared to EAM of the same age cohort (Table 2). This potentially indicates a difference in the biology of prostate cancers in AAM compared to EAM, with AAM having a more aggressive PCa phenotype.

Table 1.

Age-Specific Incidence Rates of Malignant Distant Stage Prostate Cancer in African-American (AAM) and European- American Males (EAM), SEER-13, 1995–2010

Age at Diagnosis	EAM Rate*	AAM Rate*	Rate Ratio	Ratio p-value
40–49	0.70	2.49	3.55	<0.0001
50–59	5.45	16.91	3.10	<0.0001
60–69	19.78	57.76	2.92	<0.0001
70–79	43.16	104.52	2.42	<0.0001

^*Rates are per 100,000

Source: Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence - SEER 13 Regs Research Data, Nov 2012 Sub (1992–2010) <Katrina/Rita Population Adjustment> - Linked To County Attributes - Total U.S., 1969–2011 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch, released April 2013, based on the November 2012 submission.

Table 2.

Age-specific Prostate Cancer Mortality Rates and Rate Ratios in African-American and European-American Males, U.S., 1995–2010

Age at Death	AAM		EAM		Rate Ratio	95% C.I.	Ratio p-value
	Rate	95% C.I.	Rate	95% C.I.
40–44 years	0.55	0.45–0.66	0.18	0.16–0.21	3.00	2.39 – 3.75	< 0.0001
45–49 years	2.42	2.21–2.65	0.76	0.71–0.80	3.21	2.87 – 3.58	< 0.0001
50–54 years	9.01	8.55–9.50	2.81	2.72–2.90	3.21	3.01 – 3.41	< 0.0001
55–59 years	25.30	24.40–26.22	8.21	8.03–8.38	3.08	2.96 – 3.21	< 0.0001
60–64 years	64.23	62.57–65.92	20.55	20.24–20.86	3.13	3.03 – 3.22	< 0.0001
65–69 years	131.97	129.26–134.72	46.07	45.56–46.59	2.86	2.80 – 2.93	< 0.0001
70–74 years	264.63	260.16–269.16	93.05	92.25–93.85	2.84	2.79 – 2.90	< 0.0001

Rates are per 100,000 males.

Source: Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Mortality - All COD, Aggregated With State, Total U.S. (1969–2010) <Katrina/Rita Population Adjustment>, National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch, released April 2013. Underlying mortality data provided by NCHS (www.cdc.gov/nchs).

There is a mortality rate lead time of approximately five years among AAM compared to EAM (Table 2). The mortality rate for AAM ages 40 to 44 is 0.5, similar to the mortality rate for EAM five years older (0.8 for ages 45 to 49 years). Likewise, the mortality rate for AAM ages 45 to 49 is 2.4, close to the mortality rate of 2.8 for EAM five years older (ages 50 to 54 years). The mortality rates increase among older AAM as compared to EAM, and by age 50 to 54 the mortality rate among AAM (9.0) is higher than the rate among EAM approximately five years older (8.2). Early, more aggressive PSA testing and aggressive treatment will be necessary to eliminate this PCa mortality racial disparity.

Discussion

SEER data demonstrate a lead time of approximately five years in age specific PCa mortality rate between ages 45 to 70 years among AAM compared to EAM as well as 2.4 times greater mortality rate (12, 13). Age adjusted five-year relative survival analysis of all PCa cases diagnosed from 1973 to 1994 reveals a significantly shorter survival for AAM than for EAM (p=0.0001), but no racial difference in survival is observed for PCa cases with local, regional or distant stage disease from 1998 to 2005 (14). SWOG (Southwest Oncology Group) trials S8894 (pre-PSA era) and S9346 (current PSA era) reported similar results for distant disease, shorter survival among AAM vs. EAM in the pre–PSA era, but no racial difference in survival in the current PSA era. SWOG current PSA era study also revealed less extensive disease (15). Therefore, there is evidence that PSA testing significantly reduces or eliminates racial disparity in PCa survival.

However, the most important finding in this analysis is the continued PCa racial disparity in the diagnosis of distant disease (16), and the 2.4 times greater PCa mortality rate of AAM compared to EAM (12, 13). Even though the mortality rates for AAM and EAM have decreased during the PSA era, significant PCa racial disparity persists. Since PSA testing has apparently contributed to the reduction and elimination in relative survival and age-adjusted five-year survival differences, more aggressive community education and PSA testing needs to be implemented beginning at age 40 years among AAM. SEER data demonstrate that AAM age 40 to 49 years are more likely to have a higher Gleason score compared to EAM² (p < 0.0001, Table 3), and the mortality rate is 2.8 – 3.2 times greater in AAM compared to EAM ages 40 to 49 years (Table 2) (13).

Table 3.

Gleason Score for Malignant Prostate Cancers Cases who underwent Radical Prostatectomy, SEER-18, 2004–2010

Age at Diagnosis	Gleason Score	EAM	AAM	P-value
40–49	2–6	52.4%	45.4%
	7–10	47.6%	54.6%	<0.0001
50–59	2–6	44.8%	37.6%
	7–10	55.2%	62.4%	<0.0001
60–69	2–6	37.4%	32.4%
	7–10	62.6%	67.6%	<0.0001

Source: Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence - SEER 18 Regs Research Data + Hurricane Katrina Impacted Louisiana Cases, Nov 2012 Sub (1973–2010 varying) - Linked To County Attributes - Total U.S., 1969–2011 Counties, National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch, released April 2013, based on the November 2012 submission.

In addition, we previously reported that high grade PIN and PCa started as early as ages 20 to 30 years, without any difference in prevalence or Gleason grade between AAM and EAM, but at age 40 to 49 the high grade PIN incidence was greater among AAM compared to EAM, 46% vs. 29%, respectively, and that difference persisted with increasing age (4). It has been reported that high grade PIN is associated with aggressive PCa (17). The proportion of men treated for PCa at ages 40–49 with Gleason score 7–10 at diagnosis is significantly higher among AAM compared to EAM ( p < 0.0001, Table 3). In order to prevent a 3:1 disproportionate mortality rate in the decade of 50 to 59 years, diagnosis in the decade of 40 to 49 years is required. Therefore, to eliminate or significantly reduce the mortality rate disparity, we recommend that PSA testing for AAM begin at age 40 years.

It has been suggested that if the biology represented by the grade of the disease is more aggressive among AAM then it may not be possible to eliminate PCa mortality rate racial disparity. We recently reported that PCa is growing faster among AAM versus EAM as stated above and we also reported that genes associated with more advanced PCa are more highly expressed among AAM compared to EAM and may explain our clinical findings. We examined RNA expression of genes associated with PCa in tumors of men who underwent radical prostatectomy to identify upregulated genes and associated functional gene networks and signaling pathways that may contribute to PCa progression. For example inflammatory cytokines IL6, IL8 and IL1B show significantly higher expression levels in AAM compared to EAM (18). These functional cytokines have been reportedly associated with advanced PCa. There was also increased expression of FASN in PCa specimens from AAM versus EAM in our study. FASN germline polymorphisms were reported to be significantly associated with risk of lethal PCa (19). These factors and other biologic/genetic mechanisms likely explain the racial disparity in tumor aggressiveness.

Specifically, even though we present evidence of AAM having several genetic and biological factors associated with aggressive PCa in comparison to EAM, intervention with early PSA testing and aggressive therapy prior to transformation to aggressive and advanced phenotypic expression has drastically reduced/eliminated survival disparity and may eventually nullify PCa racial mortality disparity.

The updated follow-up of the ERSPC study and the Goteborg report clearly show benefit of PSA testing based on a 31% and 44% reduction in PCa mortality rate, respectively (10, 20). However, the issue of “harm” from PSA testing is aggressively debated, specifically with regard to post-treatment erectile dysfunction and urinary incontinence. Both of these complications of therapy can be successfully treated, and some men with erectile dysfunction are not bothered by this loss. But the harm of not being tested for PCa and appropriately treated in a timely fashion has been essentially ignored, especially in a high risk population. The harm is delayed diagnosis of aggressive PCa, the serious morbidity of metastatic disease, and ultimately PCa death. The issue of overdiagnosis and overtreatment are very controversial. There is no clear definition of either. PCa is a genetic and biologic disease that is very heterogeneous and dynamic, not static (4, 18). Until the genetics and biology are understood we cannot clearly define overdiagnosis and overtreatment. In the interim the introduction of active surveillance has reduced overtreatment of low risk disease, however, one must use caution in including young AAM in surveillance. It has been reported that AAM are 3 fold more likely than EAM to have disease progression (21, 22). In addition the harms of diagnostic testing and treatment should be uncoupled. The PLCO trial reported that the percentage of complications associated with PSA blood draw and prostate biopsy was less the 1 % (7)

Piper et al. report, “Men dying of prostate cancer incur significant costs in the last year of life. Based upon recent epidemiological data the cost of death due to prostate cancer in the US is over three quarters of a billion dollars a year,” and the cost of PCa death and care before death was greater than the cost of treatment of local disease (23). The cost of treatment for metastatic PCa has increased drastically. The cost of recently FDA approved biologically targeted therapy for metastatic disease is approximately $8,000.00 per month, and the cost is significantly higher for an immunotherapeutic agent. Since AAM have a 2.4 times greater mortality rate and risk of metastatic disease than EAM, it would be prudent to eliminate this disparity and save these PCa health care system dollars. In addition, in the year before PCa death there may be co-morbidities requiring multiple hospitalizations which include pathologic fractures, lumbar nerve compression, anemia, gross hematuria, and urinary clot retention and retention from bladder obstruction, and uremia.

The cost/benefit ratio strongly favors early and aggressive PSA testing, early diagnosis and appropriate treatment of AAM given the improved survival and disparity reduction. Therefore we strongly recommend aggressive PCa education and testing at age 40 in the AA community to avoid increased PCa health care cost, co-morbidities, premature PCa death, and most importantly to continue the reduction and elimination of racial outcome disparity. This recommendation should be included and highlighted in all PCa testing guidelines.

Our study has several limitations. First, to facilitate the comparison of the racial differences during the pre-PSA era and the current PSA era we had to choose a cut off year which was 1995. Our results might differ somewhat from the use of a different dichotomy. Second, the limitations of the national SEER registry data regarding prostate cancer staging (for 1973–1994) precluded our doing either stage-stratified or stage-adjusted analyses of five-year relative survival rates during the pre-PSA era. Third, other limitations in the SEER stage data meant that for the current PSA era, we could only analyze five-year relative survival rates beginning in 1998 (not 1995).

Our study also has several strengths. First, by utilizing SEER data our results are based on very large sample sizes which are specified throughout the Results section. Second, by using the national SEER data, our results should be generalizable to the entire US population. Third, the identification of an approximate 5-year lead time in mortality among AA prostate cancer cases compared to EA prostate cancer cases is a compelling observation regarding the racial disparity we have investigated.

Conclusion

We reported the reduction and elimination of PCa racial disparity of relative survival of local, regional, and distant disease in the current PSA testing era. Therefore we conclude that PCa racial disparity reduction is a reasonable endpoint to establish the benefit of PSA testing. However, the PCa mortality rate is now 2.4 times greater among AAM compared to EAM. Therefore we recommend robust education and PSA testing as well as digital rectal examination to begin at age 40 for AAM and that this recommendation should be included in all PCa guidelines.