Dr. Powell is to be congratulated for an outstanding review article on prostate cancer in African-American men. As he points out, the age-adjusted incidence of prostate cancer in African-American (black) males is 50% higher than that in Caucasian (white) men, and black men have the highest incidence of prostate cancer in the world. Differences between blacks and whites in the probability of being diagnosed with prostate cancer (9.6% vs 5.2%), lifetime prostate cancer-specific mortality (3% vs 1.4%), and 5-year survival (65% vs 78%) are all indicative of a major public health problem in the black male population.
The etiology for these racial differences in the clinical behavior of prostate cancer is unknown; hormonal, nutritional, genetic, behavioral, and socioeconomic status factors have all been implicated. Now, in the late 1990s, as more research funding is finally being devoted to prostate cancer, it is critically important to find the cause or causes of this racial disparity. Our research group, the Department of Defense (DoD) Center for Prostate Disease Research (CPDR), is funded to study prostate cancer and disease in the DoD health-care system. Because the CPDR has a mandate to study an equal-access, geographically diverse system that provides health care for a large number of African-American men, it is a good setting in which to investigate this issue. I will review our ongoing investigations in this area and compare/contrast our findings to Dr. Powell's exemplary work.
Use of PSA in Black Men
Despite the racial disparity noted above, encouraging recent data from the Radiation Therapy Oncology Group (RTOG), US military, and Veterans Administration suggest that if black men are afforded the same access and medical care as white men, the outcome disparity may be minimized or eliminated. Armed with this information, we may be able to affect prostate cancer in this population with increased public awareness, early detection programs, and proper detection tools.
For a number of years, the American Urological Association (AUA), American Cancer Society (ACS), and American College of Radiology have recommended that early detection programs using digital rectal examination (DRE) and prostate-specific antigen (PSA) be started at age 40 in African-American men. Despite this recommendation, until now, no data existed to document the value of PSA testing for the early and accurate diagnosis of prostate cancer in this population. The finding by our group and others that black men with prostate cancer had higher PSA values at diagnosis, even after adjustment for patient age and tumor stage, grade, and volume, created an urgent need to examine the proper PSA "normal" ranges for black men.
Age-Adjusted References Ranges for Black Men
Our article, "Age specific reference ranges for serum prostate specific antigen in black men," published in the August 1, 1996, issue of the New England Journal of Medicine documents the outstanding ability of PSA to detect prostate cancer in both Caucasian and African-American men and develops age-adjusted PSA reference ranges for maximal cancer detection in this high-risk group of men. In this study, between January 1991 and May 1995 serum PSA concentration was determined for 3,475 men without clinical evidence of prostate cancer (1,802 Caucasian, 1,673 African-American) and 1,783 men with prostate cancer (1,372 Caucasian, 411 African-American). All PSA examinations were performed in a central, single laboratory using the Abbott IMx assay (normal, 0 to 4 ng/mL).
Prostate-specific antigen concentration was analyzed as a function of age and race to determine the operating characteristics of PSA for the diagnosis of prostate cancer. Serum PSA concentration correlated directly with age for both African-American and Caucasian men (r = .40, P = .001 for African-Americans; r = .34, P = .0001 for Caucasians). African-American men had significantly higher serum PSA concentrations than Caucasian men (P = .0001)(Figure 1). When sensitivity was plotted against specificity, the area under the receiver-operator-characteristic (ROC) curve was .91 for African-American men and .94 for Caucasian men, indicating that the PSA test is an excellent early detection tool. For comparison, the Pap smear for cervical cancer, which is an accepted clinical screening test, has an ROC value of only .70.
When we calculated age-specific PSA reference ranges by the identical methodology used by Oesterling and colleagues in their 1993 study of Olmstead County, Minnesota, we found very similar values for white men but higher values for black men. These ranges were 0 to 2.4 ng/mL for black men age 40 to 49, 0 to 6.5 ng/mL for men age 50 to 59, 0 to 11.3 ng/mL for men age 60 to 69, and 1 to 12.5 ng/mL for men age 70 to 79. We then tested these new ranges in our group of black men with prostate cancer to determine how the ranges would have performed if they had been used to detect the cancers. Unfortunately, these markedly higher ranges would have missed 41% of the cancers (sensitivity of only 59%).
The reason that these traditionally derived ranges performed so poorly is simple: They represent the 95th percentile of values in the black controls. Because there is more variability of PSA results in blacks without evidence of cancer, there is more skewness, which pushes the 95th percentile farther to the right (higher). This higher range, however, is not clinically useful. Accordingly, we developed age-adjusted reference ranges for black men with prostate cancer, selecting upper limits of normal PSA by decade to maximize cancer detection. In other words, we developed reference ranges by decade in the men with prostate cancer by using the fifth percentile of PSA values. Only the lowest 5% of prediagnosis PSA values in the black men with cancer are "normal," and the remainder (95%) are above normal (95% sensitivity).
We call these ranges the Walter Reed/Center for Prostate Disease Research (WR/CPDR) age-specific reference ranges for maximal cancer detection (Table 1). They maximize sensitivity (cancer detection) without undue loss of specificity (false-positive results, unnecessary transrectal ultrasound/biopsy). These values for maximal cancer detection for black and white men are compared to the traditional normal ranges (0 to 4 ng/mL) and the previously developed age-specific reference ranges in Table 2. Now that we have shown that PSA is an outstanding tool for prostate cancer detection in African-American men, as well as in Caucasian men, we need to promote public awareness and encourage African-American men to be tested, such as Dr. Powell and colleagues did in their landmark Detroit Education and Early Detection (DEED) study.
Is African-American Race a Prognostic Factor for a Worse Outcome?
There is an ongoing debate as to whether African-American race, itself, is a prognostic factor for worse outcome in men with localized (and advanced) prostate cancer. As noted above, data from the RTOG, US military, and Veterans Administration suggest that race alone is not a prognostic factor. Conversely, a recent Surveillance, Epidemiology, and End Results (SEER) database study in metropolitan Detroit found that, at all stages of disease, blacks had a poorer survival than whites; this was especially evident in younger men (less than 65 years of age). Powell et al also found that survival was worse for African-American men under age 65 but better for men over age 65, and used the phrase, "ethnic survival crossover," to describe this phenomenon. More study needs to be done to determine whether younger black men do have a worse survival, and, if so, what causes this disparity.
In an attempt to answer these questions, our group has been comparing outcome and tumor characteristics by race among men who have undergone radical prostatectomy. In a consecutive group of 91 radical prostatectomy patients, 28 (30.8%) of whom were African-American, blacks had higher tumor volumes in all clinical stage categories (Table 3). Overall, the black men had a mean tumor volume of 5.42 cc, as compared with a volume of 2.10 cc in the white patients. This greater tumor volume was associated with more adverse pathologic features of the radical prostatectomy specimens and worse outcome.
In a follow-up study, we compared 518 black and white men who had undergone radical prostatectomy at Walter Reed Army Medical Center between 1975 and 1995. Table 4 compares the pathologic variables of the radical prostatectomy specimens in these two groups. There were nonsignificant trends for blacks to have higher pathologic stage and Gleason grade. Most strikingly, the margin-positive rate was 50.5% in blacks, as compared with 38.4% in whites (P = .038). These statistics are very similar to those reported by Powell and associates.
Our survival analysis for black and white men was very interesting and points to the need for careful multivariate analysis when trying to determine whether race is a prognostic factor. Overall, the African-American men had a lower recurrence-free survival. In other words, at a mean follow-up of 23.3 months, black men were more likely to have had a recurrence (defined as a rising PSA as the first evidence of failure). Even in one multivariate analysis that adjusted for pathologic stage, grade, and pretreatment PSA and acid phosphatase, black race remained an independent prognostic factor. However, in another multivariate analysis that included margin positivity in the model, race was no longer an independent prognostic factor (P = .083).
Because clinical stage and pathologic stage categories may not accurately reflect tumor volume and subtle pathologic features, such as margin positivity and volume of high-grade cancer, it will be imperative to include these factors in any multivariate analysis aimed at determining whether race affects outcome. The Armed Forces Institute of Pathology and CPDR are currently collaborating on a prospective radical prostatectomy study to compare three-dimensional tumor volumes and careful quantitative histology with outcome in black and white men.
Is Biology or Behavior and Access to Blame for the Poor Pathologic Findings?
Even if race itself is not found to be an independent prognostic marker when these comprehensive pathologic assessments are included, we will still be left with the question of what is responsible for the worse pathologic findings in black men. Considering that the black men in our studies were 1 to 3 years younger than the white men and yet have bigger tumors with more adverse pathologic features, is biology or behavior and access to blame? My own bias is that behavior and access are largely at fault. African-American men simply have not been educated about prostate cancer and the need for early detection. If early detection programs were more universally available to black men starting at age 40 and were accepted, I believe that we would detect the majority of prostate cancers when they were smaller and with fewer adverse pathologic features and would eliminate the current racial survival disparity.
This contention appears to be borne out by the early DEED study results obtained by Powell and associates. Compared to a nonscreened population, the screened African-American men undergoing radical prostatectomy had a significantly higher rate of organ-confined disease and lower recurrence rate at early follow-up. By instituting education about prostate cancer and screening PSA in this high-risk population, the DEED project increased the organ-confined rate in these men to 65%--a gratifying result. Furthermore, at a mean follow-up of 18 to 20 months, of the 15 DEED men screened who underwent a radical prostatectomy, only 1 (7%) suffered a recurrence, as compared with 25% of the nonscreened clinic patients.
Also encouraging was the finding that the DEED screened men had lower mean and median initial PSA values. Considering that pretreatment PSA correlates directly with tumor volume in both black and white radical prostatectomy patients, the DEED data suggest that screened African-American men are being diagnosed when their tumors are smaller. It is tantalizing to speculate that if we use the lower PSA reference range of 2.0 ng/mL developed by WR/CPDR in men between 40 to 49 years old, as discussed earlier, and encourage screening in these younger African-American men, we can have even a greater impact.
Finally, I strongly agree with Dr. Powell that based on the DEED data, we cannot afford to wait to promote early detection in this high-risk population of African-American men.