Predicting survival for men with clinically localized prostate cancer

What do we need in contemporary practice?


  • Peter Albertsen MD

    Corresponding author
    1. Department of Urology, University of Connecticut Health Center, Farmington, Connecticut
    • Department of Urology, University of Connecticut Health Center, Farmington, Connecticut
    Search for more papers by this author

  • See referenced original article on pages 69–74, this issue.

In their article entitled Nomogram incorporating PSA level to predict cancer-specific survival for men with clinically localized prostate cancer managed without curative intent, Kattan et al.1 provide a tool to help men with newly diagnosed prostate cancer evaluate whether to seek aggressive intervention or to follow a program of active surveillance. They reassure the reader that the nomogram has been tested and validated and has been calibrated over a 120-month period following diagnosis. Will men find it useful? The answer depends on how well the study population mirrors the men seeking guidance.

The study population identified by the authors was drawn from 6 English cancer registries and includes men diagnosed between 1990 and 1996. The authors appropriately exclude men with clear evidence of metastatic disease, men who died of any cause within 6 months of diagnosis, and men diagnosed with other life-threatening cancers. All of the men appear to have had clinically localized prostate cancer at the time of diagnosis, but a more careful review of table 1 shows that 75% of the men had a PSA greater than 7.3 ng/mL, 75% of the men were age 67 or older, and 33% of the men received hormonal therapy within 6 months of diagnosis. Contemporary Gleason scores were unavailable for 65% of the men and clinic staging was unavailable for 40% of the men.

Why should readers consider these factors carefully? Because contemporary North American men with newly diagnosed prostate cancer frequently have very different characteristics when compared to English men diagnosed over a decade ago. How is this possible? The differences stem from the dramatically different rates of screening conducted in these 2 populations.

To understand this point, readers need to consider carefully the impact of annual testing for prostate-specific antigen (PSA) over time. PSA testing was not routinely performed in England during the 1990s. Men with clinically localized prostate cancer were frequently identified by transurethral resection of the prostate (46% of the cases in the study cohort) or by prostate biopsy to evaluate a prostate nodule. Biopsies performed to evaluate an elevated PSA test were performed infrequently during this time period. The English men included in the Kattan database were identified because they presented with disease that was clinically evident or that had grown to a sufficient volume that it could be detected at the time of transurethral resection.

Contrast these findings with the typical scenario in contemporary North America. The diagnosis of prostate cancer by transurethral resection is relatively rare and prostate nodules are encountered infrequently. PSA testing has dramatically increased the number of men diagnosed with prostate cancer such that the incidence of this disease in the US has essentially doubled since the 1980s.2 In fact, in 2007 the majority of men with localized prostate cancer are now identified because of a rising PSA test, not simply an elevated PSA test noted on an initial screening test.

As a consequence, there are a growing number of men who are diagnosed with minimal volume prostate cancer (usually involving less than 30% of one or two cores), which is often read as low-grade cancer (usually Gleason score 3 + 3 = 6). At this point we have no real understanding of the natural history of screen-detected prostate cancers, especially the low-grade, small-volume cancers.

Because of the normal variation in serum PSA levels there is a significant chance that many men will occasionally have an abnormal PSA value. Many of these men are often subjected to transrectal ultrasound and prostate biopsy and some will be diagnosed with prostate cancer. Data from the finasteride chemoprevention trial have shown us that over 20% of asymptomatic men with normal PSA values harbor small foci of disease.3 As a consequence of repeated PSA testing many of these men are being diagnosed with prostate cancer. It is uncertain how many of these men are destined to die from their disease.

Another important implication of repeated testing for PSA is lead time. Screening advances the date of diagnosis when compared with the date of clinical presentation. Draisma et al.4 estimate that at age 55 the lead time associated with PSA testing is approximately 12.3 years and the likelihood of detecting clinically insignificant disease is 27%. By age 75 the lead time associated with PSA testing decreases to 6.0 years, while the likelihood of detecting clinically insignificant disease rises to 56%. The Kattan et al. nomogram predicts disease-specific survival up to 5 years following diagnosis. Since most of the English patients were not detected as a result of PSA testing, the disease-specific survival estimates do not include the lead time typical of North American men.

Do these factors impact how North American men might use the Kattan nomogram? The answer is “probably.” North American men with newly diagnosed prostate cancer who are considering active surveillance usually have the following characteristics: diagnosis made by needle biopsy, stage T1c, Gleason score 3 + 3 = 6, percent cancer <5%, and baseline PSA <10. None of these men should be considering androgen withdrawal therapy within 6 months of diagnosis. Based on the Kattan et al. nomogram, their point score would total between 0 and 10. In fact, the only factor that will really add any points for many men is age. For a man age 50 at the time of diagnosis, his total point score would be 30, suggesting a 5-year disease-specific survival of about 94%. A man age 76 would have a disease-specific survival of 85%. These estimates are not that different from the graphs we published in JAMA that were also based on a patient cohort diagnosed on the basis of clinical presentation rather than PSA testing.5 The Kattan et al. nomogram provides important information for men who have localized prostate cancer detected clinically, but it is much less helpful for the typical North American man diagnosed with prostate cancer following repeated PSA testing who is contemplating whether the disease uncovered is clinically significant or not and whether intervention might make a difference.

Randomized trials are currently under way that will provide us with important new insights concerning how to diagnose and manage this disease, but results from these studies are still many years away.6, 7 How should we manage this disease now? Most physicians recommend aggressive therapy for men with Gleason score 7 disease or higher. Men with Gleason score 6 disease have much more to ponder. Although still controversial, a growing number of physicians are advocating a program of active surveillance for this disease. Ideal candidates are men who harbor small volumes of low-grade prostate cancer. Typically, these men have 1 or 2 cores positive out of 12 that contain less than 30% of Gleason 3 + 3 disease. These men have a very low risk of disease progression during the first 10 years following diagnosis.

To identify the subset of men with clinically significant disease, researchers have suggested obtaining serial PSA tests every 4 to 6 months and serial prostate biopsies every 1 to 4 years.8 Disease progression is signaled by a rise in serum PSA, and/or and increase in the volume or grade of cancer. Early reports from these protocols suggest that men who fit the criteria of ‘progression’ have long-term outcomes that do not appear to be compromised by the delay in surgery or radiation.

PSA testing is uncovering a large pool of men who harbor small-volume, low-grade disease. These men are different from the English men diagnosed a decade ago. Until better genetic markers become available we should use PSA values along with repeat prostate biopsies and a tincture of time to try to separate those men with indolent disease from those who require more aggressive treatment. A nomogram would be extraordinarily helpful, but unfortunately the nomogram developed by Kattan et al. has too short a follow-up and focuses on patients we see infrequently in contemporary practice.