Commentary (Kuban): Prostate-Specific Antigen as a Marker of Disease Activity in Prostate Cancer

August 1, 2002

The authors present an excellent review of prostate-specific antigen (PSA), bringing us up to date on the large body of information that has been collected since this marker came into clinical use in the mid-1980s. It is hard to believe that we have had this tool for nearly 20 years. Much has been learned.

The authors present an excellent review of prostate-specific antigen (PSA), bringing us up to date on the large body of information that has been collected since this marker came into clinical use in the mid-1980s. It is hard to believe that we have had this tool for nearly 20 years. Much has been learned.

An Invaluable Tool

We have become more knowledgeable about this serum protein in terms of laboratory assay standardization, reference ranges based on patient age, and its utility as a screening tool. At our institution, the percentage of patients with prostate cancer diagnosed solely by serum PSA-prompted biopsy (T1c) has steadily increased from 6% in 1989 to 50% in 2000. At community hospitals, which are not tertiary referral centers, this percentage may be even higher. Although many authors have questioned how many of these tumors would have been insignificant and non-life-threatening had they never been diagnosed, the latest statistics show that prostate cancer mortality rates have been decreasing by 4% each year since 1994.[1]

A recent review of our stage T1c patient group, in fact, shows a heterogeneous mixture of tumor histology and PSA levels. High-grade disease is present in 30% to 40% of patients (Gleason score 7-10), and 36% would not be considered low risk by PSA level (> 10 ng/mL). The assumption, therefore, that PSA screening is only as useful as treating the earliest stage and least aggressive tumors would seem to be erroneous based on analyses such as this. No doubt the screening controversy will continue among physicians, but patients have already been convinced of its utility. Public awareness has put PSA on the tip of every man’s tongue.

We have found PSA to be an invaluable tool in estimating tumor volume and the risk of local and distant spread and, therefore, outcome after therapy. It has gone a long way to level the playing field when comparing different therapeutic modalities. Likewise, PSA doubling time has served to predict the aggressiveness of the disease, both before and after treatment.

Important Lessons Learned

We have become more experienced in our use of PSA as a marker after various modality therapies, realizing that, in each situation, "normal" is actually unique and must be acceptably defined. Consensus on the definition of treatment failure using PSA criteria is still evolving, especially in evaluating patients after radiation. Over the last several years, as the PSA era continues and we acquire ever-longer follow-up after therapy, important lessons have been learned.

Establishing a PSA level that indicates treatment success vs failure has become much more straightforward for prostatectomy as compared to radiation. Because the prostate is totally removed by the former procedure, any detectable PSA-after accounting for female levels and any small amount of glandular tissue that might remain after surgery-is considered abnormal, and any rise in PSA level signals recurrent disease.

Defining Postirradiation Cure

Defining recurrence after radiation therapy is a bit more tricky. Normal prostate tissue with the capacity (albeit reduced) for producing PSA remains in place. Therefore, a measurable amount of PSA nearly always remains, although it seems to vary slightly between individual patients. The question of how low the PSA level must drop before a patient is considered cured has been hotly debated.[2] Data to address this issue were compiled, and a Consensus Conference convened several years ago.[3] From this conference came the first attempt at a standard definition of failure and a multi-institutional pooled analysis, which demonstrated outcome after irradiation using these parameters.[4]

One important piece of information from this study was the fact that PSA nadir was not an "all or none" phenomenon, but instead, a continuous variable that merely set the odds as to the probability of failure. For example, stage T1/2 patients with a PSA nadir £ 0.5 ng/mL had an 83% chance of being disease-free at 5 years after treatment, according to the American Society of Therapeutic Radiology and Oncology (ASTRO) definition. For those with a PSA nadir of 0.6-0.9 ng/mL, the chance was 68%, and for a nadir of 1.0-1.9 ng/mL, 56%. Even patients with a PSA nadir at or above 2.0 ng/mL had a 28% chance of being disease-free at 5 years, thus lending credibility to the ASTRO definition, which describes treatment failure as three consecutive rises in the PSA level instead of relying on an absolute value.

Although this definition has been criticized, one must remember that PSA is merely a biochemical surrogate for what eventually happens clinically over what may be a very extended period of time.[5] As such, it is only useful if meaningful correlates can be drawn.

Calculating Survival

Ultimately, the issue becomes survival, and in this patient population (which includes many who are afflicted with other life-threatening diseases), the question is, more specifically, which competing cause of death is most likely to take precedence. Studies addressing disease recurrence after prostatectomy have shown a median time from PSA elevation to death of 11 years for Gleason 7 tumors, 9.5 years for Gleason 8 lesions, and 7.5 years even for Gleason 9/10 disease.[6]

Taking into account that it is approximately 2.2 years from treatment to PSA recurrence for patients with high Gleason scores and 6 years or more for those with lower scores, the ability of PSA to predict survival may be determined largely by the patient’s age and how serious his comorbidities become within 10 years or more. As both Albertsen’s data on watchful waiting[7] and this article by Partin and colleagues point out, PSA may be most predictive of ultimate survival in younger patients with aggressive disease.


It is imperative, therefore, that our knowledge of PSA and its ability to predict outcome for prostate cancer patients be used not only academically but practically and wisely, especially as the array of treatment options for both primary disease and recurrence expands. We must continue to analyze and review the available data as it matures. With this, the authors have done a stellar job. I look forward to part 2.


1. Jemal J, Thomas A, Murray T, et al: Cancer statistics, 2002. CA Cancer JClin 52:23-47, 2002.

2. Critz FA: A standard definition of disease freedom is needed for prostatecancer: Undetectable prostate specific antigen compared with the AmericanSociety of Therapeutic Radiology and Oncology Consensus Definition. J Urol167:1310-1313, 2002.

3. Consensus statement: Guidelines for PSA following radiation therapy.American Society for Therapeutic Radiology and Oncology Consensus Panel. Int JRadiat Oncol Biol Phys 37:1035-1041, 1997.

4. Shipley WU, Thames HD, Sandler HM, et al: Radiation therapy for clinicallylocalized prostate cancer: A multi-institutional pooled analysis. JAMA281:1598-1604, 1999.

5. Pound CR, Partin AW, Eisenberger MA, et al: Natural history of progressionafter PSA elevation following radical prostatectomy. JAMA 281:1591-1597, 1999.

6. Jhaveri FM, Zippe CD, Klein EA, et al: Biochemical failure does notpredict overall survival after radical prostatectomy for localized prostatecancer: 10-year results. Urology 54:884-890, 1999.

7. Albertsen PC, Hanley JA, Gleason DF, et al: Competing risk analysis of menaged 55 to 74 years at diagnosis managed conservatively for clinically localizedprostate cancer. JAMA 280:975-980, 1998.