What's Wrong with the PSA Test?


Prostate cancer is much akin to a turtle and a rabbit in an open box.  The turtle, the indolent, nonaggressive cancer, wanders around the box, while the rabbit, the potentially aggressiveness cancer, hops around and might at any time jump out of the box.  Thus, while we can diagnose prostate cancer, we cannot distinguish which cancer is the turtle and which is rabbit, and which needs treatment.  More men die with prostate cancer than die because of prostate cancer.

PSA is a normal component of the prostate.  It is not specific for cancer.  Rather it is present in the normal, benign and cancerous prostate.  The PSA test is therefore not the best tool we have for early diagnosis as some of the “self-anointed” experts continue to preach.  The ability of the PSA test to identify men with prostate cancer is slightly better than that of flipping a coin.  The fact that prostate cancer is an age-related disease, the PSA test merely precipitates a biopsy, wherein related to the age of the biopsied individual, and “… how hard it is looked for,” i.e., the cancer (Welch. Journal of the National Cancer Institute, 97, 1132, 2005), he may or may not have cancer, in which case, if he has cancer, we cannot determine if it is a turtle or a rabbit!

The manufacturers of the PSA test sold the urological community on a test that cannot detect what it purports to detect, i.e., prostate cancer, and the media have conveyed to the average man that current treatments work.  If the later is true, why do 25-35% of men with localized prostate cancer have a recurrence within 5 years?  We have bought into the cultural myth that doing more is doing better!  In the case of prostate cancer, doing less may be better, although men with a family history of prostate cancer or those with symptoms should be more aggressively monitored by their physicians.

By the way, I discovered PSA in 1970, and while in many cases, PSA is an indicator (“harbinger”) of recurrence of disease following treatment, it is not cancer-specific and cannot, in the manner in which the PSA test is currently used, be a screen for prostate cancer.

Richard J. Ablin, Ph.D.
President, Robert Benjamin Ablin Foundation for Cancer Research


Prostate-Specific Antigen: Promises and Pitfalls

By Richard J. Ablin, Ph.D.


Heightened awareness, brought forth by the introduction and aggressive marketing of initially indiscriminate screening and media attention focused at high profile celebrity cases, has dramatically altered the epidemiology of prostate cancer (PCa) in the United States (US) within the last decade. Consequently, many view early detection and treatment as rational, ethical, economical, effective and necessary, even in the absence of definitive evidence of their accuracy and efficacy.

The principal, if not sole contributors to the rise of PCa to the most commonly diagnosed type of cancer (excluding skin cancer) among men have been the identification of prostate-specific antigen (PSA) and the PSA antigen test.


Historically, recognition in the late 1970's of the limitations of immunoassays of prostatic acid phosphatase (PAP) to provide a means for the early diagnosis of PCa, redirected the collective attention of several investigators to a prostate tissue-specific antigen previously demonstrated to be distinct from PAP. Identified in the normal and pathologic human prostate and its secretions by our group (which included the late renowned microbiologist, Ernest Witebsky), this antigen was the forerunner of the subsequently purified and characterized antigen named prostate antigen, and ultimately PSA.

A member of the kallikrein gene family, PSA is a serine protease with a molecular weight of approximately 33 kDa. It is normally produced by highly specialized secretory epithelial cells of the prostate and periurethral glands and secreted into the seminal plasma, wherein one of its functions is to liquefy the seminal coagulum (semen). In the course of a variety of pathological processes in which the normal tissue barriers between the prostate epithelial cells and capillaries are breached, PSA may enter the circulation.

The use of antibodies to PSA permitted the detection, and subsequent quantification, of PSA in the serum of patients with PCa. Based on the clinical usefulness of PSA in monitoring PCa, PSA testing was approved by the US Food and Drug Administration (FDA) in 1986, to monitor disease status in patients with PCa and in 1994, to aid in PCa detection. However, following 1986, the PSA test was widely used to screen patients for PCa. Screening resulted in the diagnosis of a substantial increase in the incidence of PCa, particularly in the number of early (presumed organ-confined) tumours, and to some, their questionable treatment. For example, between 1988 and 1992, prior to FDA approval of the PSA test for detection, the incidence of radical prostatectomy (the presumed 'gold standard" for the treatment of organ-confined tumours) rose in the geographic areas of the Surveillance, Epidemiology and End Results (SEER) Program of the National Cancer Institute from 17 cases per 100,000 to 55 per 100,000!

In the course of the foregoing, it became obvious that several confounding factors, such as elevations in the serum PSA of patients with benign prostatic hypertrophy (BPH) and other irregularities of the prostate, resulting in false-positive tests in up to 80% in some studies, PSA was not disease-specific, which placed significant limitations on its usefulness for screening. Consequently, various PSA-related concepts evolved in an effort to enhance the performance of PSA in discriminating patients with early PCa from those with BPH and other prostatic irregularities. These have generally, with exception, perhaps of age-specific PSA reference ranges, proved inadequate.

The generally accepted normal range of serum PSA is 0-4 ng/ml. However, as PSA is synthesized in epithelial cells, which increase in volume with age, some investigators have proposed the use of age-specific PSA reference ranges, where normal values are from 0 to: 2.5 for ages 40-49; 3.5 for 50-59; 4.5 for 60-69 and 6.5 for 70-79. The use of age-specific PSA ranges, when placed in a diagnostic algorithm with digital rectal examination, has been suggested to increase the sensitivity for detecting PCa at an earlier, potentially curable stage in young men, and to increase the specificity for detecting more clinically significant PCa in older men.

In addition to the use of several PSA-related concepts to enhance the performance of PSA testing, attention has been given to the different molecular forms of PSA in serum. Herein, lies one of the, if not major, pitfalls for current assays.


PSA exists in three major forms in serum. These are free PSA (f-PSA) and PSA complexed to the serine protease inhibitors alpha1-antichymotrypsin (PSA-ACT) and alpha2-macroglobulin (PSA-AMG). Complex formation between PSA and AMG results in encapsulation of PSA and complete loss of PSA epitopes. Therefore, current commercially available immunoassays detect only f-PSA and PSA-ACT. Consequently, what is currently referred to as total PSA (t-PSA) is comprised of f-PSA and PSA-ACT.

The avidity of binding studies of PSA to ACT and AMG following the addition of PSA to blood plasma shows preferential complex formation between PSA and AMG. The formation of a stable complex between PSA and AMG results in a loss of PSA immunoreactivity. This results in a loss of PSA available for quantification by current commercial assays and incorrect calculation of the level of PSA available, whether free or in complex with ACT.

In order to have an accurate determination of t-PSA, the fraction of PSA-AMG must be included. The absence of the inclusion of PSA-AMG in what is presently considered t-PSA may be a contributing factor to discrepancies between a patient's level of PSA and his disease status. For example, 25-33% of PCa patients detected by digital rectal exam and up to 40% of patients with organ-confined PCa have PSA in the normal range. Furthermore, as PSA complexes with AMG, any concomitant underlying pathology contributing to dyscrasias in AMG could conceivably effect the amount of PSA available for quantification by current assays.

With current assays for PSA, the actual amount of t-PSA is underestimated. And, in instances where the ratio of PSA-ACT:t-PSA is utilized, e.g., to discriminate between BPH and PCa, some patients may actually have lower ratios than had the t-PSA included quantification of PSA-AMG. Similarly, attempts to use the percentage of f-PSA, calculated as the ratio of f-PSA:t-PSA multiplied by 100, to provide improved discrimination between BPH and PCa, will be flawed, in the absence of the inclusion of PSA-AMG in what constitutes t-PSA.

Based on the foregoing, it appears counterintuitive that the recently introduced ACT-complexed PSA (cPSA) assay (which also detects PSA complexed with alpha1-protease inhibitor, a minor component in serum), for which its proponents state that it provides the same specificity as the ratio of f-PSA:t-PSA, which requires two separate assays to perform, but in one assay, will be any better, save for the profits generated for the manufacturers and their clinical consultants!


As we approach the next decade, their has been an unprecedented stage migration in PCa with a 50% decline in the diagnosis of metastatic (Stage D) disease. And, there is current speculation that PSA screening, irrespective of its limitations, may nonetheless be responsible for the recent slight decline in PCa mortality, even though the ability of the PSA test to reduce PCa has not been established by ongoing randomized controlled trials. On the other end of the spectrum, the effects of the "boom" in the early diagnosis and treatment of a large number, of generally younger men, for presumably localized PCa, has witnessed a recurrence of their disease.

Given the foregoing, we must understand the limitations of PSA. Therefrom, the use of PSA as a marker of prostate pathology warranting further attention and for monitoring treatment, including use of the new ultrasensitive assays, which may permit detection of biochemical relapse prior to clinical detection and possible adjuvant therapy, appear to be well founded.

Apart from its role as a "tumour" marker, a term inappropriately used by many given its absence of tumour-specificity, perhaps too much has been expected from PSA. Therefore, attention should be directed toward further studies of its physiological function(s) and its role in the pathophysiology of the prostate and therefrom, possibly new therapeutic approaches. By way of example, studies of the proteolytic activity of PSA, wherein serine proteases play important roles in the sequential events of tumour progression and metastases may be most revealing. In this regard, inhibition of PSA by zinc is particularly interesting in view of the inverse relationship between zinc concentration and prostate pathology, whereby the reduction of zinc to about one-third of the normal level in the malignant prostate may contribute to increased proteolysis by PSA of its various physiological substrates, e.g., insulin-like growth factor binding proteins, and consequences thereof.

By all means, and not considered by many of the "nouveau" authors and editorialists on PSA, we must nonetheless, treat the patient, not their PSA.




Reference: The Loop (Newsletter of the National Registry of Microbiologists), Volume 25, Number 2, Fall 1999.