Prostate-Specific Antigen: What’s New in 1997

Introduction

Prostate-specific antigen (PSA) is the most important tumor marker available today for the early
detection, staging, and monitoring of men with prostate cancer.[1] This serine protease was first identified in seminal fluid in 1971.[2] Although minimal immunoreactivity against PSA antibodies can be detected in breast cyst fluid, salivary glands, ductal cells of the pancreas, and endometrium, PSA is organ-specific for all clinical purposes. However, PSA is not cancer-specific; it is produced by nonmalignant as well as malignant prostatic epithelium.[3] Therefore, there is a substantial overlap in PSA levels between men with benign prostatic hyperplasia (BPH) and those with prostate cancer.[4]

When used alone, PSA is not sufficiently sensitive or specific to consider it an ideal tool for the early detection[5] or staging of prostate cancer.[6] In order to optimize the use of PSA, the concepts of PSA velocity,[7] PSA density,[8] and age-related PSA values[9] have been established. Most recently, the measurement of different molecular forms of PSA,[10] especially the ratio of unbound, or "free," PSA to total PSA, was introduced. Also, advances in molecular techniques, such as the reverse transcriptase-polymerase chain reaction (RT-PCR), enable the detection of minimal amounts of PSA messenger RNA (mRNA).[11]

Prostate-specific antigen, also known as human kallikrein 3 (hK3), is a member of the kallikrein family. The availability of a monoclonal antibody against human kallikrein 2 (hK2), an other member of this family that is also expressed predominantly in the prostate, may be the first step toward the development of a new marker for prostate cancer.[12]

This review will focus on these new developments and their use in clinical practice.

Total PSA

The PSA assays in clinical use since 1987 detect both the complexed and noncomplexed forms of PSA. Today, PSA assays are available that selectively detect unbound PSA (free PSA) or PSA complexed to alpha-1-antichymo-trypsin (bound PSA).[10] The sum of all measurable complexed (bound) and uncomplexed (free) forms of PSA is referred to as "total PSA."

In 1996 alone, nearly 42,000 men in the United States will die from prostate cancer.[13] A method for the early detection of men with prostate cancer is urgently required. Early studies using total PSA for the detection of prostate cancer suggested that, due to theoverlap in serum PSA levels between men with BPH and prostate cancer, PSA would not be useful in prostate cancer detection.[14] Since then, however, many studies have shown that PSA is a valuable tool for the detection of men with prostate cancer,[15] and recent data indicate that by using PSA screening for the early detection of prostate cancer, men with organ-confined disease can be better identified.[16]

Screening for prostate cancer remains controversial, however. Currently, only 60% of newly detected prostate cancers are organ-confined, and thus, potentially curable.[13] An effective screening method should have high sensitivity and specificity for the discrimination between malignant and benign disease.

The use of PSA levels alone for screening is compromised by the varying amount of PSA produced by benign prostatic tissue. Thus, this screening test has a positive predictive value of only 47%.[17] Several studies have demonstrated that PSA is a useful addition to transrectal ultrasound (TRUS) and digital rectal examination (DRE) and can increase the overall detection rates for prostate cancer among screening populations.[18-20] On the other hand, 35% of men with organ-confined prostate cancer present with PSA levels less than 4 ng/mL, and 38% of men treated with radical prostatectomy present with a palpable lesion.[6]

These data underline the fact that PSA alone is not a perfect marker for prostate cancer screening. However, the combination of PSA, DRE, and TRUS is more effective in detecting prostate cancers.[18] A recent decline in prostate cancer death rates has been noted in the United States—the first time that this has occurred.[13] These statistics suggest that screening with this combination of tests improves survival in men with prostate cancer. Nevertheless, there is marked controversy over whether or not to screen for prostate cancer.[21]

A National Cancer Institute (NCI) randomized trial to resolve this controversy has been initiated but will take several years to be completed. It is unreasonable to restrain from using PSA until a definitive answer is available. The combination of DRE and PSA determination remains the most thorough, cost-effective method for the detection of early prostate cancer. It should be noted, however, that screening is reasonable only if effective treatment options exist for the early detected disease. Today, radical prostatectomy and radiation therapy for men with localized prostate cancer are the established forms of definitive therapy for prostate cancer.

Concepts for Improving PSA Specificity and Sensitivity

In order to improve the specificity and sensitivity of PSA for the detection of men with prostate cancer, three different concepts were introduced: PSA density, PSA velocity, and age-specific PSA reference ranges.

PSA Density

Based on the observation that malignant prostatic tissue leads to a tenfold higher rise in PSA levels compared to an equal volume of benign tissue,[14,22] Benson and associates determined the ratio between serum PSA and prostatic volume assessed by TRUS, a parameter which they termed "PSA density."[23] Seaman et al were the first to compare the value of serum PSA levels and PSA density in detecting prostate cancer, with histologic confirmation of the results.[8] Using a PSA density of 0.15 as a threshold, they demonstrated the increased utility of PSA density for detecting prostate cancer in patients with serum PSA levels between 4.0 and 10.0 ng/mL.

Since then, conflicting results have been published.[24-27] Brawer et al reported that PSA density was not superior to serum PSA alone in distinguishing men with BPH from men with prostate cancer,[24] whereas Wolff et al found that the use of PSA density helped discriminate between men with prostate cancer and men with BPH.[25] Ohori and coworkers described the utility of PSA density in only a small subset of patients with an elevated serum PSA (more than 10 ng/mL).[26] Finally, Catalona et al, presenting the results of a multicenter trial of nearly 5,000 men, determined that the use of PSA density with a cut-off of 0.15 could have missed nearly 50% of the prostate cancers.[27] The differences between these studies can be explained, at least partially, by the fact that the determination of prostatic volume by TRUS is examiner-dependent and often subjective.

In summary, the value of PSA density seems to be restricted to patients with intermediate PSA levels (4.0 to 10.0 ng/mL). In this subgroup, it is helpful in identifying patients who are at a low risk of having prostate cancer.[28]

PSA Velocity

Carter and coworkers demonstrated that the rate of change in PSA over time, or PSA velocity, was more useful than PSA alone in distinguishing between men with and without prostate cancer.[7] In their study of 54 men followed for an average of 17 years in the Baltimore Longitudinal Study of Aging, differences in PSA velocity between BPH and prostate cancer patients were detectable up to 9 years before the diagnosis of prostate cancer was made. Carter et al suggested a cut-off value of 0.75 ng/mL/yr. Oesterling et al[29] and Smith and Catalona[30] confirmed a cut-off point of 0.8 ng/mL/yr for PSA velocity.

In a study aimed at determining the optimal time interval between the serum samples and the optimal number of samples for calculation of PSA velocity, Carter et al found an inverse relationship between PSA velocity and PSA sampling interval.[31] They recommended using three consecutive measurements obtained over 2 years rather than over a short-term period (3 to 6 months).[31]

Kadmon et al demonstrated that, over a 2-year follow-up period, 12.5% of 265 men without prostate cancer had a PSA increase higher than 0.75 ng/mL/yr. Thus, they recommended an observation period of at least 2 years before using PSA velocity to make clinical decisions.[32]

In summary, PSA velocity increases the predictive value for the detection of prostate cancer, as compared with PSA level alone. However, follow-up of at least 2 years is required before therapeutic decisions can be made.

Age-Specific PSA Levels

The normal PSA reference range of 0.0 to 4.0 ng/mL does not account for age-dependent prostatic growth and variations in PSA secretion and production.[33] Oesterling et al assumed that a PSA level of 5.5 ng/mL had a different significance in a 75-year-old man than in a 51-year-old man. Thus, in a prospective study, they evaluated men without evidence of prostate cancer and determined age-specific reference ranges for PSA.[9] Compared to the standard PSA reference range, age-specific PSA was a more sensitive tumor marker in men younger than 60 years and a more specific marker in men older then 60 years.

Partin and coworkers demonstrated that, in younger men, the use of age-specific reference ranges detected more potentially curable prostate tumors with favorable pathology. The tumors missed in older men by age-specific PSA ranges also had a favorable pathology in 95%.[34] According to Partin et al, if urologists want to maintain sensitivity for the detection of nonpalpable cancers in men over age 60 years, they may use the standard PSA reference range.

In contrast, based on an analysis of 6,630 men participating in a PSA early detection study, Catalona and associates concluded that a PSA cut-off of 4.0 ng/mL was superior to all other cut-off values for all age groups.[35] Their study demonstrated that the number of biopsies could be decreased in older men.

Morgan et al showed that the use of age-specific reference ranges for PSA are different in Caucasian and black men. In their study, 41% of the prostate cancers detected in a group of 411 black men would have been missed if the traditional reference ranges had been used.[36] To maintain a 95% sensitivity among black men, the age-specific reference ranges have to be higher than in corresponding groups of white men.[36]

DeAntoni and coworkers determined race-specific reference rates for white, Latino, black, and Asian men.[37] The higher upper limits of race-specific reference ranges for black men, for example, are due to a greater PSA variance. The higher reference range would raise the cut-off level of PSA in black men, however, thus making the test less sensitive.

In summary, the use of age-specific reference ranges seems to be clinically useful at least for patients younger than 60 years. The clinical significance of race-specific PSA reference ranges remains to be determined.