The sensitivity of prostate-specific antigen has never been questioned. It is the test’s lack of specificity (i.e., the number of false-positives that result in invasive evaluations) that has caused most concern. In an attempt to improve specificity, several investigators have tried to modify the interpretation of prostate-specific antigen values. These modifications have been termed prostate-specific antigen derivatives. Numerous prostate-specific antigen derivatives have been developed and tested clinically with the purpose of optimizing the use of prostate-specific antigen as a screening tool. This can only be achieved by increasing the test’s specificity while preserving its sensitivity.
Prostate-specific antigen Density
Prostate-specific antigen density is one such variation and is defined mathematically as total prostate-specific antigen (ng/mL) divided by the volume of the prostate gland (cc). The concept of Prostate-specific antigen density is based on the assumption that the Prostate-specific antigen density calculation would standardize the amount of prostate-specific antigen produced per cubic centimeter of prostate tissue. A volume occupied by cancer will result in a higher serum prostate-specific antigen than a volume occupied by benign tissue. It was postulated that Prostate-specific antigen density would be able to identify which patients had an elevated prostate-specific antigen secondary to benign enlargement, benign prostatic hyperplasia, versus those with elevations secondary to prostate carcinoma.
The concept of Prostate-specific antigen density was introduced at Columbia-Presbyterian Medical Center by the authors in collaboration with Dr. William Cooner of Mobile, Alabama. The objective was to derive a means of decreasing the number of false-positive prostate-specific antigen results and thus the number of unnecessary biopsies.
TABLE. Age- and Race-Specific Reference Ranges for Prostate-Specific Antigen
|
|
Serum prostate-specific antigen(ng/mL) |
| Age |
Caucasian |
Japanese |
African American |
| 40-49 |
0-2.5 |
0-2.0 |
0-2.0 |
| 50-59 |
0-3.5 |
0-3.0 |
0-4.0 |
| 60-69 |
0-4.5 |
0-4.0 |
0-4.5 |
| 70-79 |
0-6.5 |
0-5.0 |
0-5.5 |
An analysis of 773 patients with a normal digital rectal examination and a prostate-specific antigen between 4 to 10 ng per mL who subsequently underwent transrectal ultrasound led to the proposal of a Prostate-specific antigen density cut off value of 0.15 mg per mL. In this preliminary study, only patients with an abnormal transrectal ultrasound (hypoechoity or asymmetry) underwent an initial biopsy. The positive prostate biopsy rate in this group was 6% for patients with a Prostate-specific antigen density less than 0.15 versus 18% for patients with a Prostate-specific antigen density greater than 0.15. Another investigation by Bazinet and colleagues confirmed these findings. In their study, 142 patients with a negative digital rectal examination and a prostate-specific antigen between 4 to 10 ng per mL underwent prostate biopsy. They noted that only 2 patients who had a Prostate-specific antigen density less than 0.15 biopsied positive, while 20 patients had positive biopsies with a Prostate-specific antigen density greater than 0.15.
Despite these encouraging initial data, other studies have failed to confirm improved cancer detection rates using Prostate-specific antigen density compared with total serum prostate-specific antigen. Factors responsible for the conflicting results and conclusions may involve anatomic and technical difficulties in determining prostate volume, lack of uniformity regarding the statistical analysis applied to the different studies, and the fact that the epithelial-stromal ratio differs considerably from patient to patient. Differences in the amount of epithelium versus stroma in an individual’s prostate allow for a wide range of prostate-specific antigen production in prostates of similar volume. Because of these observations, Prostate-specific antigen density as a means of increasing the specificity of serum prostate-specific antigen testing is not universally accepted.
Age-Specific prostate-specific antigen
An age-specific reference range for serum prostate-specific antigen is a variation of total prostate-specific antigen which was designed to increase sensitivity in younger men and to increase specificity in older men. The concept of age-specific prostate-specific antigen follows that as most men get older they develop benign prostatic hyperplasia, with a resultant increase in their prostate size and an increase in their serum prostate-specific antigen. In a community-based study, Oesterling and colleagues enrolled 537 men aged 40 to 79 into a screening protocol that included serum prostate-specific antigen, digital rectal examination, and transrectal ultrasound. Of the 537 men, 471 had all three tests performed without any evidence of prostate cancer. Utilizing this subset of patients, they correlated serum prostate-specific antigen with age and prostate volume. The results indicated that prostate-specific antigen increased by 0.04 ng per mL (3.2%) per year. From these data, using 95th percentile confidence limits, age-specific reference ranges for serum prostate-specific antigen were developed.
Several investigators have tested this hypothesis and examined its clinical usefulness in lowering the normal prostate-specific antigen value in younger men and raising the normal value in older men. It appears that lowering the normal range in younger men is valid and appropriate. In a multi-institutional study, Catalona and colleagues utilized a patient population base consisting of 6630 men undergoing prostate-specific antigen and digital rectal examination screening followed by transrectal ultrasound, with prostate biopsy for those with abnormal results. They reported that a cutoff of 3.5 ng per mL in men 50 to 59 years old resulted in a 15% increase in cancer detection versus a cutoff of 4.0 ng per mL. Partin and colleagues studied a population of clinical stage Tic prostate cancer who had undergone radical prostatectomy. The authors concluded that in men younger than 60 years, a significant number of additional tumors would be detected by using age-specific ranges. Thus, it appears that the prostate-specific antigen cutoff of 4.0 ng per mL may be too high for younger men. However, it must be taken into consideration that the added sensitivity of a lower prostate-specific antigen cutoff would be at the cost of decreased specificity and the resultant added morbidity (more false-positive results leading to more negative biopsies).
The validity of increasing the normal prostate-specific antigen range in older men is less established. Had age-adjusted reference ranges been used in the Catalona study, men aged 60 to 69 would have undergone 15% fewer biopsies. However, 8% of organ-confined tumors would have been missed in this age group. In men older than 70, 44% fewer biopsies would have been performed with 47% of organ-confined cancers missed. Partin and colleagues also made the observation that raising the prostate-specific antigen range for older men would result in decreased sensitivity, but they also raised the question of clinical significance and clinical consequence of the missed tumors.
In summary, investigators support the theory that the use of age-specific reference ranges would improve the sensitivity of prostate-specific antigen in younger men, allowing for the diagnosis of more organ-confined prostate cancer. For now, most experts are unwilling to raise the upper limit of normal in older men.
Race and Age-Specific prostate-specific antigen
The database that led to the development of age-specific ranges for prostate-specific antigen was composed almost entirely of Caucasian men. It has been known for a long time that Asian men generally have smaller prostates and a lower incidence of prostate cancer than Caucasian or African American men. African American men have the highest incidence of prostate cancer in the world. These facts led investigators to evaluate the effect of race on age-specific ranges. Oesterling and colleagues performed a study similar to the one outlined above in an attempt to clarify age-specific reference ranges in Japanese men.
Morgan and colleagues completed a study examining age-specific reference ranges in African American men. This study confirmed that the prostate-specific antigen concentration correlates with age, and that the upper limit of normal serum prostate-specific antigen should also be age dependent in African Americans. Morgan determined that if traditional age-specific reference ranges were used as outlined by Oesterling for Caucasian men, 41% of cancers in their African American cohort would have been missed. The differences appear minor, but they have tremendous clinical significance because it is in African Americans that prostate cancer is more prevalent; it occurs at an earlier age, and it has the highest mortality rate.
Prostate-specific antigen Velocity
Another derivative used to improve prostate-specific antigen screening is known as prostate-specific antigen velocity. Utilizing data and frozen serum from the Baltimore Longitudinal Study of Aging, Carter and colleagues were able to plot the prostate-specific antigen values of 73 men, aged 60 or older, over a 7-year period. They observed that men without prostate symptoms or prostate cancer had little change in their prostate-specific antigen value over time. Patients with benign prostatic hyperplasia had a linear slope of prostate-specific antigen velocity, while patients with prostate cancer had an initial linear component that became exponential. Investigators calculated prostate-specific antigen velocity with an equation utilizing at least three separate points and suggested that a prostate-specific antigen velocity greater than 0.75 ng per mL per year was suspicious for prostate cancer.
The “normal” prostate-specific antigen velocity was examined by several studies including one by Smith and Catalona, which prospectively enrolled 982 men to examine the efficacy and utility of prostate-specific antigen velocity This study calculated that for patients with a prostate-specific antigen less than 4.0 ng per mL, the cutoff prostate-specific antigen velocity predictive of cancer was 0.75 ng per mL per year. However, for patients with prostate-specific antigen greater than 4.0 ng per mL, the cutoff point which predicts cancer was 0.4 ng per mL per year.
The major question surrounding prostate-specific antigen velocity is how many serum measurements are required and how far apart they should be spaced. Carter and colleagues addressed these issues by retrospectively examining serial prostate-specific antigen measurements in 806 men. They focused on the number of prostate-specific antigen determinations, the time interval between prostate-specific antigen measurements, and the effect on the calculated prostate-specific antigen velocity. The authors concluded that three or more values with time intervals greater than 6 months between readings are necessary to determine velocity.
This severely limits the clinical utility of prostate-specific antigen velocity, in that it is of greatest value retrospectively, and of least value when the prostate-specific antigen determinations occur within 1 year of each other. Therefore, at a minimum, 2 years of prostate-specific antigen determinations are required to accurately stratify patients by prostate-specific antigen velocity. With this lengthy time frame, patient anxiety and the possibility of disease progression become issues which push toward an early biopsy. As a result, prostate-specific antigen velocity would appear to be of greatest utility in men who have already undergone a biopsy which proved negative for cancer.
Free to Total prostate-specific antigen Ratio
One of the more recent prostate-specific antigen derivatives is the comparison of unbound (free) prostate-specific antigen to the total amount of prostate-specific antigen in the serum. The prostate-specific antigen can exist as free prostate-specific antigen and as prostate-specific antigen bound to a2-antichymotrypsin (ACT) or a2-macroglobulin (A2M).’ The prostate-specific antigen bound to A2M (prostate-specific antigen-A2M) is anti-genically shielded and not measurable by any prostate-specific antigen assay. The prostate-specific antigen-ACT complex, however, is immunoreactively unique and can be measured in the serum as a separate moiety. As a result, it is possible to compare the amount of free prostate-specific antigen to the total amount of prostate-specific antigen (free + prostate-specific antigen-ACT). Lilja and colleagues documented that the majority of prostate-specific antigen in the serum is complexed to ACT, accounting for approximately 85% of the total serum prostate-specific antigen. They later compared the ratio of free:total prostate-specific antigen in men with benign prostatic hyperplasia to the free:total prostate-specific antigen of men with prostate cancer. They found that the free:total ratio was significantly lower in men with prostate cancer than men with benign prostatic hyperplasia (18 versus 28%, respectively, p < .0001). Importantly, this difference was present for prostate-specific antigen values above and below 10 ng per mL. They concluded that the use of free:total prostate-specific antigen ratio would allow for a differentiation of elevated prostate-specific antigen levels secondary to benign prostatic hyperplasia and prostate cancer without decreasing the sensitivity of prostate-specific antigen.
The explanation of why prostate-specific antigen secreted from benign prostatic hyperplasia is less likely to be bound to ACT (higher free:total ratio) may be found within prostate cancer cells. Bjork and colleagues found that prostate cancer cells produce not only prostate-specific antigen, but also ACT. This coexpression of prostate-specific antigen and ACT may allow for an increased likelihood of a prostate-specific antigen-ACT complex when prostate-specific antigen is expressed from a cancer cell as opposed to a benign cell. However, this is only speculative and the reason or reasons behind the observed free:total differences remain unknown.
Regardless of the explanation for the increased prostate-specific antigen-ACT complex in prostate-specific antigen expressed from prostate cancer cells (lower free:total ratios), free:total prostate-specific antigen ratios improve prostate-specific antigen specificity for patients with serum prostate-specific antigen levels in the 4 to 10 ng per mL range. The use of Hybritech’s free:total prostate-specific antigen has been granted approval from the Food and Drug Administration, and free:total ratios from other companies are expected to be approved in the near future. However, free:total prostate-specific antigen, like the other prostate-specific antigen derivatives, is not without controversy. A universal free:total prostate-specific antigen cutoff has yet to be established, and it appears that like Prostate-specific antigen density and age- and race-specific prostate-specific antigen, the resultant increases in specificity are at the cost of decreases in sensitivity.