The American Urological Association (AUA) has released a policy report on the use of prostate-specific antigen (PSA) in screening for and monitoring of prostate cancer. The document was developed to serve as a resource for primary care physicians and urologists. Published in the February 2000 issue of Oncology, the report is called a best-practice policy because of the lack of adequate data to formulate evidence-based guidelines on the use of PSA. The multi-disciplinary panel that developed the policy statement included a family physician, two internists, a radiation oncologist and four urologists. The recommendations are based on review of the literature and expert opinion.

The report discusses the use of PSA for prostate cancer screening, pretreatment staging and follow-up after treatment of prostate cancer. The following summarizes the AUA report.

Use of PSA for Prostate Cancer Screening

The report states that the combination of PSA testing and digital rectal examination is the best method for early detection of prostate cancer. According to the policy statement, transrectal ultrasonography adds little to the combination of PSA determination and digital rectal examination.

Studies indicate that digital rectal examination may reveal evidence of prostate cancer in some patients with normal PSA levels (less than 4.0 ng per mL). One study revealed that approximately 20 percent of prostate cancers with aggressive features occur in men with normal PSA levels. According to the report, nearly 75 percent of cases of prostate cancer are associated with an abnormal PSA level.

PSA testing has been instrumental in the increased detection of prostate cancer at an early stage. Before the introduction of PSA testing (around 1986), approximately one third of patients thought to have cancer confined to the prostate were found to have positive lymph nodes and approximately two thirds had pathologically advanced disease. In contrast, lymph node involvement is now found in only 5 percent of patients.

The report recognizes that there is no definition for what constitutes clinically significant and insignificant prostate cancer. While tumor grade, with the Gleason score the most common grading system, and tumor volume can be used prognostically, they cannot be determined noninvasively. Gleason scores of 2 to 4 generally indicate lower aggressiveness, scores of 5 or 6 are considered to signify intermediate aggressiveness and Gleason scores of 7 or greater are considered to indicate biologically aggressive tumors. Tumor volume in excess of 0.5 mL is considered by many experts to predict clinical significance.

In a discussion about the sensitivity and specificity of the PSA level, the report states that PSA testing in patients with normal levels has a sensitivity of about 67.5 to 80.0 percent. Thus, 20 to 30 percent of tumors will be missed when PSA is used alone. Sensitivity could be improved by using an age-adjusted PSA level, such as a level of 2.5 ng per mL or less in men in their 40s. Another method for improving sensitivity is monitoring the PSA level over time. While these two methods would increase the detection of prostate cancer, they also would increase the number of men who undergo biopsy.

Data indicate that the specificity of the PSA level is 60 to 70 percent when the PSA is greater than 4.0 ng per mL. Only one prostate biopsy in four reveals prostate cancer. As is the case with sensitivity, age adjustment would improve specificity because serum PSA levels tend to increase with age. The use of higher normal levels in older men would result in fewer biopsies.

Another method of improving specificity would be to use the free-to-total PSA ratio. Patients with prostate cancer tend to have lower free-total ratios than do patients with benign disease. However, the optimal cutoff point for free-total PSA is unknown. Adjusting the PSA value in relation to the size of the prostate would be another way of improving specificity of PSA testing.

The likelihood of prostate cancer in patients with abnormal PSA levels depends on the degree of PSA elevation. PSA levels greater than 4.0 ng per mL in men more than 50 years of age are associated with a 20 to 30 percent likelihood of prostate cancer. Levels greater than 20 ng per mL have been found to be associated with prostate cancer beyond the confines of the gland in 80 percent of patients.

The section on the use of PSA testing for prostate cancer screening concludes with the recommendation that PSA testing and digital rectal examination be offered to any asymptomatic man more than 50 years of age with a life expectancy of 10 years. Testing can reasonably be offered at an earlier age in men at increased risk of prostate cancer, such as black men and those with a first-degree relative with prostate cancer.

Use of PSA for Staging of Prostate Cancer

The report notes that PSA levels correlate with the risk of extracapsular extension, seminal vesical invasion, and regional and distant disease. Patients with PSA levels of less than 10.0 ng per mL are most likely to respond to local therapy. The policy also states that routine use of bone scanning is not required for staging of prostate cancer in asymptomatic men with clinically localized cancer if the PSA level is 20 ng per mL or less. Similarly, computed tomography (CT) and magnetic resonance imaging are not indicated for cancer staging in men with clinically localized prostate cancer and PSA levels of less than 25 ng per mL. Data show that the incidence of lymph node involvement is less than 5 percent in such patients. Even in patients with PSA levels of more than 25 ng per mL, the sensitivity of CT scanning for the detection of positive lymph nodes is only 30 to 35 percent.

With regard to pelvic lymph node dissection in staging, the report states that it may not be required in patients with PSA levels of less than 10.0 ng per mL. It also does not seem to be required when the PSA level is less than 20 ng per mL and the Gleason score is 6 or less.

Use of PSA in Post-treatment Follow-up

The policy statement advocates periodic PSA determinations to detect recurrence of prostate cancer. PSA should be undetectable after radical prostatectomy. According to the report, detectable PSA after prostatectomy is associated with eventual disease recurrence in most patients. After radiation therapy and cryotherapy, the PSA concentration should fall to low levels, but the acceptable PSA level is debatable. While no consensus has been reached as far as what should be the target nadir value, a PSA of less than 0.5 ng per mL (or undetectable levels) is not likely to be associated with disease recurrence within five years of treatment.

The policy statement also points out the recommendations of the American Society for Therapeutic Radiology and Oncology (ASTRO), in which biochemical recurrence is based on three consecutive rises in the serum PSA above nadir. ASTRO recommends that PSA be determined no more often than every three to six months to detect meaningful increases in the PSA level.

Distant disease is more likely if the PSA does not fall to undetectable levels after surgery, rises after radiation therapy or cryotherapy, rises within 12 months of any form of treatment or doubles in less than six months.

According to the report, the nadir PSA and the percentage of PSA decline at three and six months predict progression-free survival in men with metastatic prostate cancer treated by androgen ablation. With metastatic prostate cancer, undetectable PSA levels or PSA levels that decrease by more than 90 percent at three and six months signify an increased likelihood of prolonged progression-free survival. Studies suggest that a 50 percent or more decrease in the PSA level at eight weeks after the initiation of secondary therapy is associated with improved survival.

The report notes that the PSA level that should trigger a bone scan following initial treatment of localized disease is uncertain. Rarely is early biochemical failure associated with a positive bone scan. Studies have shown that the probability of a positive bone scan is less than 5 percent until the PSA level reaches between 40.0 and 45.0 ng per mL. Gleason score, pathologic stage, preoperative PSA and time to recurrence have not been found to be helpful in predicting bone scan results.

Screening for prostate cancer is controversial and neither the U.S. Preventive Services Task Force nor the American Academy of Family Physicians has recommended screening because of the known risks and uncertain benefits.