CLINICAL GUIDELINE: PART III: Screening for Prostate Cancer

Screening Test or Diagnostic Test?

The traditional distinction between a screening test and a diagnostic test can seem blurred when testing for prostate cancer. Screening involves testing persons who have no symptoms of the condition in question. The purpose of screening is to detect disease early, when treatment might be more effective or less expensive. However, about one third of men who are older than 50 years of age have lower urinary tract symptoms consistent with benign prostatic hyperplasia [15, 16]. Is testing for prostate cancer among such men screening or diagnosis? Recent studies strongly suggest that men who have lower urinary tract symptoms consistent with benign prostatic hyperplasia are not at greater risk for prostate cancer, aside from the risk conferred by their age [17]. In this guideline, screening for prostate cancer is defined as testing in men who do and do not have lower urinary tract symptoms consistent with benign prostatic hyperplasia.

Screening Tests

Digital Rectal Examination

Physicians have traditionally used digital rectal examination in early detection efforts, despite a lack of evidence from controlled studies showing that the procedure reduces disease-specific mortality or net morbidity rates for prostate or rectal cancer [18, 19]. To do a digital rectal examination, the physician places a gloved finger inside the rectum and palpates the posterior aspect of the prostate gland. This test is not associated with any significant immediate risks in apparently healthy men. It requires little time and has historically been integrated into many physicians' routine periodic health examinations for middle-aged and older men. As such, it does not generally incur any extra financial cost.

In relatively unselected populations, 7% to 15% of men older than 50 years of age may have suspicious results on a digital rectal examination if the criteria for abnormal results are broadened to include induration and marked asymmetry in addition to frank nodularity [20-22]. Interrater reliability for this test seems to be only moderate, even among urologists [23, 24]. The true sensitivity and specificity of digital rectal examination remain unknown because previous studies [which were primarily done with volunteers and were therefore subject to referral bias] have not 1) uniformly done biopsies on men with glands that were normal on palpation or 2) used long-term observation of all screened men as a preferred criterion standard. Data from community-based studies suggest that the positive predictive value of digital rectal examination for prostate cancer is 15% to 30% [20-22, 25] and varies relatively little with age [20]. In general, digital rectal examination has a 1% to 2% detection rate when used alone in men who are older than 50 years of age [20-22, 25]. Abnormal results on digital rectal examination increase the odds of a clinically significant intracapsular prostate tumor (> 0.5 mL) 1.5- to 2-fold (see Table 3 in part I of the background paper) and increase the odds of extracapsular disease 3- to 9-fold. However, results on digital rectal examination that do not indicate cancer do not appear to substantially reduce the post-test odds of clinically significant prostate cancer. In other words, the negative predictive value of digital rectal examination for prostate cancer seems low, largely because of poor sensitivity.

Measurement of Prostate-Specific Antigen

Another way to test for prostate cancer is to measure the level of prostate-specific antigen (PSA) in serum. Levels greater than 4.0 ng/mL are often considered suspicious in the most commonly used assays. Age-specific thresholds for abnormal PSA levels have been proposed by some authorities [26-28], but disagreement about their value remains [29, 30].

Measurement of PSA is not associated with any immediate risks other than those that are associated with simple phlebotomy. Clinically important elevation of PSA levels is not seen if digital rectal examination is done shortly before the blood is drawn [31, 32]. Although PSA is generally specific to prostatic tissue, it is not specific only for prostate cancer. Because many older men develop benign prostatic hyperplasia, which often elevates PSA levels, the specificity of PSA measurement decreases with age. Serum levels of PSA can also be elevated for several weeks after acute prostatitis, transrectal needle biopsy, acute urine retention, and prostate surgery [33, 34]. The average cost of PSA measurement is $30 to $40. Because of the substantial cost related to evaluation and treatment that can follow abnormal results of PSA measurement (and digital rectal examination), the total discounted cost per patient for a combined digital rectal examination and PSA measurement is estimated to range from $161 to $414, depending on the age of the patient.

No published studies of PSA measurement in unselected populations have applied an acceptable reference standard. Moreover, substantial uncertainty remains about the ability of PSA measurement and histologic grade to predict which newly diagnosed prostate tumors will lead to clinically significant illness and premature death. Thus, the true sensitivity and specificity of PSA measurement are unknown.

Gann and colleagues [35] used long-term follow-up as a reference criterion for PSA. However, the proportion of men whose PSA levels were elevated at enrollment who were later found to have advanced cancer but had potentially curable cancer at the time of blood sampling remains unknown.

In studies of community volunteers, an abnormal PSA level (> 4.0 ng/mL) has been seen in about 15% of men who are older than 50 years of age [20, 36-38]; the proportion increases substantially with age [20]. The probability that a man who is older than 50 years of age has prostate cancer if his PSA level is elevated (positive predictive value) is approximately 20% to 30%; however, the likelihood of cancer depends on the degree of elevation in the PSA level [20, 36-38]. For levels between 4 and 10 ng/mL, the positive predictive value is about 20%; this value increases to between 42% and 64% if the PSA level is greater than 10 ng/mL. Positive predictive value appears to vary little with age, which suggests that increased disease prevalence is balanced by decreased test specificity in older men [20]. Use of PSA measurement alone results in a cancer detection rate of about 3%. Elevations in PSA level between 4 and 10 ng/mL increase the odds of clinically significant intracapsular prostate cancer 1.5- to 3-fold and the odds of extracapsular cancer 3- to 5-fold (see Table 4 in part I of the background paper). Although PSA levels that exceed 10 ng/mL may still reflect only benign prostatic hyperplasia, the odds of extracapsular cancer at this PSA level are increased more than 20-fold. Although PSA levels greater than 10 ng/mL may increase the odds of significant intracapsular cancer 3-fold, older data [39] suggest that elevations of this magnitude reduce the likelihood that any cancer present is still contained within the gland. Thus, most cases of prostate cancer in men who have PSA levels greater than 10 ng/mL are extracapsular and much less likely to be curable. Results of PSA measurement have a greater influence than the results of digital rectal examination on the odds of clinically significant prostate cancer, especially for extracapsular tumors (see Tables 3 and 4 in part I of the background paper).

One way to reduce the proportion of false-positive results on PSA measurement is to use PSA measurement in conjunction with digital rectal examination [17]. In one study [17], when the results of both tests were suspicious, the positive predictive value of an elevated PSA level among community volunteers increased from 32% to 49%. However, restricting transrectal needle biopsy to those men who have abnormal results on both tests significantly reduces the test's sensitivity. Today, most physicians engaged in the early detection of prostate cancer use both digital rectal examination and PSA measurement and recommend transrectal needle biopsy if the results of either test are abnormal [2, 40]. Suspicious results on digital rectal examination or a PSA level greater than 4.0 ng/mL increases the odds of an intracapsular (> 0.5 mL) prostate tumor between two- and threefold [20] (see Tables 3 and 4 in part I of the background paper).

Transrectal Ultrasonography

Transrectal ultrasonography should be reserved for further investigation in patients who have abnormal results on digital rectal examination or PSA measurement and to guide biopsies. It should not be used as a primary screening test.

Transrectal Needle Biopsy

Transrectal needle biopsy is not recommended for screening.

Potential Risks and Benefits of Screening

The potential benefit of screening for prostate cancer is that early detection may save lives and may avert future cancer-related illness for some men [40]. If PSA measurement and digital rectal examination are used aggressively, the pathologic stages at which cancer will be treated will apparently be earlier compared with historical control cases [17, 20]. However, evidence of “stage shift” alone is insufficient proof that early detection reduces disease-specific mortality rates. The decision model presented in the background paper suggests that, under a set of assumptions that favors screening (in particular, that men with organ-confined disease would be cured), men 50 to 69 years of age who have one-time (prevalence) screening for prostate cancer would appear to gain, on average, several weeks of life expectancy (see Tables 2 and 3 in part II of the background paper). For men who subsequently receive a diagnosis of cancer, aggressive treatment with radical prostatectomy or radiation therapy potentially adds 3 more years of life for men in their 50s, 1.5 years for men in their 60s, and 0.4 years for men in their 70s. Data to model potential benefits for men who are younger than 50 years of age or to model the results of repeated (serial) screening are currently insufficient.

In exchange for these potential but unproven benefits, prospective candidates for screening face a 15% to 40% chance (depending on age) of requiring a biopsy with the attendant discomfort. Minor, self-limiting complications have been reported in as many as 40% of men who have had biopsies [41].

The risks of radical treatment of prostate cancer are well established. Men who are found to have clinically localized cancer as a result of screening and those who have had radical prostatectomy face greater than a 50% chance of permanent sexual dysfunction [40, 42, 43], a 20% to 30% likelihood of some degree of chronic urinary incontinence [40, 42, 43], and a 0.5% to 1% chance of perioperative death [42, 44] (see Table 1 in part II of the background paper).

Radiotherapy generally poses lower associated risks for these adverse effects but confers risks for other chronic sequelae, including an approximate 10% chance of bowel dysfunction [40, 42, 45].

Using the best available evidence, the analysis of risks and benefits of screening using the decision model suggests that, under a set of assumptions that favor screening and treatment, one-time testing with digital rectal examination and PSA measurement may eventually be proven to have a clinically important net benefit in terms of life expectancy and a reasonable cost in men of average health who are 50 to 69 years of age (see Tables 2 and 3 in part II of the background paper). Early detection efforts, especially for men in this age range, cannot be dismissed out of hand as clearly ineffective or cost-inefficient. However, considerable illness will be engendered among those who are aggressively treated for prostate cancer discovered as a result of testing. Moreover, with assumptions that are less favorable to screening, the estimated net benefit in life expectancy would decrease significantly. Similarly, the ratio of costs to health benefits would escalate dramatically.

Because of the uncertainty about which combination of assumptions is valid, this analysis cannot be used to effectively argue for or against screening in all circumstances. However, the accompanying analysis strongly suggests that screening men who are older than 69 years of age will generally increase life expectancy by only a few days, even with generous assumptions. If these assumptions are less favorable, screening could result in net harm in men older than 70 years of age (see Table 4 in part II of the background paper).

Recommendations

Recommendation 1: Rather than screening all men for prostate cancer as a matter of routine, physicians should describe the potential benefits and known harms of screening, diagnosis, and treatment; listen to the patient's concerns; and then individualize the decision to screen. Factors to consider include the following.

The principal reason for caution is that the balance of benefit and harm from early treatment is unknown. The risks of treatment are known, but the benefits are not known because no controlled studies of the effect of early treatment on death from prostate cancer have been done.

Indirect evidence based on decision models suggests that if screening and treatment prove beneficial, men who are 50 to 69 years of age will enjoy most of the benefit from screening. Men older than 69 years of age will gain little from screening.

It is important to note that because additional information is obtained from digital rectal examination (for example, information on occult blood in the stool, masses, rectal fissures, fistulas, and hemorrhoids), digital rectal examination is valuable for evaluating conditions other than prostate cancer.

In issuing a general recommendation against the routine use of such tests, the College acknowledges that it may be reasonable for a physician to recommend that an individual patient be screened for prostate cancer. The area of greatest controversy is screening for men who are between 50 and 69 years of age. For men in this age group, the physician should be particularly guided by the patient's preference and by the physician's and the patient's interpretation of the risk–benefit equation. Clinicians should be prepared to discuss this issue with their patients, to provide counseling on an individual basis, and to document these discussions. Potential benefits must be balanced against the potential morbidity and mortality related to treatment by radical prostatectomy or radiation therapy.

Men who have first-degree relatives with prostate cancer [46, 47] and black men [48, 49] have a higher lifetime risk for developing prostate cancer. Some authorities have recommended that such men be tested earlier (in their 40s) or more aggressively [6]. However, no direct or indirect evidence from large studies quantifies the yield and predictive value of early detection efforts in such men. Proof that screening on a one-time or serial basis reduces disease-specific mortality rates and that aggressive treatment of localized prostate cancer provides a net benefit awaits the findings of longitudinal controlled trials [9-12]. Black men and men with a family history of prostate cancer should be made aware of their higher lifetime risk. However, available evidence does not suggest that they should be cared for differently from men at average risk.

Recommendation 2: The College strongly recommends that physicians help enroll eligible men in ongoing clinical studies.

Given the uncertainty about the net benefit of early detection and treatment of prostate cancer, the College encourages physicians to help enroll eligible men in ongoing clinical studies when feasible. The National Cancer Institute's Prostate, Lung, Colorectal, and Ovarian Screening Trial is a 10-center study that is designed to measure the net benefit of screening for prostate cancer [9]. Physicians can request more information or direct their patients to call the enrollment center directly at 800-4-CANCER. The Prostate Cancer Intervention Versus Observation Trial (PIVOT) is a recently initiated collaboration between the Veterans Administration Cooperative Studies Program and the National Cancer Institute [10]. Men who are 75 years of age or younger and have clinically localized prostate cancer of all histologic grades are eligible. Participants are randomly assigned to receive radical prostatectomy with additional treatment for evidence of residual or recurrent disease or expectant management with treatment for evidence of symptomatic progression or metastatic disease. Physicians may contact the PIVOT Chairman's office at 612-725-2000, extension 4586, for further information.

Counseling Patients

All men who are considering having digital rectal examination and PSA measurement should understand the potential risks and benefits of screening and participate with their physicians in deciding whether to be tested. Before any testing occurs, patients should be fully informed about the following:

1. Prostate cancer is an important health problem.

2. The benefits of one-time or repeated screening and aggressive treatment of prostate cancer have not yet been proven.

3. Digital rectal examination and PSA measurement can both have false-positive and false-negative results.

4. The probability that further invasive evaluation will be required as a result of testing is relatively high (see Table 1 in part II of the background paper).

5. Aggressive therapy is necessary to realize any benefit from the discovery of a tumor.

6. A small but finite risk for early death and a significant risk for chronic illness, particularly with regard to sexual and urinary function, are associated with these treatments (see Table 1 in part II of the background paper).

7. Early detection may save lives.

8. Early detection and treatment may avert future cancer-related illness.

Routine PSA measurement without a frank discussion of the issues involved is inappropriate. Patients who elect to be screened, either by digital rectal examination or PSA measurement, should provide verbal informed consent. In making this recommendation, the College acknowledges the logistical difficulties of trying to incorporate an informed decision-making approach to screening for prostate cancer into a busy practice [50]. Various strategies, ranging from fact sheets [51] to videotapes [52], have been devised to assist clinicians in providing an unbiased summary of the potential benefits and harms that can result from screening. No published data have indicated the optimal way to facilitate such physician–patient communication. As with much of clinical medicine, the interpretation of the general information that is available in brochures or videotapes must be tailored to the individual patient. This interpretation is probably best done in the context of an ongoing relationship with a primary care provider.