Feature Story

Getting the most out of prostate workups

Any discussion about detecting and treating prostate cancer starts and ends with prostate-specific antigen. In its most highly publicized solo role, as a screening tool, PSA’s performance gets rave reviews.

"PSA screening is clearly effective in detecting prostate cancer early," says Peter Scardino, MD, chair of the Department of Urology and head of the prostate cancer program at Memorial Sloan-Kettering Cancer Center, New York. But PSA doesn’t exit the scene after prostate cancer has been detected. It also acts as part of an ensemble in a less dramatic but equally crucial activity—estimating prognosis and risk of recurrence. And PSA returns to the stage after curative surgery as the optimal technique for monitoring for recurrence.

"Following radical prostatectomy," Dr. Scardino says, "PSA is by far the best indicator of recurrence. It gives six to eight years of warning ahead of conventional studies like physical examination or bone scan."

While serum PSA’s varied repertoire earns it the starring role in prostate cancer management, its less famous co-star—surgical pathology—is equally indispensable. "We actually make the definitive diagnosis of cancer," says Richard Muller, MD, attending pathologist at Hartford Pathology, Hartford (Conn.) Hospital. Determining whether prostatic carcinoma is present is not simple, says Jonathan Epstein, MD, professor of pathology, urology, and oncology at Johns Hopkins Medical Institutions, Baltimore. Although Dr. Epstein consults on 40 to 50 specimens from around the country each day, he finds that "prostate cancer is probably one of the most difficult cancers to diagnose." One reason for this difficulty is that the pathologist gets a limited amount of tissue. To try to compensate for this, urologists tend to take more needle cores. Ultimately, however, Dr. Epstein says, the key to avoiding underdiagnosis is "to be very vigilant, looking for literally just a few malignant glands."

After diagnosis, the surgical pathologist’s analysis of the biopsy and surgical specimens is an essential component of estimating prognosis and risk of recurrence. "In both specimens, we make an estimate of the malignant potential or biological aggressiveness of the tumor," Dr. Muller says. Lawrence True, MD, associate professor of pathology and director of male genitourinary pathology at the University of Washington, Seattle, expands on this theme. "Once a pathologist makes a diagnosis of cancer, it is important to provide the clinician with information to put the patient into a management category," he says. "Either the cancer should be handled with the prospect of cure or the patient is advised that the likelihood of progression is less and he might not need radical treatment with its potential complications."

When determining treatment options, clinicians rely on the Gleason score from the biopsy specimen, serum PSA level, and clinical grade. "Typically these days," says Dr. True, "particularly if radical surgery is indicated, the clinician will use the Partin nomograms to present information about the relative likelihood of the patient’s cancer being more extensive." Named after urologist Alan Partin, MD, of Johns Hopkins, the Partin nomograms are tables in which PSA, clinical stage, and Gleason grade in the biopsy specimen are integrated to predict pathologic stage. In separate nomograms, developed by Michael Kattan, MD, pretreatment PSA, Gleason score, and clinical stage are used to predict the probability of biochemical recurrence preoperatively. Once the surgical specimen is analyzed, the pathologic findings are used to predict recurrence postoperatively.

Screening for PSA in asymptomatic men may be a successful way
to detect prostate cancer early. Ninety percent of patients diagnosed based on an abnormal serum PSA test have localized disease, says Dr. Scardino, and 80 to 85 percent have pathologically confined cancer. "Even though randomized prospective trials have not yet been done to show that PSA screening reduces overall or prostate-cancer-specific mortality," he says, "because it is capable of detecting prostate cancer earlier and because results of treating those cancers with surgery or radiation are so good, we think patients should know about the possible advantage."

The National Cancer Centers Network recommends that, beginning at age 50, men with a life expectancy of 10 or more years have an annual digital rectal examination and serum PSA test. If either is abnormal, prostate biopsy should be done. This recommendation is supported by the American Urological Association and the American Cancer Society. Men at high risk—African-Americans and men with a strong family history of prostate cancer—should have DRE/PSA screens beginning at age 45.

Dr. Epstein says some Johns Hopkins clinicians follow a more liberal policy. "Perhaps because we get selected cases, we not infrequently see prostate cancer in men in their 40s," he says. "So some of our urologists now offer PSA once at age 40, then again at age 45. If both tests are low, they test again at age 50. If PSA is under 2 ng/mL at age 50, our studies show you can wait two years to do another screen."

Despite PSA’s widespread acceptance as a screening tool, controversy remains regarding the extent of its benefit. "It is still too early to tell asymptomatic men that a routine PSA test will lower the death rate for prostate cancer," says Gerald Chodak, MD, clinical professor of surgery at the University of Chicago and director of the Midwest Prostate and Urology Health Center at Louis A. Weiss Memorial Hospital, Chicago. Dr. Scardino agrees. "That means that you can’t fault a physician who sees a 55-year-old man and doesn’t recommend DRE and PSA testing," he says. "You wouldn’t consider that medical negligence because those tests are not of clearly proven medical benefit. We recommend that men be screened, but screening is not yet ready for prime time as a public health measure." To handle this ambiguity, Dr. Chodak has introduced consent forms into his practice. (See "PSA screens and consent.")

In a recent update, the U.S. Preventive Services Task Force reiterated that evidence is lacking as to whether routine screening for prostate cancer improves health outcomes. At the center of this dilemma is the question of whether all prostate cancers detected following an elevated screening PSA need to be treated. "The majority of cancers detected by PSA appear to be significant cancers," Dr. Epstein explains. "But with screening, we are detecting an increasing minority of small cancers that, when the prostate is taken out, may not have needed to be removed." In his practice, Dr. Epstein finds that about 25 percent of prostate tumors are 0.5 cc or less in volume. He says it is "borderline" whether these tumors could have been left behind without clinically affecting the patient. Men with such small tumors theoretically are candidates for watchful waiting.

Dr. Muller refers to autopsy studies to make the same point. "If you take men dying at around age 65, say of a car wreck, about half on autopsy have a small prostate cancer," he says. "Most of these men did not have clinical signs of prostate cancer. Yet if you biopsied this same group, 50 percent would end up with radical prostatectomy." Dr. Muller also sees postsurgical specimens of less than 0.5 cc with low Gleason grade (<6) confined to the prostate. "Such a low-grade tumor has a long doubling time," he says. "In a patient with a life expectancy of less than 20 years, that tumor may not have become manifest in the patient’s lifetime."

Unfortunately, there is no certain way before surgery to distinguish possibly indolent tumors from aggressive ones. In Dr. Muller’s practice, the most common Gleason scores on biopsy are 3+3, 3+4, 4+3, or 4+4, with the first two scores accounting for about 80 percent of patients. "Gleason grade itself doesn’t allow prediction of biological potential for most patients," he says.

"Our challenge," says Dr. True, "is to stratify men who are found to have prostate cancer into those whose cancer should be dealt with and those whose cancers—based still on limited outcomes data—don’t need to be treated. We know that many men in the population have cancers that won’t progress, but we are far from confident in identifying those individuals."

"We have tried to develop algorithms to identify these insignificant cancers," Dr. Epstein says. He and his colleagues have identified a cohort of men whom they are following with repeat biopsies. "We are more comfortable watching a tumor in an older man," he adds. In a younger man, a tumor has a longer time to grow. On the other hand, surgical sequelae, chiefly incontinence and impotence, have greater impact on younger men.

These uncertainties led to the U.S. Preventive Services Task Force’s Dec. 2, 2002 update reiterating that "there is insufficient scientific evidence to promote routine screening [for prostate cancer] for all men and inconclusive evidence that early detection improves health outcomes" (www.ahcpr.gov/clinic/uspstf/uspsprca.htm). The scientific evidence included in the update makes a number of points, such as: "...up to 25 percent of apparently PSA-detected tumors and more than 25 percent of apparently DRE-detected tumors were likely in fact to have been detected by serendipity, that is, an incidental finding from a blind biopsy."

Two large randomized trials in the United States—the National Cancer Institute-sponsored Prostate, Colorectal, Lung and Ovarian Screening Trial and the Prostatectomy Intervention Versus Observation Trial are testing whether screening followed by radical treatment reduces death. Also ongoing is the European Randomized Study of Screening for Prostate Cancer. None of these trials will produce outcomes data for several years.

Even if the trial results are unequivocal, they may not affect clinical practice. "There is a certain psychological element that can obviate all our clinical studies," Dr. Muller says. "The clinician is up against it in advising a patient with low-grade tumor on biopsy, slightly elevated PSA, and our assessment that the tumor is not very big. Often those patients don’t want to hear that they have only a 10 percent chance of having a biologically aggressive tumor. They say, ’I am not going to wait for this to kill me.’ They want definitive treatment."

"Until we have a very accurate way of predicting which tumors are biologically aggressive," says Dr. Muller, "most men will be treated."

Attempts to improve the specificity of PSA screening have focused on following up serum total PSA results in the equivocal range—4 to 10 ng/mL, where the probability of cancer on biopsy is 25 percent or less—with measurement of percent free PSA (%fPSA), the form not complexed to proteins. Performing biopsies on only those samples with <25% fPSA avoids about 20 percent of biopsies in this group while missing five percent of cancers.

"There was great enthusiasm a couple of years ago that the free-to-bound ratio would further stratify patients into likelihood of cancer being present and being more extensive," Dr. True says. "My impression is that that test is not routinely used now because it is not all that powerful." Dr. Scardino finds %fPSA useful primarily for the patient who has borderline PSA, such as 4.1; or who has a very large prostate on DRE, where he suspects PSA may be low for the size of the prostate; or who has a negative biopsy, to help decide whether to perform repeat biopsy.

Further compromising PSA’s accuracy, values below the accepted cutoff may not be safe. Dr. True has seen cases with total PSA levels under 4 ng/mL in which cancer is fairly extensive and may even have high-grade components and higher stage. "These cases are rare," he says, but they exist.

Dr. Partin recently studied the use of %fPSA testing on total PSA results between 2 and 4 ng/mL (J Urol. 2002; 168: 504-508). In this setting, three to five biopsies were needed to find one cancer, compared with three biopsies needed to detect one cancer using %fPSA on total PSA in the 4 to 10 ng/mL range. Of 41 cancers detected when total PSA was 2 to 4 ng/mL, six had a Gleason score of five and 28 had a Gleason score of six.

Data have also suggested that the converse of %fPSA, which is complexed PSA (cPSA), can replace total PSA and %fPSA in the diagnostic setting. Dr. Scardino acknowledges that claims for cPSA are "scientifically correct." However, he says, "I think its appeal is going to be limited." First, physicians are already widely informed about total PSA and would have to be re-educated about the meaning of a given level and cutoffs. Second, "The value of total PSA extends well beyond diagnosis to prognosis and following patients after therapy," he says. "To shift to complexed PSA just for one setting, we would have to redo all staging and prognostic nomograms."

Before prostate biopsy, a rise in total PSA must be confirmed.

"PSA can vary from week to week and month to month," Dr. Scardino notes. If the rise is confirmed, ultrasound-guided needle biopsy of the prostate is performed. "We recommend taking at least six and preferably 10 to 12 cores," he says. "Cores should be labeled as to where they came from to help plan treatment or if results are equivocal, repeat biopsies can be done in the same area."

Immediately the pathologist’s first hurdle arises—the small amount of material available. Each core is about one-half inch by a few millimeters and represents about 1/1,000 of the prostate gland. To compensate, urologists are expanding the amount of sampling. Dr. True cites a urologist colleague’s opinion that 12 biopsy cores are now standard.

Urologists at Johns Hopkins also think that six biopsy cores are inadequate, says Dr. Epstein, and take 12 or even 18. For a man who is considering watchful waiting, for whom it is critical not to miss an aggressive tumor, clinicians may take up to 30 cores under outpatient anesthesia. Still, Dr. Epstein says, "even with more extensive sampling, the problem of limited tissue persists."

Making diagnosis even more difficult, says Dr. Epstein, is that prostate cancers often do not form a solid mass like breast or colon cancers, "but grow by insinuating themselves among benign tissue. So a needle biopsy often shows only a few malignant glands among an overwhelming number of benign glands." Prostate cancers also can have a deceptively benign appearance. And there are numerous mimics of high- and low-grade prostate cancer, such as adenosis and atrophy.

The key to recognizing prostate cancer is to look for a constellation of features, says Dr. Epstein. Architectural factors arguing in favor of cancer include small, crowded glands between larger benign glands. Positive nuclear features are prominent nucleoli, nuclear enlargement, hyperchromasia, and mitotic figures. Dark cytoplasm is also characteristic of malignancy, as are intraluminal crystalloids and blue and dense eosinophilic intraglandular or intraluminal secretions.

Three features are pathognomonic of prostate cancer—perineural invasion, mucinous fibroplasia, and glomerulations (much like a kidney glomerulus)—but they are infrequent.

Findings that should make the pathologist hesitate to call cancer are atrophic features, inflammation, and adenosis, all of which are benign mimickers of prostate cancer. High-grade prostatic intraepithelial neoplasia may be a precursor lesion of prostate cancer or a mimic of invasive cancer, "but you would not call it invasive cancer," Dr. Epstein says. "As a precursor lesion, its natural history is unknown, so you wouldn’t want
to subject the patient to surgery or radiation if only PIN is present."

Diagnosis is aided by ancillary immunochemical tests that label basal cells, which are present in benign glands but absent in cancer. Two basal cell markers are 34betaE12 (high-molecular-weight cytokeratin) and a relatively new marker, p63. "We use both in a given case," Dr. Epstein says. "One may work better." Although it may sound straightforward to say that a biopsy containing tissue in which both stains are absent indicates the presence of cancer, Dr. Epstein warns, "Many benign glands include some mimickers of cancer where one does not always see these markers. So the absence of basal cells is not always diagnostic."

An even newer marker called AMACR (alpha methylacyl-CoA racemase), one of the first developed using molecular techniques, is the first marker present in cancerous but not benign glands. Again, however, in practice, interpretation is not quite so straightforward. Some noncancerous lesions are positive for AMACR, most typically high-grade PIN but sometimes even adenosis or atrophy stains.

"You have to use all immunostains in light of what the lesion looks like under routine light microscopy with H&E," Dr. Epstein says. Using histological features plus existing markers will resolve the "vast majority" of prostate biopsies, all but about five percent, he says. If results are mixed, Dr. Epstein reports it as "small focus of atypical glands suspicious of cancer but not diagnostic." Typically, he adds, unless the patient is very old, repeat biopsy is recommended since approximately 50 percent of these men will be found to have cancer on repeat biopsy.

After diagnosis, the next step is to estimate the probability of progression based on total PSA plus Gleason score, clinical stage, and amount of cancer tissue present. Extent of cancer is expressed as linear length of tissue or an estimate of percent of malignant tissue present, such as the fraction of cores involved or how many millimeters of cancer are seen in each core.

No markers currently help with prognosis. "Even though we have identified candidates, right now we have no markers that we can say with confidence stratify patients," Dr. True says. "DNA ploidy and proliferation and cell cycle markers, such as ki67 and p27, have been proposed, but none has been validated prospectively in multi-institutional studies."

Dr. True is part of a research group, the Pacific Northwest Prostate Cancer SPORE (specialized program of research excellence), that received NIH funding to investigate markers for prostate cancer likely to metastasize to bones and to grow through androgen blockade, thereby exhibiting hormonal independence. The group will search for new gene products using the same molecular methods used to discover AMACR. "Our goal is to identify new immunohistochemical tools for practicing pathologists," Dr. True says.

Dr. Scardino estimates that at Memorial Sloan-Kettering Cancer
Center about 600 men per year are treated with surgery, about 500 with external beam radiation therapy, and about 100 with brachytherapy seed implants. Radical prostatectomy is effective for prostate cancer localized to the prostate or immediate periprostatic tissue, stages T1-T3a, while external beam radiation therapy can be used in locally extensive cancers, stages T3b or T4. "In some patients you could flip a coin," Dr. Scardino says. "Others are better suited for surgery or radiation."

If a man who is receiving anticoagulation for atrial fibrillation has radiation therapy, he may suffer radiation proctitis and rectal bleeding that would be difficult to control. If he has surgery, it would be safe to stop Coumadin for a short time and resume it post-surgery. On the other hand, Dr. Scardino would be reluctant to perform surgery on a man with a history of pulmonary embolism, who is more appropriately treated with external beam radiation therapy. Brachytherapy is appropriate for a limited group of patients—men with favorable low-risk cancers who do not have a large prostate or a high level of bladder outlet obstruction.

Dr. Scardino believes most prostate cancers suitable for surgery are candidates for a nerve-sparing operation, with the neurovascular bundle needing to be resected in roughly 10 to 30 percent of patients. Determining where the tumor extends requires DRE and biopsy results, sometimes supplemented with endorectal MRI. Sparing one or two nerves raises the probability of recovering erectile function from virtually zero to 30-50 percent and 70-90 percent respectively. "These are extremely small, delicate nerves," Dr. Scardino says, "and it is very easy to injure them. So while they are often preserved, they can be damaged, which compromises recovery of function." Success of nerve-sparing surgery thus depends on the surgeon’s skill. Dr. Scardino and his colleagues published data last year showing substantial variations in outcomes from radical prostatectomy (N Engl J Med. 2002; 346: 1138-1144). "Generally the busier surgeons get better outcomes," Dr. Scardino says. "Still, even among the busier surgeons there were disturbing variations. Radical prostatectomy is an unforgiving operation."

Hormonal therapy is important in treating metastatic prostate cancer. "Hormonal therapy is very effective in placing prostate cancer in remission and in reducing its size," Dr. Scardino says. "But it is not curative. Given enough time, prostate cancer will always learn how to grow in the absence of androgens." No data demonstrate a clear advantage to giving hormonal therapy early after diagnosis. Because of its adverse effects—reduced libido, osteoporosis if given long-term, and decreased muscle mass—Dr. Scardino favors saving hormonal therapy until later, but certainly before the disease becomes symptomatic. "That way we gain as much length of life but avoid side effects," he says.

Hormonal therapy is also used to treat localized cancer in conjunction with radiation. Administered before radiation, it shrinks tumor and improves long-term disease-free survival in men with large high-grade cancers or high PSA. A few studies also suggest that men with localized but aggressive tumors benefit further when given hormonal therapy for two to three years after radiation therapy.

As use of high-dose 3-D conformal radiation therapy increases, however, use of adjunctive hormonal therapy may decrease. "At high radiation doses, such as 81 Gray, beneficial effects of hormonal therapy before radiation are doubtful," Dr. Scardino says. Hormonal therapy confers no apparent advantage when used before surgery, according to the results of five large trials.

With the resected radical prostatectomy specimen, the pathologist can make a much better determination of prognosis. One criterion is degree of extension: Is the tumor confined to the prostate? Has it invaded the seminal vesicles or pelvic lymph nodes? Did the surgeon remove the entire lesion or is the tumor at the margin? A more accurate determination of Gleason grade is also possible. "After surgery, we have the entire specimen to look at," says Dr. Epstein, "whereas on needle biopsy, we are confronted with sampling error and might undercall the Gleason score."

Dr. Muller points out that no posttreatment specimen is available after radiation therapy. "That makes the estimate of Gleason grade in biopsy tissue of paramount importance," he says. Gleason grade by itself without pathologic stage is an indicator of biological aggressiveness. "But," Dr. Muller says, "it is much more satisfying to me to get the radical prostatectomy specimen and determine pathologic stage." Recent work has shown that the combination of pathologic stage and Gleason score on the prostatectomy specimen predicts a patient’s probability of biochemical failure (Cancer. 2001;91: 1414-1422).

Dr. True notes the importance of being able to recognize prostate cancer after localized or systemic therapy. "Many times the clinician will go back to biopsy the prostate to evaluate success of treatment," he says. Histologic changes caused by therapy can make this a challenging task. With hormonal blockade or brachytherapy, cancer cells are typically shrunken and inconspicuous. "After brachytherapy especially, benign glands often have a degree of atypia that we are not used to seeing in benign glands," Dr. True says. To avoid mistaking therapy-induced changes for cancer, the pathologist must become familiar with those changes. Immunostains for basal cell markers are also helpful in this context.

Assuming that post-therapy PSA has declined below detectable
limits, any subsequent rise in PSA is interpreted as a sign of recurrence. Radical prostatectomy removes the entire prostate gland, so the only source of PSA would be recurrent tumor, Dr. Muller notes. With radiation therapy, however, some prostate tissue remains, so it is a little less clear-cut that every PSA rise after radiation therapy is a sign of recurrent tumor and a biopsy may be performed.

Dr. True adds one more consideration. "We are becoming more aware that there is a spike of PSA about 18 months after implantation of brachytherapy seeds," he says. "We have now had anecdotal experience in up to a dozen patients that if nothing further is done, PSA will drop to undetectable levels over the next year."

For recurrent cancer after radiation therapy, hormonal therapy is the standard treatment, Dr. Scardino says. For recurrence after radical surgery, treatment depends to a great extent on when the PSA rises and how rapidly this occurs. If PSA levels start to climb within two years of surgery and increase rapidly, that suggests metastasis, for which hormonal therapy is indicated. PSA levels that rise later and go up more slowly suggest local regrowth, and it is reasonable to use radiation to kill that tumor. A look back at the pathology can also be useful. For a more aggressive cancer—high-grade, poorly differentiated, growing into seminal vesicles or lymph nodes—a recurrence is more likely to be metastatic. What is not usually helpful is radiology: When PSA is first elevated, the tumor is so small that it is impossible to detect with a radiologic study. "If you do give radiation for recurrence after surgery," Dr. Scardino says, "we know that you must start before PSA gets to 2 ng/mL. After that, virtually no patients can be cured. And we prefer to start radiation before PSA reaches 0.5 ng/mL."

To better detect malignant prostate lesions in the future, Dr. Muller looks to the detection of oncogenes. "My hope is that we could do an analysis on the biopsy specimen of genetic lesions in each individual’s tumor," he says. Combining that genetic analysis with classic parameters—serum PSA, Gleason grade, and extent of tumor—may provide a more accurate estimate of each tumor’s biological potential and make it possible to tailor therapy based on that patient’s risk.

"We are just now starting to study those genetic parameters," says
Dr. Muller.

William Check is a medical writer in Wilmette, Ill.