Better Send it to Epstein

By Janet Farrar Worthington

Prostate cancer cells may be the hardest of all to analyze. and they can spell big trouble. One pathologist has deciphered more of these splotchy microscopic patterns than anyone else in the world.

You are a man who may or may not have prostate cancer, and you are sweating bullets. You're still sore from the needle biopsy, and you're waiting for some guy you've never met-some faceless pathologist- to make the call. Is it cancer? Is it treatable? Now, switch roles: You are that pathologist, and a man's life may depend on what you can glean from a few tiny cores of prostate tissue. Did the urologist who sent you the biopsy nail the cancer, or miss it entirely? Consider that the prostate gland is roughly the size of a large strawberry and in it, a patch of cancer-the average cancerous prostate has about seven-is about the size of a strawberry seed. The cancer cells that are found generally tend to be hard to interpret; thus, biopsy is often a hit-and-miss affair. And now it's in your court. Your judgment will be a major part of the treatment decision-making. Is it cancer? Maybe yes, maybe no. The best you can determine is that it's "atypical." Better send it to Epstein.

Jonathan Epstein is world-renowned for his expertise and accuracy in judging prostate cells. He has probably examined more prostate tissue than any other pathologist. A pioneer in the budding specialty of urologic pathology, Epstein also has the distinction of holding the world's only endowed chair in the field; this year, he became the Rose-Lee and Keith Reinhard Professor of Urologic Pathology.

"He's unsurpassed in what he does," says Patrick C. Walsh, director of the Brady Urological Institute. The idea of a chair for urologic pathology was Walsh's idea-a way to honor and support Epstein's considerable research contributions. "In general, pathologists are the unsung heroes," Walsh explains. "Patients sit in their doctor's office and receive some of the most important information they'll ever hear in their lives-the verdict on whether or not they have cancer-and it comes from the pathologist, a person they'll never know." Walsh wants Epstein to be, well, sung. "Jonathan Epstein wrote the books on the diagnosis and prognosis of prostate cancer. He's one of the reasons our work in prostate cancer is so well respected: Our patients start with the scientific opinion of the best pathologist in the world."

Selecting the slides he'll use for a lecture at another university.

It is Epstein's pathology work, for example, that gives remarkable continuity to the Partin tables, developed at Hopkins by urologists Alan Partin and Walsh as a means of estimating the exact extent of prostate cancer. In the span of a few years, these tables have become indispensable, with many patients as well as doctors trying to chart the right course of treatment in a sea of confusing choices. (To give you an idea of their importance as a layman's tool, a recent Google search on "Partin Tables" turned up 1,220 entries on the Internet in English alone.) The tables correlate three key pieces of the prostate cancer puzzle-clinical stage (the estimated extent of the cancer), the grade of cancer (based on the biopsy), and the amount of prostate-specific antigen (PSA, an enzyme made by the prostate) present in the blood- with the actual pathologic stage, determined when Epstein examined 5,000 surgically removed prostate specimens. The tables predict, with 95 percent accuracy, a man's likelihood of having organ-confined disease, cancer that has penetrated the prostate wall, or cancer that has spread to the nearby seminal vesicles or lymph nodes.

Epstein discovered that he had a knack for deciphering prostate cancer cells in 1982, during his pathology internship at Hopkins. Fresh out of the Boston University School of Medicine, he was tapped by Joseph Eggleston, then head of surgical pathology, to work with urology. "They had a project, a very tedious project, where you actually had to circle cancer cells under the microscope. It would take hours."

Epstein recalls it vividly-mainly because he had a case of shingles at the time. "I was in torturous pain, and I was sitting there without a shirt on, because I couldn't have a shirt touch the shingles, circling these cancer cells forever." The study was published in Cancer, and the urologists, impressed with his stoicism as well as his work, kept right on collaborating with Epstein. He finished his residency at Hopkins, did a fellowship at Memorial Sloan-Kettering, came back to Hopkins to serve as chief resident in pathology, joined the faculty, and now is a professor in pathology, urology, and oncology.

Epstein was a full-time urological pathologist before there was such a thing. "When I started out, there was absolutely no such specialty," he says. At national pathology meetings, "there would maybe be three talks and a handful of posters on prostate cancer," and Epstein's own work would account for at least two of those. "Last year, there were 150 posters and talks. There were no fellowships back then. Now there are seven or eight in the country. There's a Urological Pathology Society, there's now the Journal of Urological Pathology-so it's really coming into its own." Epstein gets calls from around the country from hospitals looking for urological pathologists, and as yet, "there are not enough of us around."

For hundreds of pathologists and urologists, and thousands of their patients, Epstein remains the fallback position-a rock in the unsteady world of hollow-core needle biopsies. Which begs the question: What is it about prostate cancer that makes it such a stinker for pathologists to interpret?

"Of all biopsies," explains Epstein, "prostate biopsies are probably the hardest. You're dealing with such a limited amount of tissue." Until very recently, a prostate biopsy consisted of six measly needle cores of tissue, taken from throughout the gland; It's now routine at any sophisticated medical center to take 12 samples, more if a prostate is particularly enlarged. Plus, he adds, prostate cancer tends not to form a conveniently solid mass; instead, it's more like a veneer over an expanse of healthy tissue. (Using the strawberry image, it's not only a tiny dot, it's a flimsy one.) And the veneer is often maddeningly ambiguous-so, not only can the biopsy needle overshoot and miss the cancer, the cancer cells it does get don't always match the pictures in the textbook.

Part one of the challenge for pathologists is simply knowing how to make sense of the cells they see. Under the microscope, prostate cancer looks like a splotchy work of modern art: Imagine countless shades of gray, some nearly white, some nearly black, most subtle variations of indeterminate shades somewhere in between, vomited onto a canvas. It's a mess, a disgraceful jumble of cells that run the gamut from the almost ordinary-looking to runny, malignant blobs, cells that are so poorly differentiated and obviously diseased that they could never be considered normal. Worse, cells taken from one part of the prostate may look one way, and those from another part may look completely different.

One result of this is the biopsy labeled "atypical"-a diagnosis that appears in about 5 percent of biopsies at most institutions, says Epstein. "What that means is that a pathologist will see something that could be cancer, but doesn't feel comfortable saying it's definitely cancer." The next step is having a repeat biopsy-and the value of this is often questionable, he says. "The problem is, in about 20 percent of cases, the biopsy can miss cancer-so even if it's negative, it doesn't mean the patient doesn't have cancer. In fact, the cancer can be extensive. We've seen some missed entirely; they were called totally benign, yet they were cancer."

Second opinion pathology makes the difference between the right and wrong treatment. The two images were taken from the same case. On left: An image from a prostate needle biopsy that was originally diagnosed as carcinoma. Numerous crowded glands (arrows) with central lumina are strongly suggestive of prostate cancer. Treatment typically would consist of either radical removal of the prostate or radiation therapy. On right: On second opinion examination of the case, subtle features under the microscope were recognized that showed the lesion was adenosis, a close mimicker of prostate cancer. A special stain detected the presence of basal cells (arrows), which are absent in prostate cancer, and verified the benign nature of the case.

DIAGNOSING PROSTATE CANCER CAN BE like trying not to fail a hellish multiple-choice test-the kind with bewildering answer options like "All of the above" and "Other." Epstein has shed considerable light on the "Other"-called PIN, or prostatic intraepithelial neoplasia. PIN cells, found in the tissue lining the prostate, are "funny-looking." They're not cancerous, but they're not benign, either. They're abnormal, and they are strongly linked to prostate cancer. Like cancer, PIN has its own distinct patterns, and high-grade PIN, Epstein believes, is a marker for prostate cancer. "We'll often find high-grade PIN next to cancer." If high-grade PIN is found but cancer is not, instead of indicating that a man's prostate is cancer-free, the more likely scenario is that cancer may have been missed, and a man should have another biopsy. Epstein's research continues to define these not-quite-cancerous cells; in fact, he just discovered a new form of high-grade PIN, which he named after its resemblance to a hobnail.

The other end of this vexing spectrum, Epstein has found, is that many pathologists seem just as likely to overdiagnose cancer. "There are many mimickers of prostate cancer under the microscope, and people not as familiar with prostate biopsies can diagnose cancer when it's not." About six to eight men who come to Hopkins each year with a diagnosis of prostate cancer are found to have been misdiagnosed. But even biopsies that seem straightforward deserve another look.

Epstein and colleagues at Hopkins have done numerous studies looking at the reproducibility of Gleason scores (named after another legendary pathologist, this is the grading system for prostate cancer) in the general pathology community, comparing the biopsy's Gleason grade to the actual prostate specimen removed during surgery, "and found that by and large, the Gleason grading that's performed is disappointing. All across the map, it doesn't correlate with what you see in a radical prostatectomy. People are having decisions made based in part on a Gleason grade, when it's not accurate at all."

Part of the problem is that at most hospitals, pathologists don't have the luxury of specializing in prostate cancer. "Even in academic centers, maybe 5 percent would have somebody whom I would say is a urological pathologist."

Epstein recommends that any man about to undergo treatment for prostate cancer get a second opinion from a pathologist. "It's just as important as getting a second opinion for surgery or radiation," he notes. "You could have the best surgeon in the world, but if you don't have the right pathology, you could have the wrong thing done for you." In a recent six-month period, Epstein and colleagues looked at 3,000 consults, "almost 700 of which were sent at the request of either the patient or the urologist. Overall, we changed the diagnosis about 35 percent of the time-which to me was quite striking because these were not cases where the pathologist even had a problem."

In an effort to improve prostate cancer diagnosis, Epstein and colleagues devised a tutorial for the Internet-a website for pathologists. "We just did it on our own," he says, "because we think pathologists can do better than they have been. The key is education, not just getting frustrated." Epstein has tried other approaches, such as articles and teaching. In addition to working with medical students, urology and pathology residents, he gives lectures to pathologists, urologists, oncologists and even patient groups worldwide.

The online course, the first of its kind, takes about an hour. First, pathologists are asked to grade a set of biopsies. Then, they're shown some of the telltale signs of various grades-Epstein calls them "tricks of grading"-and taught how to interpret another set of biopsies. Finally, they are asked to re-evaluate the biopsies. "We've found that pathologists can make a dramatic improvement, just in this brief tutorial." (Pathologists interested in learning more can find the website at www.pathology.jhu.edu/prostate.) So far, more than 2,000 pathologists have taken the short course. Epstein is encouraged by this, and by a change he has noticed in the kinds of biopsies he's receiving.

"Even five years ago, when I got consults, often there were cases with extensive cancers that pathologists still just couldn't recognize, and larger cancers than you would have thought would have given them difficulty. There were some difficult cases, but there were also many cases where it was clearly cancer." Now, the consult cases that he is getting are more difficult-a good sign, Epstein believes.

Much of Epstein's research has attempted to predict the behavior of prostate cancer-"which is tremendously variable, much more than most tumors," he says. "There's such a dramatic difference." Some prostate cancer is indolent; it grows slowly, and may never need to be treated. (Going back to the canvas image, these are the most well-differentiated and ordinary-looking cells.) Other prostate cancer (the most poorly differentiated cells) attacks the body with the viciousness of a pit bull. Without treatment, this most aggressive, out-of-control cancer can kill a man within several years of its initial clinical presentation.

"It's a constant striving to find out which patients need the treatment, and to make sure they get it, but also to keep someone from being overtreated," Epstein says. "In the past, doctors used to think that certain findings were invariably associated with an adverse outcome, and were treating these patients very aggressively."

For example, urologists believed that men who were found to have tumor outside the prostate, or to have positive surgical margins-cancer cells on the edge of the surgically removed prostate-should also receive external-beam radiation. Epstein's years of studying prostate specimens and, with urologists Walsh, Partin and H. Ballentine Carter, correlating the findings with the long-term outcome of the patients have taken much of the fear out of the words "positive margins," and have saved many men from unnecessary radiation.

Sometimes, margins are too close to call. What then? In an ideal world, the pathologist would immediately send a triumphant report to the surgeon: "I've looked at the prostate tissue you removed from Mr. Smith, and all the edges are clear. You've removed all the cancer!" At Hopkins, fortunately, it often happens this way; fewer than 10 percent of men who undergo radical prostatectomy, surgical removal of the prostate, are found to have cancer at the margins. Recently, Epstein has shown that these close margins are nearly always negative. He studied men whose tumors were particularly close-a hair's breadth, less than two tenths of a millimeter-from the surgical margin. Even though there wasn't a comfortable cushion of tissue between the tumor and the edge of the prostate, he found, "those patients do fine after surgery and have an extremely low risk of recurrence. Because of that, we don't even specify how close the tumor is from the margin in our specimens; we just say the margin is positive or negative, so as not to unduly worry patients."

And now, it's author's confession time: When I found out that my father had prostate cancer five years ago, the first person I called was Patrick Walsh, with whom I have written two books about prostate cancer. The second person I called, after my parents Fed-Exed Dad's biopsy slides to Hopkins, was Jonathan Epstein. I called him early, because he likes to start work in the predawn hours when Hopkins is at its quietest, when the phone rarely rings, the only e-mails he gets are from Europe, and "I can sit there and do my work." In his office, surrounded by photos he's taken from his many trips overseas (often with his wife and two sons) he confirmed the diagnosis made by the local hospital in South Carolina- Gleason 6. But he also cautioned that when pathologists examine the removed prostate, the Gleason grade often turns out to be a notch higher than the biopsy suggested. He was right. Dad's prostate, removed by Walsh and thoroughly examined by Epstein, turned out to be Gleason 7-higher, but still within the range considered curable. He remains cancer-free today.