In breast cancer diagnoses, don’t balk at talk

April 2007
Cover Story

Karen Titus

Medicine is terribly complex, yes, but parts of its practice should be simple. For instance, information between physicians should flow freely, like beer from a barrel.

It stands to reason that pathology should not be practiced in a vacuum. With a field as complicated as breast pathology, it stands to reason that pathologists and radiologists should be sharing information with each other. And it stands to reason that both, in turn, should be speaking regularly, and intelligibly, with oncologists and surgeons.

What seems reasonable is now being supported by evidence. An article published last fall in Cancer (Newman EA, et al. 2006;107:2346–2351) looked at how multidisciplinary tumor board case review affects surgical management of breast cancer patients.

In the study, which took place at the University of Michigan, Ann Arbor, review of pathology resulted in interpretive changes in 43 of the 149 patients (29 percent) whose cases were reviewed. Thirteen patients (nine percent) had surgical management changes based solely on pathologic reinterpretation.

Review of imaging studies resulted in changes in interpretations in 67 of the 149 patients (45 percent), resulting in surgical management changes in 16 patients (11 percent). Overall, the researchers reported, a second evaluation by the multidisciplinary tumor board led to changes in recommendations for surgical management in 77 of the 149 patients (52 percent).

Pathologists have greeted this study with, well, not a yawn, exactly—more like a knowing nod. They see the problems mirrored in their own experiences.

“It really resounded with me when I read the study. This is exactly our experience,” says Michael D. Lagios, MD, director of breast services at St. Mary’s Medical Center, San Francisco, and clinical associate professor of pathology, Stanford University and University of California. “There are significant errors that are evident on review of material, very commonly in radiology and equally commonly in pathology”—about five percent of cases reviewed, he says.

“Let’s see: Nine percent had a surgical change—that’s about right,” says Ira Bleiweiss, MD, looking at the Michigan pathology figures. The 29 percent figure “is about midrange.” Dr. Bleiweiss is director of surgical pathology, director of the Division of Breast Pathology, and professor of pathology at Mount Sinai Medical Center, New York. “In my practice, I would say that probably 50 to 60 percent of my consults have—I don’t want to call it error—have something different in my interpretation relative to what was submitted.”

The Michigan study essentially blended two forms of review: multidisciplinary tumor board review and second opinion by an outside institution. The study retrospectively looked at the medical records of 149 consecutive patients who were referred to UM’s Comprehensive Cancer Center after being diagnosed with breast cancer and initially evaluated elsewhere.

Is it possible to separate those two strands? As Elizabeth L. Wiley, MD, points out, breast cancer is the source of innumerable study protocols. Some are the province of individual institutions, while others are tucked away at special oncology groups, which makes it impossible for the average physician—including pathologists—to know about them. “So very minor things may impact on treatment assessment for patients when they come to another institution,” says Dr. Wiley, professor and director of surgical pathology, University of Illinois Medical Center at Chicago.

It is important to make the distinction between the two types of review, says Dr. Lagios. Multidisciplinary tumor boards, with independent review of material, have advantages in that experts review all the material separately.

Dr. Lagios does second opinion reviews of his own, apart from his participation in the tumor review board at St. Mary’s, as medical director of his private practice, Breast Cancer Consultation Service. In his own practice, he notes, the interpretation changes in approximately 15 percent of cases, a third of which are significant—benign versus malignant, for example, or invasive versus in situ. Some 40 percent of his cases come from medical colleagues, including pathologists but primarily radiation oncologists, medical oncologists, and surgeons, and 60 percent are second opinions requested by patients.

For Dr. Lagios, the real surprise comes from those patient-referred cases. While cases from colleagues are often referred because they’re difficult, those from patients, who generally aren’t schooled in diagnostic dilemmas, tend to be more mundane, at least on the surface. “You’d be surprised at the number of times where straightforward bread-and-butter invasive carcinomas are misinterpreted in one respect or another, either in terms of size, which is very common, or the grade or margin status.”

The real issue, it emerges, isn’t type of review, but rather the quality of each. The best second opinions are far more than simply popping your head in the office next door. At a high level, second opinions can function almost like a miniature tumor board, with review of clinical information and imaging studies as well as the pathology. Likewise, the best tumor boards should be more than a dog and pony show. For both, successful breast pathology boils down to something simple: talk early and often. But talk about what?

One of the Cancer study’s authors, Lisa A. Newman, MD, MPH, director, Breast Care Center, and associate professor of surgery, University of Michigan Comprehensive Cancer Center, says borderline breast malignancies are a good place to start. “The distinction between some cases of severe atypia versus ductal carcinoma in situ versus microinvasion is not always straightforward, but the treatment recommendations will vary significantly on the basis of the pathologist’s final determination,” she says. Pathologists should let surgeons know of any uncertainties in a diagnosis and suggest ways to resolve them, such as obtaining additional tissue for analysis (especially important if the initial biopsy was via needle, she says), obtaining a second opinion, or both.

Dr. Newman stresses the need for clear descriptions of margins for cancerous lumpectomy specimens. Decisions about additional breast surgery “require detailed information about the minimum width of margin negativity, and if there are areas where the margin is ‘close,’ we would like to know the measurement for this margin and whether it spans over multiple foci versus a single focus.”

“It’s not enough to say that the margin is ‘close.’ We have to say how close,” agrees Celina Kleer, MD, lead breast pathologist and associate professor of pathology, University of Michigan, and another of the study’s authors. “Is it 1 mm? 2 mm?” Invasive carcinomas with margins greater than 2 mm have a lower chance of tumor recurrence compared with tumors with margins less than 2 mm. “If that information is not conveyed properly to the surgeons, they may not offer reexcision, and then the tumor may recur.”

Breast biopsies for microcalcifications also warrant special attention, Dr. Newman says. “If cancer is identified within the specimen, then we need to know whether the calcifications were associated with this cancer.”

The more specific pathologists can be with their language, the better, says Dr. Kleer. “For example, we need to be specific when discussing invasive tubular carcinoma versus invasive ductal carcinoma with tubular features.” In the former case, she says, chemotherapy may only be offered to patients with larger-sized tumors; in the latter, chemotherapy may be an option with smaller tumors.

When they talk about communication, particularly with radiologists, most pathologists have at least one favorite case they like to point to, which they recount as if they were cops swapping arrest stories.

Dr. Bleiweiss’ illustrative case occurred more than a decade ago, though the particulars remain clear in his mind.

The matter involved a very large, well-circumscribed breast mass that was excised without biopsy. In microscopic appearance, it was a dead-ringer for invasive lobular carcinoma. “If you took an isolated core biopsy of this big mass and showed it to anybody, they would have called it invasive lobular,” says Dr. Bleiweiss.

“And they would have been wrong.” Invasive lobular is never well circumscribed, he notes.

The mass (described in Arch Pathol Lab Med. Garfein CF, et al. 1996;120:676–680; Dr. Bleiweiss was the senior author) was a myoid hamartoma, an uncommon, completely benign breast lesion. But without the correlative radiology report, no pathologist then or now would have been able to make the correct diagnosis on a core biopsy.

Dr. Lagios mentions two cases, both of which appeared to be straightforward. One was a patient with a T1c-sized intermediate-grade invasive cancer, who was node negative. The radiation oncologist who referred her to the tumor board wanted to know if the margins were clear. A look at the pathology report revealed spotty documentation: The tissue had merely been sampled, and no specimen x-ray had been taken. The operating report revealed, remarkably, that the localization wire was transected; the surgeon had to fish around, Dr. Lagios says, for the distal half of the wire and hook.

All this is pertinent, Dr. Lagios says, because the initial margins were involved over a large area. Anyone reading the pathology report would immediately have thought the additional tissue, the second specimen, was a reexcision of the clearly involved margin. It wasn’t, of course—it instead comprised the little pieces picked up by the surgeon at the site where he found the needle tip. In fact, it wasn’t at all certain that the margins were clear. Additional imaging studies were required to confirm the lesion had been completely removed.

Another case—“one of my favorites,” says Dr. Lagios—involved a case of DCIS. “Reading the path report, you would have believed this was very small focus, of 3 mm size, with wide margins.” A review of the imaging material on the multidisciplinary board showed at least 25 mm of microcalcification scattered in a segmental fashion. Though the patient had undergone a core biopsy, the subsequent needle-localized lumpectomy was sampled rather than fully processed.

“It’s true that there was only one focus of residual DCIS, but we don’t know what else was present. We do know that there were other areas of calcification that were not sampled from the tissue that was submitted. And the rest of the unsampled tissue, of course, is thrown away,” Dr. Lagios says. “So this lady is going to need additional imaging, most likely an MRI, to see what her real status is in terms of excision.”

For breast core biopsies, “You need to know what the radiologist is seeing,” Dr. Bleiweiss says. He runs down the list: Is it solid? Is it cystic? Is it partially solid, partially cystic? Is it well circumscribed or irregular? Did the lesion shrink or disappear? Is the specimen x-ray available for calcifications?

Since time immemorial, x-rays accompanied needle- or wire-localized specimens when they made their way back to the surgeon or the pathologist. Today, those same lesions are still being targeted, with a different methodology. “It stands to reason you want to do the same thing with core biopsies,” says Dr. Bleiweiss. “And it’s amazing that it doesn’t happen routinely.”

“We occasionally will get specimens that are done by radiologists who, for whatever reasons—insurance purposes—send a specimen to a lab that then subcontracts to us,” he continues. “So occasionally we’ll get a specimen that is done for calcifications, coming from some radiologist whom I don’t work with routinely, or whom I’ve never heard of. And he’ll write ‘calcification’ on the form, and I’ll immediately make a phone call and say that I need to see the specimen x-ray. And he’ll say, ‘Why? Nobody’s ever asked me that before.’”

The answer, says Dr. Bleiweiss, is simple: You need only one bad case to know that getting the x-ray is the right thing to do.

He points to calcification cases where the radiologist does a core biopsy of linear-branching calcifications but also removes other, less suspicious or nonsuspicious calcifications from the same area, all of which are seen on the specimen x-ray. If the specimen arrives and no specimen x-ray is available, the potential for problems is enormous.

“We get four levels on each block,” he explains. “Say the calcifications are all in benign tissue. And so I write a report: ‘fibrocystic changes with calcifications.’” Looking at the x-ray, if it’s available, will confirm the low-suspicion calcifications found in benign tissue. But it may also reveal a linear-branching structure that can’t be explained by his initial slides. Dr. Bleiweiss will then radiograph the paraffin blocks. “Lo and behold, that little linear-branching structure is still in the paraffin block. So then I’ll go and cut deeper levels, five, 10 more levels. And surprise surprise, it’s now DCIS.

“Now, if I don’t have that specimen x-ray with me, I don’t know that I have to do that,” he continues. Alarm bells should go off for the radiologist, who receives a report of benign but was expecting a malignancy. “But I don’t know that all radiologists are that obsessive-compulsive, shall we say. I’ve had this happen a couple of times. In fact, it’s probably a reason for some of the issues in the published literature, about upstage rates of certain lesions, such as LCIS on core biopsy.”

Dr. Bleiweiss says he or one of his colleagues speaks to their radiologist colleagues “about every single core biopsy case.”

It’s not as onerous as it sounds. Some of those interactions are the equivalent of speed dating. “That discussion could be two seconds long, because some cases are very obvious—it’s clear what they’re after, the correlation is perfect,” Dr. Bleiweiss says. The radiologist could be targeting an oval, solid, well-circumscribed nodule, for example, and the core biopsy’s a fibroadenoma. “End of discussion,” says Dr. Bleiweiss. “On the other hand, if the radiologist says that the lesion is an irregular solid mass, and all I see is benign fatty breast tissue with a tiny little fibroadenoma, then I’m going to have a discussion, because I’m pretty confident she missed the lesion.”

Tumor review boards are an opportunity for meaningful discussion, but not all tumor boards are created equal. “Sometimes the tumor board is simply a show-and-tell session,” Dr. Lagios says. “The pathology department presents a couple of representative slides and proclaims what the findings were, and that’s just accepted on face value. So there are different levels of review.”

At his institution, the treatment management conference, held every Monday, is a teleconference between four different hospitals. In that sense, he says, it really tends to be more like an independent board, “because we’re dealing with four different entities.”

That setup came about after Dr. Lagios joined the hospital in 1994. Establishing the board was a joint goal of his and the administration, he says. Was it difficult to achieve? Replies Dr. Lagios: “If you were to see me in person, with the thinning gray hair and the scars, you would realize how difficult.” He doesn’t laugh when he says this.

“Some of my colleagues are very innovative and forward-thinking, but to be honest, those are the minority,” he says. The ongoing frustration for physicians everywhere, he says, are recalcitrant colleagues. “There are still surgeons who, for example, don’t do sentinel node biopsy, who still use non-image–guided—that is, blind—reexcision of an imaged abnormality because they can feel the ‘wound site,’” he says.

At her last post, at Northwestern University’s Feinberg School of Medicine, Chicago, Dr. Wiley says the breast tumor board functioned more as a working conference for planning patient therapy than as a teaching tool. “As a result, there was high interest in the pathologic assessment of the cases,” says Dr. Wiley. That was quite a wakeup call for her and her pathologist and radiology colleagues, who quickly discovered that much of the information demanded by their oncologists was available but simply hadn’t been included in the report. Mammographers, for example, had neglected to include an objective measuring instrument on the radiograph to denote lesion size.

Dr. Wiley found another way to help oncologists plan treatment: While at Northwestern, she began doing tumor staging summaries and tumor markers on the cores, which let oncologists know the profile tumor before starting chemotherapy. “In fact, what happened was a couple of the surgeons actually converted to core biopsy, instead of doing FNAs on large, palpable tumors, simply because they would have that information up front to talk to their patients about planning therapy.” She also established cumulative pathologic staging for the cores, a practice she eventually hopes to bring to UIC. “My colleagues in oncology, in radiation oncology particularly, appreciated having the cumulative summary, so that they didn’t have to go leafing through multiple specimens, multiple notes, to find out the definitive pathologic staging.”

These weren’t mind-boggling changes to make, and changing tumor board focus is well within the reach of any institution. It’s not that the educational component is neglected. Instead, the end-of-presentation discussion is aimed at giving direction to the next group of caregivers. The key, Dr. Wiley says, is scheduling patients in a timely fashion so their cases are available for discussion. “You have to have an administrator or a nurse practitioner, someone, who can take ownership of scheduling patients for review on the board.” Once the ball got rolling at Northwestern, she says, the breast surgeons made it a priority to bring their patients’ cases for tumor board discussion early.

Dr. Kleer says the University of Michigan was one of the first institutions to create a multidisciplinary tumor board to discuss breast cancer cases. It was her mentor, pathologist Harold Oberman, MD, and his clinical and surgical colleagues who established in the early 1980s the breast care conference—a forum for discussing cases with the aim of planning treatment. Pathologists were involved in the decisionmaking from the start.

You don’t need a lot of resources to simply talk to physicians, says Dr. Kleer, who adds that her colleagues “are always grateful when we discuss a case with them.

“Don’t be afraid to pick up the phone and exchange information. Maybe there’s not a chance of getting everybody together in a room with a projector and microscope, but if questions or unexpected diagnoses arise, then it’s very important to discuss the case with the radiologists, clinicians, and/or surgeons.”

It may, in fact, be easier to simply schedule regular tumor board review meetings. At Michigan, the weekly meetings last an hour and a half. But that covers a large caseload—20, 25 cases—which would not be the case at every institution.

Regardless of the approach, the goal is the same: Keep talking. Breast pathology, as it turns out, is a convivial affair—even without the beer.