March 2010
Feature Story

Karen Titus

In the battle that is breast cancer, micrometastases are a little like Liechtenstein: a principality that may not be principal, despite appearing to be smack in the center of things.

Physicians want to know two things about micrometastases, says Aysegul Sahin, MD, of the University of Texas M.D. Anderson Cancer Center, Houston. “What is the biological significance? And what should we do?” says Dr. Sahin, professor of pathology and section chief of breast pathology. Do they indicate systemic disease, requiring systemic therapy? And should surgeons complete an axillary lymph node dissection if micrometastases are found in the sentinel lymph node?

The last decade has brought a little clarity, but hardly enough. The NSABP’s B-32 protocol, a randomized, phase III trial to compare sentinel node biopsy to axillary dissection, was begun more than 10 years ago, and hopes were high provisional data might be available before the decade’s midpoint. But 2004 came and went without it, as did 2005, and 2006....

The problem, says Donald Weaver, MD, protocol pathologist for the trial, was that researchers needed enough recurrence events (50 percent) to trigger the statistical analysis. Dr. Weaver has a personal interest in this as well as a professional one: He’ll be writing up the results of the trial, once the 50 percent figure has been reached and the databases are “cleaned up” prior to the analyses. At one point he thought he’d be devoting last summer to writing. This winter, he was still wondering. Would it be late this spring? Summer? “I need to clear my schedule to be able to write,” says the bard of B-32, who is also professor of pathology, University of Vermont College of Medicine, Burlington, and attending in pathology and director of surgical pathology, Fletcher Allen Health Care.

A proposed study, the American College of Surgeons Z-11 sentinel node trial, could have provided much-needed insight into whether axillary dissection was necessary in patients with positive sentinel lymph nodes. But the trial closed early for lack of patients. Those with node positivity didn’t want to be randomized to the “no further surgery” arm of the trial. “Most patients insisted on further node dissection,” says Roderick Turner, MD, adjunct member, John Wayne Cancer Institute at Saint John’s Health Center, Santa Monica, Calif.

Then there’s the 7th edition of the AJCC Cancer Staging Manual, which took effect Jan. 1. Depending on whom you ask, the manual has either helped or hindered matters.

Dr. Turner puts himself in the former camp, saying it provides clearer definitions of isolated tumor cells and micrometastases. The prior edition, he says, caused vexing reproducibility problems for pathologists.

He published on this (Turner RR, et al. J Clin Oncol. 2008;26:258–263), reporting that the 6th edition had led to wide variability in distinguishing between micrometastases (pN1mi) and isolated tumor cells [ITC; pN0(i+)], while more clearly defined histologic criteria, perhaps not surprisingly, led to more reproducible results. Those criteria are part of the 7th edition.

The updated manual also adopted an added criterion: In any given lymph node profile histologic section, the presence of more than 200 tumor cells is classified as micrometastasis, rather than ITC. “That specifically addresses the problem area of lobular carcinoma metastases,” says Dr. Turner, noting that this commonly, with lymph node involvement, demonstrates a dispersed pattern, rather than cohesive clusters of carcinoma cells. The new criterion is well described in the 7th edition, he says, which should lead to more consistent reporting.

Or maybe not. Asked about the 7th edition, Dr. Sahin says, “Actually, in my opinion, it makes matters a lot more difficult.” The manual lacks clarity, she says. “And some of the comments are, in my opinion, not very practical.”

The new edition, for example, says cluster sizes less than 0.2 mm should be called isolated tumor cells. Those larger than 0.2 mm should be classified as micrometastases.

“But in practice, what happens is sometimes you can see more than several clusters,” says Dr. Sahin, the lead author of a recent review paper looking at current controversies regarding micrometastases and sentinel node biopsies (Arch Pathol Lab Med. 2009;133:869-878). It’s difficult to know whether such clusters should be assessed together, or measured individually, she says; the new guideline calls for the largest cluster to be measured. “But there are many, many lymph nodes where you know there’s metastatic tumor, especially for patients with invasive lobular carcinoma. There are multiple clusters. And each cluster measures less than 0.2 mm. But we never hesitate to call that metastatic carcinoma, because it’s the known metastatic pattern of invasive lobular carcinoma.” In these cases, she fears, strictly following the latest AJCC definition would require classifying the clusters as isolated tumor cells. “From the biology, we know that’s not correct,” she says.

Dr. Turner and Dr. Sahin aren’t firing from the fringe. Dr. Weaver, who helped write the 6th and 7th editions, says he and his colleagues did indeed wrestle with the ITC definition. In a minority of cases of invasive lobular carcinoma, he explains, a lymph node can contain singular dispersed tumor cells. This could signify a fair amount of tumor burden in the node, but couldn’t be measured as being larger than 0.2 mm. “That drove some people nuts,” says Dr. Weaver. “People like pretty firm rules, and they don’t like to break them, I find, when talking about staging.” Multiple single tumor cells in a lymph node put them in a difficult situation.

That led to the change mentioned by Dr. Turner. “We wanted to give people an out,” Dr. Weaver says, one that didn’t diminish the importance of tumor volume. “We said, ‘OK, if you’re willing to count the cells in a single cross-section of the lymph node, and find more than 200 cells in that cross-section, don’t call it an ITC; call it a micrometastasis.’” The authors based their reasoning, in part, on rough volume calculations. A single ITC cluster of 0.2 mm contains about 1,000 cells, Dr. Weaver says; finding 200 in a single cross-section indicates a high probability of there being more than 1,000 in the entire node.

Such discussions are typical of the topic of micrometastases and sentinel node biopsies. There are endless threads to pluck, like unraveling the causes and goals of the Thirty Years’ War. But even after you’ve pondered the various parts, it’s still hard to figure out the whole, with so much data MIA.

When they do turn up, it’s possible they’ll be greeted by a collective yawn. It’s not that the data will be suspect. But breast cancer treatment is often shaped by the axiomatic reasoning that more is better. Medical oncologists and surgeons, says Dr. Sahin, often err on the side of overtreatment, in part spurred by patients who want to do whatever they can to prevent a recurrence. “So, many oncologists say, ‘OK, we don’t have the data, but let’s treat you.’ And then you’ll never have the answer, because there’s no randomized study.”

Dr. Sahin isn’t the first person to make this observation, but it bears repeating: Breast cancer patients often become significantly involved in their care. And since physicians may lack concrete data, she says, their guidance is often couched as, “This is my recommendation; this is what we do as routine clinical practice.” Patients then choose a course of treatment, aided in part by their own enquiries. “Sometimes they do not want to change their opinion, even after hearing the recommendation of the surgeon or the medical oncologist,” says Dr. Sahin.

Patients are also using their own personal barometers to gauge risk, she says. The same information, with the same numbers, may imply high risk to one patient and low risk to another.

All this is good, in Dr. Sahin’s view. It even provides opportunity for pathologists, who, after all, know a thing or two about transforming lab data into practical guidance. “It forces physicians to present the facts in a very logical fashion,” she says.

It’s an unfortunate reality of medicine, she says, that physicians tend to interpret and present data based on their own beliefs, whether they’re talking about hypertension, high cholesterol, or cancer. “You introduce some bias into it,” she says. So when breast cancer patients educate themselves about their disease, “It forces physicians to make sure they interpret the data in a non-biased, clear fashion.”

Pathologists can help by reminding their clinical colleagues of the finer points of detection methods, Dr. Sahin says.

Current wisdom among nonpathologists is that either immunohistochemistry or molecular methods are much more sensitive than histologic evaluation. “New” means better sensitivity and specificity, goes the thinking. Likewise, she says, for surgeons, the number of sections evaluated tends to be the endpoint. “When you go to clinical conferences, everybody says, ‘How many sections are evaluated, and how many immunohistochemistries are done?’” says Dr. Sahin. “Because most people superficially understand that the more advanced techniques that you use, the more sections you look at, the more metastases you’ll pick up. Which is superficially true.”

It’s true, that is, if the lymph node is grossed appropriately to start with. But if it isn’t, the search is less rewarding, like selecting the finest Bordeaux at Walgreens. The lymph node needs to be sectioned very thinly to start with, Dr. Sahin notes. The goal is 2-mm sections, since macrometastases are defined as being more than two mm.

But what about micrometastases and ITCs? Dr. Weaver weighs in, drawing on information from the B-32 trial that he’s spoken about in public presentations, but not yet published. To reassure physicians and patients that there is a very, very low likelihood of a micrometastasis being in the nodes would require pathologists to look at sections every 200 microns—in other words, at least 10 sections systematically spaced through a 2-mm-thick section of the node. Dr. Weaver expresses doubt that such effort is worth it, either medically or economically. “And if you wanted to say, ‘I didn’t miss any isolated tumor cell clusters,’ now you’ve got to cut 10 microns all through the block. And that’s just ridiculous,” he says. “That’s 200 sections. No one’s going to do that.” That may be one reason no one’s talking about chasing down ITCs, he says.

In the NSABP protocol, 10 sections per block were out of the question, given the cost. The researchers instead took extra sections at a half-millimeter and one millimeter into the block, primarily to see if they were missing macrometastases, Dr. Weaver says. But it also let them pick up “a lot” (around 14 to 15 percent) of micrometastases and isolated tumor cell clusters simply by chance, he says. The implications aren’t known yet, since the findings have not been correlated with outcome. “We’re all sort of slightly holding our breath,” he says.

If the B-32 trial is positive—if it provides results that physicians and patients want to act on, in terms of treatment decisions—that could be a reason to provide the two additional sections, Dr. Weaver says, since this was done in the trial. Anything beyond that would be speculative.

There might be another reason to consider doing the extra cuts. Not everyone achieves the 2-mm goal. Making thicker cuts—a not uncommon occurrence in labs, says Dr. Weaver—means some macro­meta­stases could go undetected.

Pathologists could also remind their clinical colleagues that IHC is not, in fact, formally recommended or considered mandatory, despite its widespread use.

That being said, Dr. Sahin says she’s a big fan of IHC. “From the pathologist’s point of view, it is a lot easier, no question. And it really increases your speed.”

She’s fully aware of the arguments over IHC’s worth and cost-effectiveness. “Those are issues people argue, and they’re valid,” she admits. “But as a pathologist I like IHC.”

At Fletcher Allen, Dr. Weaver shuns routine use of IHC. “People are always surprised by that,” he says. The decision was tied to the B-32 protocol, where pathologists could use IHC to evaluate a suspicious finding on an initial section but did not have carte blanche. When the trial was over, Dr. Weaver approached his colleagues in surgery and oncology and asked if they wanted the lab to start doing IHCs (and deeper cuts). “They looked at me with mouths gaping open and said, ‘Are you crazy?’” Granted, he says, his institution is the only tertiary hospital in the area, and thus is not competing with other places that might offer IHC analysis. Nonetheless, he says, his colleagues willingly explain to patients that they don’t think IHC adds much.

At John Wayne, Dr. Turner and his colleagues section the sentinel nodes no thicker than two mm, then prepare two histologic levels, separated by 200 microns with H&E and cytokeratin immunostain at each level. With standard H&E pathology, small metastases are variably detected, he explains; adding cytokeratin immunostains helps reduce that variation.

The risk, however, is that minimal findings or debris detected by IHC can be mistaken as a rare tumor cell. Dr. Turner says he’s encountered this problem not so much in his own practice as in the literature, where the notion of so-called benign transport became, for a time, a matter of much hand wringing among national experts. Now, he says, “I think people have gotten tired of debating that issue.” There’s just no scientific way of answering the question of whether a tumor cell was displaced by a prior needle biopsy and floated into the lymph node without actively metastasizing.

Dr. Turner helped author another paper published in Journal of Clinical Oncology (Hansen NM, et al. 2009;27:4679–4684), which looked at the survival impact of micrometastases and ITCs in the sentinel node of breast cancer patients. Among other findings, he and his colleagues observed a slight bias toward prescribing adjuvant therapy in patients with sentinel node involvement, even though there’s no consensus agreement as to the significance, if any, of sentinel node ITCs in breast cancer, he says.

The B-32 trial, along with the Z-10 study at Dr. Turner’s institution, could provide answers. Or, again, not. Most patients in this country who have breast cancer are treated based on the primary tumor findings, and nearly all patients receive some sort of adjuvant therapy, making it possible, and even likely, says Dr. Turner, that the studies will show no proven significance of ITCs. “But we won’t know whether that was an effect of the treatment intervention or a lack of significance of the finding in the first place.”

The aforementioned JCO study drives that point home. More than 90 percent of the patients who had ITCs or micrometastases received some form of adjuvant treatment (hormonal, chemotherapy, or both) as well as radiation. That’s why, Dr. Turner says, “Some of the older literature that has come out in support of the significance of micrometastases and isolated tumor cells might be, in the end, most meaningful as far as determining the significance.”

A more recent article, by a group from the Netherlands (de Boer M, et al. N Engl J Med. 2009;362:653–663), has raised interest—and a few eyebrows. The paper drew conclusions about adjuvant therapy, but Dr. Weaver says many are looking at it to draw conclusions about the importance of micrometastases. “When I took the paper apart in my own mind,” says Dr. Weaver, “there was a significant bias: Patients who recurred had large tumors and a higher grade than the group that didn’t have recurrences.” That suggests to him that factors other than micrometastases might be driving the recurrence risks; the micrometastases might instead be an epiphenomenon of the other factors.

But, he adds diplomatically, “I do think it’s a provocative paper. It helps us continue to think about the question.”

He finds the population-based analysis using the SEER database a more useful way to think about the prognostic significance of micrometastases (Chen SL, et al. Ann Surg Oncol. 2007;14:3378–3384). This showed that for small tumors, there was only a one percent decrement in survival. (The SEER database doesn’t separate ITCs and micrometastases, he notes.) “The study kind of crossed over between the pre-sentinel node era and the post-sentinel node era,” Dr. Weaver says. “So I thought it had at least some kind of reasonable expectation of being closer to reality than some of the other studies.” The results are exactly what one might expect, assuming nodal disease is a continuum of risk: The smallest amounts of metastases would have lower risks than increasingly larger sizes or volumes of metastases.

He expects the B-32 data—when it does emerge—to show similar results. Since the study is so big (4,000 patients), researchers should be able to demonstrate a statistically significant difference.

Whether that translates into clinical significance is another matter. Suppose they find a one to two percent difference in recurrence rate? “Nobody would treat based on that,” Dr. Weaver says.

A certain amount of battle fatigue appears to be setting in. “Stuff is so entrenched now,” says Dr. Weaver. One new study will not a new practice make. “If it doesn’t come out the way people individually expect, they’re going to say, ‘Ah. Well, it’s just one study.’”

With so much focus on minimal findings, prognosis, and survival, some physicians are missing the point when it comes to sentinel node biopsies, says Dr. Turner.

“Studies tend to focus on survival data,” he says. “I think the better question to be asking is, if there are minimal findings in the sentinel lymph node, which patients will benefit from having further axillary dissection to reduce the risk of axillary recurrence?” Naturally, no one knows the answer to this question, either. With few exceptions, the standard practice is for those patients to have further lymph node dissection, though it’s not clear who benefits. Remember that Z-11 trial that closed for lack of accrual? “Nearly all wanted more lymph nodes removed,” Dr. Turner says.

Nevertheless, physicians continue to debate whether removing the lymph nodes is purely a staging procedure, or whether it has therapeutic benefit. Put another way, is it safe to leave disease behind, in the axilla?

A big-picture, population-based look at this (courtesy of the B-32 trial) shows the false-negative rate for sentinel node biopsy is 9.6 percent. (This is not true for experienced surgeons, Dr. Weaver points out, a bit wearily. “The surgeons will say, ‘Oh, that’s totally wrong! My false-negative rate is four to five percent.’” The B-32 figure reflects the general rate in the United States.) He thinks it’s possible to get away with such a rate, given that radiation and chemotherapy can, and probably do, treat the nodes. “If leaving nodal disease behind 10 percent of the time was a bad thing, we should see lots of recurrences in the axilla,” which does not seem to be the case.

There’s another way to think about all this. Are primary tumor characteristics more important than the presence or absence of micrometastases?

“We’ve learned so much about the molecular phenotyping of breast cancers—ER, PR, HER2,” says Dr. Weaver. Those factors—or maybe some other as-yet-unknown factor—are probably going to be more relevant to fine-tuning treatment, he says. “I’ve heard a few of the more progressive medical oncologists say that under certain circumstances, they don’t even care about knowing the nodal status. It’s irrelevant to them.” Micro­metastases in the nodes may not be an indication that the tumor is bad; instead, says Dr. Weaver, they may indicate the tumor has taken up long-term residence in the patient, giving it a chance to get into the nodes.

The Chen paper, in Dr. Weaver’s opinion, “tells us a lot, because it’s population-based.” There was only a one percent difference in outcome, at both five years and 10 years, between patients with micrometastases and those with negative nodes. Micrometastases appeared to have more significance in patients with larger tumors. “Maybe micrometastases are a surrogate for identifying aggressive tumors,” Dr. Weaver says.

It’s well known that women who recur, who have large tumors, have aggressive cancers: triple (ER, PR, HER2) negative, or the so-called basal phenotype, and HER2-positive. “If we can really pick this apart, we might find that micrometastases are more common in those two tumor types,” Dr. Weaver says. Would that confer additional clinical value? “My guess is probably not,” says Dr. Weaver. “But I don’t have the data.” Once again.

In this particular drama, micrometastases may turn out to be mere groundlings. Even as the impressively large, highly regarded, eagerly anticipated NSABP trial draws to a close, its star might already be fading a little. “If you’d asked me at the 6th edition [of the AJCC staging manual] whether I ever thought nodal status was going to start becoming a less important factor, I probably would have defended nodal status as the most important prognostic factor available to us,” Dr. Weaver says. Bringing molecular definitions to breast cancer “totally blew that out of the water,” he cheerfully admits, adding, “That kicked the legs right out from underneath me.

“Now I’m the person who’s saying, ‘Gosh, maybe the status of the nodes is just not going to be that important in the future,’” he says—even as he prepares to dive into the data one more time.