Molecular tumor boards: fixture or fad?

“It’s really nice, from the laboratory side, to hear when there are new options for patients based on the data we’re providing,” Dr. Hill continues. Case in point: Dr. Hill recalls the surprise he felt at a recent molecular tumor board when a medical oncologist colleague responded to a report that a patient had tested positive for a lung cancer resistance mutation. To Dr. Hill, the news seemed bad: The patient was failing therapy, and here was the proof. “But my colleague was very happy.” Not only had his suspicions been confirmed, but the patient now qualified for a new drug specifically tailored to this situation: patients with the mutation who were failing therapy. “Now we are very careful to communicate that kind of information as rapidly as we can, because we’ve seen how they may use it to qualify patients for a different therapy.”

He’s also noticed that while presenters at the molecular tumor board come bearing similar research, the face-to-face discussions yield slightly different responses. “Everybody’s looking at the literature, but with a slightly different point of view. It’s actually a bias, but we’ll call it a point of view,” he says with a laugh. “So the board gives us a well-rounded picture of the new information.”

Dr. Hill says the molecular tumor board, which meets monthly (that may change to weekly at some point) to discuss three or four cases, on average, has been “incredibly valuable. It gives me an opportunity to be more up-to-date on how information coming out of my laboratory is being used.” Like many of the tumor boards, the format presents clinical findings (in this case, from a hematology-oncology fellow) followed by pathology and molecular reports.

Dr. Hill doesn’t see either traditional or molecular tumor boards disappearing. Like many others, he expects more molecular findings to wind up in traditional discussions. “But unless we start significantly reducing the number of patients who present in late-stage disease, I don’t think we’ll see the molecular tumor board go away.” Even as the board continues to cope with the high-tech challenge of molecular testing, one low-tech challenge persists, he says. Doctors are busy; time is fleeting. “But we find it valuable enough that we make time to do this.”

So important is the molecular tumor board at MD Anderson that even the institution’s president, Ronald DePinho, MD, finds time to periodically attend, says Dr. Alexander Lazar. “This is something that he’s particularly interested in and supportive of.”

Some 150 clinical cases are sequenced each week using multiple next-gen panels of approximately 30 to 400 genes; the board meets monthly for 90 minutes. There’s no need here for even a quick calculation. “Clearly all of those aren’t going to be discussed at this conference,” Dr. Lazar says. Instead, he and his colleagues try to pick several particularly thought-provoking cases—a novel mutation, intriguing mutation patterns, comparisons between a primary and metastatic tumor. The conference also is a place to discuss changes to the panels. Some 25 to 50 people attend the tumor board. In the near future, these discussions will likely expand to sister institutions within the MD Anderson network through videoconferencing.

In a typical discussion, a clinician presents the clinical features of a case, followed by someone else presenting a literature review of the tumor’s genetic features. A pathologist might fill in with additional details about allele frequency or pitfalls in interpretation, for example. Then comes discussion and an attempt to synthesize the information, all with an eye toward dealing with similar cases in the future.

Dr. Lazar uses the example of BRAF to describe this trickle-down approach. It’s commonly mutated in melanoma, but perhaps sequencing has turned up a variant that’s not well described in the literature. “Someone might discuss everything that’s known about where BRAF is mutated, including the areas that are really common, such as the V600 region in exon 15,” Dr. Lazar says. “Then we’d discuss what we know about alternative mutations in exons 15 or 11. Are they good activators of the BRAF protein and ERK pathway, are they not good activators of the protein, do they seem like they are potential driver mutations, or could they be passenger mutations? From there, we’d narrow it down to looking at the particular mutation in the case we have to discuss. If there’s not a lot in the literature, what would we predict about the case based on all the information we just reviewed? And that would lead to a discussion of which of the mutations tend to respond well to the family of BRAF or MEK inhibitors, and which ones seem not to. And then what is going on in clinical trials?”

Not all molecular results require that level of discussion. More routine testing, with straightforward results, easily makes its way into traditional tumor board discussions, as it has for years with microsatellite instability testing in colon, Dr. Lazar says.

The Institute for Personalized Cancer Therapy has become another piece of figuring out how to annotate and interpret next-generation sequencing, he says. Right now, by agreement with the clinicians at MD Anderson, molecular reports from pathology do not include information or suggestions about clinical trials that might be useful based on testing data. The institute group—including molecular pathologists, clinical teams, bioinformaticists, and researchers—is creating a continually updated guide for interpreting molecular test results. As the database goes live, one gene after the next, attendees at the molecular tumor board are watching closely. “It’s somewhat of a proving ground to discuss what’s going into the database.”

The molecular tumor board has also become a place to discuss so-called bucket trials, in which patients are treated based on mutations as well as histology. Patients with a similar mutation, regardless of tumor site, might be treated with the same therapeutic agent that inhibits a certain pathway. “We’re trying to find ways to identify patients with certain types of mutations, and then quickly try these different targeted therapies to see what histologic context they’re going to work in, and what ones they’re not,” he says.

Molecular tumor boards may also serve as a sort of refuge for physicians, Dr. Lazar suggests. “We’re basically making new rules for how to manage information and apply it clinically. And that’s incredibly exciting. But that can also create uncertainty. So by discussing it from every aspect—from the tests we want to do, to interpretation, to the limits of particular test technologies, to clinical trials, to treatment response—we can create a system of best practices that makes sense for everybody.”

Given the nature of what they’re trying to accomplish, it only makes sense that molecular tumor boards would evolve. In a field that changes this rapidly, a static setup would be as valuable as tracking websites with a card catalog.

When the molecular tumor board was first launched at MD Anderson, for example, the discussions were much more formal and didactic than they are now. They had to be, Dr. Lazar says. “We were explaining how sequencing works, and the quirks of different platforms, and why some tests were good for certain events but not others.”

The first change for many institutions, however, will be setting up a molecular tumor board. Even Dr. Hensing, with all his satisfaction with the current traditional tumor board, sees that day approaching for NorthShore. It could come when next-generation sequencing comes onboard, bringing with it those pesky variants of unknown clinical significance. “As clinicians, we don’t necessarily know best how to deal with that information. There needs to be a mechanism to discuss these cases.” At the same time, he says, those mechanisms should start within disease-specific tumor boards, “because it’s got to start with how to handle the diagnostic material.”

When should an institution consider adding a molecular tumor board? Not everyone has the resources of an MD Anderson. As Dr. Lazar puts it, “Our breast medical oncology department is larger than most people’s entire cancer centers.”

That’s almost beside the point, though. Nearly everyone makes it abundantly clear, with the repetitiveness of a Schubert symphony, that molecular results are best handled in a multidisciplinary way. “Everybody can’t be good at everything,” says Dr. Lazar. So regardless of size, every institution needs pathologists who are very familiar with molecular testing—how it’s done, how it’s interpreted, and what it means. “They have to present a face to the treating clinical team.” In addition, he says, pathologists need to be present to listen to what their clinical colleagues need from molecular tests and communicate what is possible on a practical level.

And again, if it’s not already clear, this whole field—“personalized cancer therapy, targeted therapy, molecular medicine, precision medicine, whatever you want to call it,” says Dr. Lazar—starts with results from tests that are performed by pathologists, on tissue that pathologists diagnose, curate, and maintain. “So our ability as pathologists to explain these tests, their importance and limitations and clinical justification, is absolutely critical for patient care.”

[hr]

[typography font=”Cantarell” size=”18″ size_format=”px”]Next month: Molecular tumor boards at Dana-Farber Cancer Institute and Brigham and Women’s Hospital[/typography]

[hr]

This is true even if the molecular testing is sent out. It may not make sense for an institution to offer the testing locally, Dr. Lazar says, but “you absolutely have to have someone in your practice—frankly, as many people as you possibly can—who will get interested in this, even if they’re not producing the information in their own department. Personally, I think every pathologist should be interested in this and be able to explain it.” Pathologists need to oversee the preanalytic variables, but also figure out how to incorporate it into the medical record and how to help clinicians figure out how to use it in patient care. “Treating clinicians will be much happier trying to discuss this with a local pathologist who’s an expert, rather than trying to get answers from a reference lab across the country,” Dr. Lazar says.

And for those who already have a molecular tumor board or two, Dr. Heusel shares his thoughts about how he’d like to see things evolve.

First, he says, there needs to be standardization in how variants are reported and in the criteria by which they are considered actionable. “For a long time, people were counting the variants they were finding in p53 as being actionable. Well, it’s hard to make that claim at this point.” Similarly, he says, BRCA1 and BRCA2 have “many, many VUS, and very few of them are probably actionable.”

He’d also like to see molecular tumor boards pay closer attention to data-set integration, making sure that cytogenetic microarrays are coupled with the genome and the transcriptome, and possibly the methylome, in a way that meets the standard of patient care. “So that we’re not all rushing around and chasing the next best thing before we’ve even decided whether it meets clinical standards.”

Finally, he asks, what’s the best way to improve understanding of gene-gene-gene interactions and the intricacies of pathway interactions? When do passenger variants become modulators, and when do modulators become drivers? “We are still operating in 2D space, relatively speaking, mapping variants in single genes to diseases or phenotypes.” But pathology is a 3D, possibly even a 4D, endeavor, he argues.

Molecular testing is only one bud on the branch, and for all its complexities, says UCSD’s Dr. Hansel, “Right now it’s actually easy to do. The material is pretty stable, and it’s not extraordinarily expensive to do.” As tests delve deeper, however, it might become harder to make clinical connections. Correlations between genome alterations and mRNA alterations tend to be fairly linear, but at the protein level, that correlation starts to break down. “You could have genomic alterations that don’t necessarily predict what happens on the protein level, but it’s the protein level you’re targeting,” Dr. Hansel says. “So not all genomic alterations carry through to what the targeted therapy is predicted to be.” In that sense, “We still have a lot of work to do.”

Dr. Hansel is unwavering in one final hope. No matter how molecular tumor boards evolve, she says, pathologists need a seat at that table. “Because change is going to happen with or without us.” [hr]

Karen Titus is CAP TODAY contributing editor and co-managing editor.