Home >> ALL ISSUES >> 2015 Issues >> Genetic profiling vies with IHC in retune of CUP testing

Genetic profiling vies with IHC in retune of CUP testing

image_pdfCreate PDF

Anne Paxton

March 2015—Tesla beats Camry. Online catalogs replace paper. Keurig edges out Chemex. Mobile trounces landline. When paradigms shift, the theory goes, we can only cling to the technology in the outbox for just so long.

But that’s a theory that may not apply to diagnostic testing for cancer of unknown primary (CUP). Microarray-based gene expression profiling (GEP) has recently gained a foothold in the quest to identify origins and therapeutic targets for metastatic cancer, but traditional immunohistochemistry is not about to be sidelined. In research and in the field, oncologists and pathologists are still weighing whether GEP has advantages over IHC in diagnosing CUP, or if IHC should continue to retain its central role with GEP serving as an adjunct.

Relative cost, precision, timeliness, and efficiency are among the competing concerns. And they reveal as much about pathologists’ changing role in diagnosis, and relationships between pathology and oncology, as they do about scientific advances in diagnosing and treating cancer.

Between three percent and five percent of metastatic cancers are CUPs, generally poorly differentiated tumors that present a diagnostic challenge and often a dismal prognosis for the patient.

“Tumors of unknown primary are a thorn in your daily diagnostic practice. It’s a frustrating clinical problem that we all, pathologists and oncologists, face on a fairly routine basis,” says Lynette Sholl, MD, assistant professor of pathology at Harvard Medical School and a pathologist at Brigham and Women’s Hospital in Boston. “You’re kind of treating a black box, and the usual regimens just don’t apply, whether you’re approaching it from a diagnostic or a therapeutic standpoint.”

Immunostains have been the primary diagnostic tool for CUPs for decades. Gene expression profiling came on the scene about five or six years ago, with manufacturers offering to diagnose CUP at a cost of $3,000 and up. The commercial landscape now includes the ResponseDX Tissue of Origin Test (developed by Pathwork which declared bankruptcy in 2013 and was acquired by Response Genetics), Cancer of Origin Test made by Rosetta Genomics, and CancerType­ID made by bioTheranostics.

Dr. Handorf

Dr. Handorf

GEP has had a somewhat hard time catching on, says Charles R. Handorf, MD, PhD, vice chair of the National Comprehensive Cancer Network unknown primary panel and a professor of pathology at the University of Tennessee College of Medicine. Standard-setting organizations like NCCN and the National Cancer Institute have not begun to call for use of GEP in diagnosing CUP. But the dramatic growth of new targeted therapies, some of them for previously hopeless cancers, is intensifying interest in GEP as a tool for finding the targets. “As we get more targeted therapies,” Dr. Handorf notes, “it moves us to do a better job of learning the biological behavior and sensitivity of a particular cancer, and not just simply writing it off because it’s poorly differentiated or in an advanced clinical stage.”

Dr. Handorf, who was first author on a 2013 study comparing GEP and IHC for the identification of the primary site in metastatic tumors (Am J Surg Pathol. 2013;37[7]:1067–1075), believes that early on, the commercial companies making expression assays mistakenly took a combative view of the pathology world. “Instead of trying to work with pathologists on the best use of the assays, they came out, guns blazing, claiming that pathologists do IHC only because they make money doing it.”

“So all these moving parts are going on at the same time. We wanted to see if there was a meeting ground on the interface between IHC and GEP, where maybe one leaves off and the other picks up. That was the focus of our study.”

The multicenter study was prospective. He and colleagues took known primary cancers, presented them to study participants as unknown, then asked pathologists to proceed through their normal rounds of immunostains. “Across the board, with both difficult and easy cases, metastases of all kinds, the pathologists got the right tissue of origin just on the H&E evaluation about 70 percent of the time. They would then get it up to 80 percent after a round of immuno­stains; further rounds of IHC didn’t improve on that. GEP was able to assign the correct origin about 90 percent of the time.”

One major conclusion from the study was that “after one round of stains you are much less likely to get a definitive answer, so stop and consider using GEP to get an answer,” Dr. Handorf says. That would avoid spending several thousand dollars on a GEP in many cases.

Perhaps more important, he emphasizes: It’s a mistake to consider finding what organ a cancer arises from as the central goal. “We pathologists are trained that way, because there are different treatment protocols directed against the organ if it’s ovarian, breast, or stomach cancer. But we have to realize when we have targeted therapy, it’s therapy against some druggable target, a mutation in the gene sequence of the tumor. So where the tumor comes from is not going to be nearly as important as the mutation that you know the drugs work against.”

In five percent to 10 percent of CUP cases, after all the best efforts, “we still won’t know the primary organ.” But even in some of those cases, “if you can find that druggable target in the genes, you could cure the cancer—it can happen.” This is in contrast to 10 years ago, he adds, when the treatment options were few and the patient would likely survive only a few months.

It’s important for pathologists and the diagnostics industry to work together on CUPs, Dr. Handorf stresses. “The biggest mistake as these tests were being developed—and that’s not pathologists’ fault—is the companies were working at odds with pathologists. It was industry’s miscalculation of how the world really works.”

Cancer of unknown primary could be seen as the epitome of personalized medicine, says Gauri R. Varadhachary, MD, professor of GI medical oncology at the University of Texas MD Anderson Cancer Center. “There is no one designated recipe for first-line/second-line treatment with an unknown primary. Given that the patients’ cancers are very different from each other, there is no primary tumor, and they’re all stage IV presentations,” each case is unique, she says.

Dr. Varadhachary

Dr. Varadhachary

The lack of a systematic, tiered approach to conducting IHC is a chronic issue in pathology, in her view. “The stains today do tend to be a little all over the place,” says Dr. Varadhachary, who has been studying CUP for the past 15 years. “The pathologists are very removed from the clinician who is seeing the patient, and they get a small piece of tissue without any history or additional data. So some do too many stains, some don’t do enough, some don’t do the right ones. There isn’t any clarity on how it should be uniform, but it would benefit all of us to really understand how we can use the least number of stains to get the most out of it.”

Exhausting the tissue is only one of the dangers. “The more complicated the diagnosis, the more undifferentiated the tumor, the more stains we seem to do. But we really don’t learn more from that.”

Wayne J. Lennington, MD, coauthor of another study finding that molecular profiling complements IHC for CUP (J Natl Cancer Inst. 2013;105 [11]:782–790), agrees. “That’s one of the things that was really apparent from our study, because we saw material from a large geographic area, many different practices. People are all over the board with what they order and what they don’t. If you don’t have a refined, almost standardized approach to how you deal with these things, you’re going to miss some tumors that you could identify,” Dr. Lennington says.

Bigger and more academically oriented practices tend to be more uniform. “Certainly some practices don’t see many tumors that fall into this category, and smaller practices may not have the kind of immuno capability that larger ones do, so they tend to be more rudimentary in their ordering practices, which makes it tougher to identify less well differentiated tumors,” he says.

When Dr. Lennington was in training, “we would order 50 immunostains on a case, just to have it or to see how it worked. You can’t do that anymore. You have to think about what you’re going to order, to try to tailor it so that you have tissue you can still use and you don’t end up having the patient re-biopsied just to get tissue for molecular profiling.”

Dr. Lennington, who practices with Associated Pathologists (PathGroup) in Nashville, says one of the goals of his study “was to show pathologists this [molecular profiling] is a tool you can use, but it’s certainly not the be-all and end-all. We’ve had some cases where molecular profiling was really, really helpful, and we’ve probably had an equal number of cases where it was virtually no help—in particular for sarcomatoid carcinoma or carcinomas with metaplastic changes. I generally try to stay away from GEP on those kinds of carcinomas because the tumor profiles tend to segregate them into sarcomas, and morphologically that’s clearly not what they are.”

In Dr. Varadhachary’s own practice, she says, “I look at the IHC, I look at the patient’s overall presentation including risk factors and imaging, and if there is significant room for doubt about the working diagnosis, and I believe the patient is going to have several options of therapies, I integrate tissue-of-origin testing in that patient. It’s not cheap and it’s not for each patient who walks through the door.”

Her previous research focused primarily on abdominal CUP cancer presentations. “If you take all the patients who present with CUP to my clinic, which is based in the Department of Gastrointestinal Medical Oncology, more than half of them have a cancer process involving either liver and/or the lining of the bowels.” Based on IHC—and in some patients, molecular assays—patients diagnosed with an intestinal profile now have many more therapy opportunities than 10 years ago, she says. “We use colon-cancer–like drugs and they tend to have a more favorable outcome.”

The premise of gene expression profiling is that metastatic tumors have molecular patterns that match their primary origin. “That’s not proven, but it is extrapolated from other cancers,” Dr. Varadhachary says. “If you look at nodes from breast cancer patients, their genetic signature will resemble the primary. So because metastatic cancers do tend to retain some part of the primary signature, one can probably apply the same to CUP.”

“There’s sometimes a discordancy that we see as CUP clinicians that makes CUP diagnosis both intriguing and confusing,” she says. “For example, you might have a patient, chronic smoker, with a chest mass that looks like lung cancer on imaging, but then you do IHC and that is non-specific and negative for lung cancer stains, and a genetic molecular profiling test suggests an intestinal cancer. That’s where the discordancy comes in. The clinical presentation fits lung cancer, endoscopy and colon­oscopy do not reveal a primary, and additional tools tell you it’s something else. Which one do you believe? We are sometimes at such a crossroads with these presentations.”

Fifteen or 20 years ago, it didn’t matter that much, because there were limited options, and everyone got the same treatment, Dr. Varadhachary adds. “But today we have more and more subclassifications of cancer and the therapies can become more specific.”
She believes there has to be greater integration between oncologists and pathologists. “I had never heard of going straight to GEP in practice, but for several of the patients who come to see me, the oncologist has already ordered the gene profiling test.” Pathologists may need to adapt to this changed environment. “If pathologists believe they are doing too many IHCs to get some level of confidence in placing the lineage of the cancer, they can stop early, do additional tests like mutational profiling or tissue-of-origin profiling, then come back to perform additional IHCs later.”

The clinical utility of GEP in increasing life expectancy of patients, in her view, remains to be seen. “Also, there tends to be some hype in believing that next-generation sequencing is clearly the wave of the future. That is absolutely a wave we want to ride, but the question remains exactly what impact it’s going to have in the quality and duration of life of our patients. On that point we are still in the trial design phase and will have to wait for the outcome phase to have some answers. How the therapies are going to be different is what we’re teasing out right now.”

It may be that the putative primary site isn’t critical to treatment. “The unknown primary situation may be unique in a way, but not unique enough that you need to make the distinction,” Dr. Varadhachary says. Her own current research on CUP, still in the planning stages, will be less treatment based and more translational biology based. “We want to leapfrog into looking at circulating DNA in the blood to see if that gives us better information than the tumor tissue to determine the tissue origin and the mutational profile, and to inform therapies.” This “liquid biopsy platform” for unknown primary would be much easier for patients and allow more real-time testing, but it’s in its infancy, she adds.

Too frequently, pathologists adopt the stance that “here’s a tumor of unknown primary, let’s start with a large panel of IHC markers,” says Fan Lin, MD, PhD, director of anatomic pathology at Geisinger Medical Center, Danville, Pa.

CAP TODAY
X