Karen Titus
February 2023—Like a thriving expat, Deborah Dillon, MD, is comfortable moving within worlds both old and new. Specifically, as a breast and molecular pathologist at Brigham and Women’s Hospital, she appreciates the biomarkers she and her colleagues grew up with, so to speak, as well as those that are part of a more recently arrived-at scenery.
Not everyone finds both worlds equally riveting. “A lot of people are much more interested in, and excited by, new markers,” she says. “When I talk to people from pharma, this is what they want to hear about.” So do many pathologists, oncologists, and patients—new markers and new therapies have a way of updating hopes.
Dr. Dillon understands the persistent thrill of the new, why people want her to talk the language of PIK3CA, PARP inhibitors, MMR, NTRK fusions, ESR1, and the like. But as an in-demand speaker as well as in a recent interview with CAP TODAY, she also advocates for making the old—the longstanding trinity of ER, PR, and HER2—seem new again.
“We would probably make a larger difference in a larger number of patients if we just make some improvements in ER and HER2 testing,” says Dr. Dillon.
She’s seen that firsthand. One case that remains vivid in her mind involved a patient who was dying of metastatic disease. After reviewing the patient’s HER2 results, Dr. Dillon realized the patient had been treated with HER2-targeted therapy even though the cancer was not HER2 amplified. The patient was also ER negative. When the patient was subsequently switched to a therapy for triple-negative cancer, “She responded immediately.”
[dropcap]I[/dropcap]t’s essential for all pathologists to be conversant in both spheres, says Dr. Dillon, who covered classic and new predictive breast cancer biomarkers in two wide-ranging talks at an annual biomarker conference last fall.
Dr. Dillon says that in her consult work, she still sees room for improvement “just on the very basic things.” Immunohistochemistry for ER, PR, and HER2 remains “incredibly important in terms of directing therapy. And yet there are still challenges in those areas.” As with the British monarchy, longevity has not smoothed out all the rough bits.
Simply staying current with changing guidelines and cutpoints can have a profound impact on care, helping clinical colleagues navigate treatment decisions, Dr. Dillon says.
She notes, for example, that though the cutoff for a positive ER result has been one percent since 2010, updated ASCO/CAP guidelines from 2020 (Allison KH, et al. Arch Pathol Lab Med. 2020;144[5]:545–563) added a new category: low positive (one to 10 percent).
The new category should clear away some inconsistent practices, she says, noting that historically, some pathologists have called cases negative at the 10 percent threshold, while others reserve that designation for cases below one percent. “Which is actually the way it should be done.”
“That’s why this new category was formed,” Dr. Dillon says, adding that the category of cases that are classified as low-positive ER likely consists of those that truly have low-positive expression but also those with artifacts of staining, or overstaining.
[dropcap]I[/dropcap]n HER2 we have a similar situation,” Dr. Dillon says, “with changes over the years in the ASCO/CAP guidelines for what’s a positive HER2 value.”
The most recent guidelines (Wolff AC, et al. Arch Pathol Lab Med. 2018; 142[11]:1364–1382) reclassified several of the less common FISH result types. Most cases—some 90 percent—fall into group one (clearly positive) or group five (clearly negative). Groups two through four reflect equivocal and nonclassical groups, or, in her words, “weird combinations of results in the middle.” With more outcomes data, these classifications will likely be further refined, she says. But for now, the guidelines are helpful in directing pathologists to parse these cases, in part because they also incorporate IHC more thoroughly. Any time FISH is done—as the guidelines make clear—it must be interpreted in the context of IHC.
Indeed, an unwritten but pragmatic category, Dr. Dillon says, might be, Make sure you’re careful not to overcall a HER2-positive result on IHC alone. “You can always get confirmation by ISH if there is any doubt in your mind,” she says.
One of the biggest errors she and her colleagues see in consults is overcalling the HER2 immunohistochemistry. HER2 IHC 3+ cases aren’t routinely confirmed by FISH, and the patients may be inappropriately treated as if the tumor is HER2 positive even when it is FISH negative.
The message here is simple: Go FISH. “Really, if you have any hesitation at all, get the in situ hybridization done,” says Dr. Dillon. “We still have more false-positives than we should.”
Another new category, called HER2 low, emerged from Destiny-Breast04, a phase three clinical trial looking at new antibody drug conjugates (Modi S, et al. N Engl J Med. 2022;387[1]:9–20). In these cases, distinguishing between 0 and 1+ HER2 IHC appears to be important—something current assays aren’t designed to do. It’s also possible that the antibody drug conjugates are effective in all cancers, including HER2 0. “That would be great,” Dr. Dillon says, “but until we know more, we will probably need new HER2 tests to make these distinctions.” (For a more in-depth discussion of these issues, see “Breast cancer breakthrough sparks HER2 quest,” CAP TODAY, June 2022.)
[dropcap]P[/dropcap]athologists also need to develop fluency in molecular testing, even when they don’t perform the tests in-house. “You have to be able to speak up in tumor board and know whether to suggest sequencing or a multigene predictor on a case,” Dr. Dillon says, noting, for example, that most oncology guidelines support the use of a multigene assay to guide chemotherapy and/or extended hormone replacement therapy in ER-positive, HER2-negative cancers.
Again, this is not necessarily new ground. “We’ve been sending material out for the multigene recurrence predictors for many years now.” But she gets asked about this frequently; specifically, colleagues want to know, What is the role of pathologists for molecular send-out tests?
Her answer is straightforward. First, make sure clinicians know about testing for newer markers. Then, understand what kind of tissue is needed for performing a certain test.
She continues: “That’s extremely important—to choose the area of the tissue with the highest concentration of invasive cancer” and excluding DCIS and other noninvasive proliferations “as best they can.”
It’s also helpful, she says, for a group’s pathologists to mark the best block for future molecular testing in their notes, or as part of the pathology report. “I have seen a number of places doing this,” she says. That means pathologists won’t have to review the whole case when someone requests molecular testing for multigene recurrence predictors or sequencing. “It’s really helpful. Because otherwise somebody has to pull the entire case and figure out which block to send”—extra work that no one seeks.
The accuracy of these send-out molecular tests depends on the tissue being used, she says. These generally are not diagnostic tests; results are only valid within the context of an invasive cancer. “But I’ve seen cases where benign things have been sent accidentally, and they’ll give you an answer.” One consult involved a benign complex sclerosing lesion, for example, that had been diagnosed as invasive carcinoma and sent for multigene recurrence predictor testing. The result? The lesion was given a 17 percent risk of distant recurrence.
The tests do not localize a molecular finding in the context of the morphology. Rather, they measure the biomolecules within the entire tissue content, which can include tumor cells, stroma, inflammatory cells, and normal epithelium.
Problems typically arise for certain cancers, including low cellularity (especially with lobular carcinomas), some sort of prominent lymphocytic infiltrate, a DCIS of another subtype (such as HER2 positive with IDC negative), or a large core needle biopsy site relative to tumor size. “The results of a molecular test may not reflect the invasive cancer if you don’t have high enough cellularity of your invasive cancer,” she says.
When the results of a send-out test come back, pathologists need to ask whether they make sense. “This is something that we started doing,” Dr. Dillon says.
She elaborates: “If you send out a low-grade lobular cancer, and you knew from your own immunohistochemistry that it was ER positive, and that the mitotic rate was low, results from a recurrence predictor with a high proliferation rate simply don’t make sense.” That’s worth a second look, she says. “Maybe the proliferation of the inflammatory cells in the biopsy site drove up the proliferation rate in the recurrence assay.”
Other red flags include a low-grade invasive mucinous cancer or a low- to intermediate-grade lobular cancer being identified as high risk.
Sometimes she and her colleagues will see lobular cancers that are clearly ER and PR positive by IHC, “but when we get the results of the recurrence predictor, they have numbers for ER or PR that are very low. If you didn’t see that in your immunohistochemistry, you should know there’s some discordance.” The algorithms for these predictor assays will drive up the risk score if they’re unable to detect the ER or PR signal because of low tumor cellularity, she adds.
Understanding these issues is critical, she says. “It’s important for community pathologists to know that a lot of these molecular tests are things where you scrape tissue off the slide, and you’re really taking everything that’s there. So you can lose the signal of your invasive cancer cells if you’re not paying attention to what went into the assay.”
Don’t assume all the testing pieces will fit together automatically. In fact, she says, with Oncotype DX, it was discovered early on that if the sample includes the biopsy site in the scraped tissue, “it can drive up the proliferation markers in the assay due to contamination from proliferating inflammatory cells,” resulting in an incorrect “high-risk” designation.
From time to time, she says, she’ll get a call from an oncologist confused by the discordant results—who is told by a recurrence gene predictor assay that the cancer is high risk, though the pathology report characterized the tumor as low grade and ER positive. Pathologists need to make sure the information makes sense as a whole, including what they know from the histologic review and the ER/PR/HER2 assays.
In some cases, the recurrence predictor assay’s limitations may be coming to the fore. But in other cases it’s possible the histology is in error. “The thing to do at this point is, obviously, go back and look at the slides again, and ask, for instance, Did I make a mistake on the grading?” Checking the tissue sample is important. “And if necessary, double-check with the outside lab that the correct block was tested, that the labeling was correct.”
In short, she says, certain cancers should not be called high-risk cancers. “And if they are, then it’s most likely somebody made a mistake somewhere.”
The mistakes can be devastating if not caught in time. Recalling the aforementioned case of the patient wrongly given HER2-targeted therapy for a TNBC, she says she’s occasionally heard pathologists say they “pushed” the HER2 results upward a bit to (as Dr. Dillon puts it) give the patient a chance.
Dr. Dillon’s response to this sleight of hand: “Don’t give her a chance; treat her with the right therapy.”
[dropcap]F[/dropcap]or those who are diagnosed with metastatic disease, new markers and approaches are bringing new hope to the conversation. For pathologists and oncologists, they’re also easier to talk about conceptually, Dr. Dillon suggests. The biggest question is easy to frame: Whom do you treat?
The last five years have seen a number of new FDA-approved targeted therapies, she says, though the word still needs to be spread about their availability, as does emphasizing their use for advanced tumors.
BRCA1 and BRCA2 are already well known, given the role germline mutations in these genes play in hereditary breast cancer. They’re also now being used to predict response to therapy with PARP inhibitors, Dr. Dillon says, based on results of the OlympiAD trial (Robson M, et al. N Engl J Med. 2017;377[6]:523–533) that showed improved survival (median progression-free survival of seven months versus 4.2 months) in patients with germline BRCA1 or 2 mutations and HER2-negative advanced breast cancer treated with olaparib versus standard chemotherapy. Beyond germline mutations, there’s interest in whether somatic BRCA1 and 2 mutations—which are present in about three percent of all breast cancers—will predict response to PARP inhibition to breast cancer (as they do in ovarian cancer).
PIK3CA is also emerging as an important marker, Dr. Dillon says. In 2019 the FDA approved the PI3K inhibitor alpelisib in patients with the gene mutation; the SOLAR1 trial showed that in patients with advanced ER-positive/HER2-negative breast cancer, median progression-free survival nearly doubled compared with endocrine therapy alone (André F, et al. N Engl J Med. 2019;380[20]:1929–1940). Some patients don’t tolerate the drug very well, though, and Dr. Dillon predicts new PI3K inhibitors are likely to become available. Molecular testing for the trial used a real-time PCR assay for 11 common mutations, though most academic sites now use next-generation sequencing, which, Dr. Dillon notes, reveals other, less common mutations with unclear clinical significance.
In recent years the FDA has also approved several new therapies based on so-called tissue-agnostic markers, which can be present in a variety of advanced solid tumors and will be relevant to some breast cancers, says Dr. Dillon. “These include markers predicting response to immune checkpoint inhibition,” she says, such as defective MMR (mismatch repair deficiency), present in about two percent of breast cancers; TMB-H (high tumor mutational burden), present in about five percent of breast cancers (but almost 20 percent of cases of metastatic triple-negative breast cancer); and PD-L1.
There’s also “the whole world of very, very rare alterations,” such as NTRK fusions, which are present in less than one percent of cases.
Given their newness, one of the challenges is that oncologists and pathologists simply need to know that these therapies exist, Dr. Dillon says. The literature has shown that many patients with advanced tumors aren’t even tested for these targeted therapies, she says. “They just get standard chemotherapy for metastatic disease.” The reasons vary. Educational efforts always take time, and these recent approvals may have bumped up against the COVID era.
While these cases are rare, if a patient is getting an NGS panel, they’ll likely be tested for most of these markers. “The message here is, get panel sequencing for any advanced cancer,” Dr. Dillon says. “Because there are so many of these tissue-agnostic approvals now.” Of PD-L1, she adds, “Personally, I’m hoping we don’t have to rely on PD-L1 for much longer.”
The field is fast-moving, and Dr. Dillon is keeping her eye on markers likely to enter routine clinical use. The most recent targeted therapy in breast cancer is elacestrant for advanced ER-positive, HER2-negative tumors with ESR1 (estrogen receptor) mutations in the ligand binding domain. The Food and Drug Administration approved it in late January. ESR1 mutations are present in about a third of all metastatic ER-positive cancers but, interestingly, are not present in most primary tumors. “We think they’re acquired by patients being treated with endocrine therapy,” Dr. Dillon says. While molecular testing for ESR1 up to this point has been only for patients on clinical trials, it will now be standard of care for all patients with advanced ER-positive, HER2-negative disease, says Dr. Dillon.
HER2 mutations are in researchers’ sight lines. They’re particularly present in high-grade lobular cancers. “That’s interesting, because it’s usually in tumors that don’t have HER2 amplification,” she says. It gives the cells an alternative mechanism to activate HER2 in these tumors, which show less sensitivity to the traditional HER2 targeting agents but are sensitive to other, different HER2 targeted agents, including neratinib. Additional drug development in this area should result in more drugs with fewer side effects.
She also sees a future for serial monitoring, taking pathology beyond the realm of diagnosis. Breast cancers in general tend to be less aggressive than, say, lung cancers, “and I think there’s a lot of opportunity for monitoring, for the detection of resistance mutations and other things that will trigger changes in therapy.”
[dropcap]I[/dropcap]n breast cancer, the old and the new remain entwined, like fog and a vineyard. Unlike many other cancers, where advances have largely been molecularly driven, in breast, “We also have these other markers that have been around so long,” Dr. Dillon says.
For the most part—at least currently—newer markers are aimed at advanced cancers. “That’s why I continue to talk about ER and HER2,” she says. “Because otherwise you’re only talking about patients who’ve recurred or metastasized.”
And if you miss out early on with ER/PR/HER2, you’ve increased the chances of recurrence. “Absolutely,” Dr. Dillon says. “Because then you’re giving the patient the incorrect therapy.” Though not a precise analogy to the ounce of prevention/pound of cure saying, appropriate early treatment is more likely to avert a desperate search for later cures. “I’m absolutely convinced this is true,” she says. “If you get that right, upfront, you’ll have fewer patients who’ll eventually get metastatic disease.”
Karen Titus is CAP TODAY contributing editor and co-managing editor.