Advances in Precision Medicine Propel Personalized GI Cancer Care

Eileen O’Reilly, MD

Consistent and accurate germline and somatic testing practices help determine the most effective targeted therapies for patients with gastrointestinal (GI) cancers, although advances in this field hinge on multidisciplinary collaboration and the widespread application of both tissue and liquid biopsies to supplement germline testing results, according to Eileen O’Reilly, MD, and Benjamin Weinberg, MD.

“The topic of biomarker testing in GI malignancies is important, as we’re now in the era of precision medicine and beginning to refine treatment options for patients based on characteristics beyond clinical parameters,” O’Reilly said in an interview with OncLive^®.

Benjamin Weinberg, MD

“It’s incumbent upon us as the treating oncologists to know what’s driving a patient’s tumor and what targetable options are out there,” Weinberg added in the interview.

In the interview, O’Reilly and Weinberg discussed the importance of conducting biomarker testing as the GI cancer treatment arena continues to evolve, the need for increased gene fusion testing, and the negative effects of inadequate biomarker testing in the GI cancer population. 

O’Reilly is the Winthrop Rockefeller Endowed Chair of Medical Oncology, co-director of medical Initiatives at the David M. Rubenstein Center for Pancreatic Cancer Research, and section head of Hepatopancreaticobilary and Neuroendocrine Cancers at Memorial Sloan Kettering Cancer Center (MSK) in New York, New York.

Weinberg is an associate professor of medicine in the Division of Hematology and Oncology at the Lombardi Comprehensive Cancer Center at Georgetown University in Washington, DC.

OncLive: How has the emergence of different biomarkers increased the importance of biomarker testing in GI cancers?

O’Reilly: For almost all GI malignancies, we want to make sure a patient has germline testing looking for actionable alterations in the germline, [which may have] implications for the family. In the tumor, [we are] looking at tissue-based testing, which is primarily DNA-based testing that is increasingly complemented by liquid biopsies and circulating tumor DNA testing, which, in some GI malignancies, may have value when we have challenges obtaining tissue.

Weinberg: We are now in the era of precision medicine and precision oncology. We’re jealous of our lung cancer colleagues on the immunotherapy and targeted therapy sides because they have many actionable, targetable alterations and a slew of drugs that are the right drugs for the right patients at the right time.

[Testing is] complicated in GI cancer because we treat [patients with] several different malignancies. Upper GI cancer, esophageal cancer, and gastric cancer are different than pancreatic cancer, which is different than colorectal cancer [CRC]. As much as we’d like to have a molecular subtype to unify these tumor types, they are different cancers. That drives the type of testing that is standard, guideline-approved, and actionable.

In the upper GI space, for example, we care about HER2 positivity and PD-L1 positivity, and maybe claudin 18.2 [CLDN18.2], with some newer drugs under development. In CRC, we’re mostly still driven [by markers] like KRAS, NRAS, BRAF, HER2, and rare [mutations] like NTRK. However, there are some unifying features like microsatellite instability [MSI] high [and] mismatch repair deficiency [dMMR]. That [applies] across tumor types, not just GI cancers, and has changed the paradigm [because we now give those patients] immunotherapy. Biomarker testing is still somewhat dependent on the disease site of origin, at least in 2023.

O’Reilly: Think about [testing] early in GI malignancies. We have many challenges, sometimes, regarding tissue. It can require considering an additional biopsy. Previously, that wasn’t considered appropriate, but that is changing, and it needs to change because some of these biomarkers can require a second and occasionally even a third attempt. That’s not justifiable all the time, but in certain populations that may be enriched for actionable findings, this may be a key consideration.

Weinberg: Generally, if there’s available tissue, we’ll send broad-based molecular sequencing, usually DNA- and RNA-based next-generation sequencing [NGS], for any patient with an advanced upper GI cancer at diagnosis, because some [mutations] are actionable down the road. For example, in BRAF V600E–mutant CRC, we have an approved option in the second- and third-line settings that might affect how I treat a patient in the frontline setting. A clinical trial might use BRAF-targeted therapy in the frontline setting.

There are many reasons to test up front. In this era of liquid biopsies, patients who have had good up-front tissue-based sequencing may then go to more liquid-based sequencing after exposure to targeted therapies to look for resistance mechanisms and novel therapeutic targets.

Why should medical oncologists be concerned about testing for pathogenic gene fusions? What methods are available to test for these?

Weinberg: You need to know what you’re looking for. Unfortunately, in the broader oncology community, we’re still not testing for these appropriately. While some can be identified on tests like DNA sequencing, we often need whole transcriptome RNA-based sequencing to find novel fusion partners.

If you’re looking for a specific fusion, there may be a gene connected to another gene that’s never been seen before, a novel fusion, that you’ll miss if you’re not doing broad transcriptomics. If there is available tissue to do the broad DNA/RNA-based sequencing up front, these will often also be missed in liquid biopsies. Some drugs in development can target specific gene fusions. We won’t find them unless we look for them in the right way.

O’Reilly: You don’t know what you don’t know. You need to think about the clinical context. For example, we mostly see these fusions in the setting of non–RAS[-mutant disease]. That’s a subset of patients where it may be worth a deeper look. Sometimes you need orthogonal testing to understand the significance of your findings, including DNA- and RNA[-based sequencing].

However, looking for the ultimate downstream consequences is the functional aspect of immunohistochemical testing. It can require careful consideration and discussion with your molecular pathologists in the right context to find these fusions. Even when we know they might be present, we sometimes miss them because of the limitations of tissue. This gap is not just in tissue biopsies but also in liquid biopsies. Sensitivity for fusion testing is a work in progress.

Weinberg: Even if you have identified a fusion, and it’s a known fusion that is present in the literature, it often is still not clear how clinically actionable it is, or what the best way to approach it clinically is. If you take a step further and find a novel fusion that’s never been identified before, it’s difficult for those of us in the clinic to know what to do. We lean heavily on our molecular pathologists and those who we collaborate with in the molecular testing arena to help us know what these downstream effects are and how clinically actionable these fusion findings are.

What are the consequences of not testing for actionable alterations?

O’Reilly: Not testing for actionable alterations is a missed opportunity to affect the patient care that comes with FDA- and guideline-endorsed and accepted alterations. There are also several alterations for where there isn’t an FDA-approved guideline or endorsement, but there might be an off-label option. That’s in the context of some rare fusions in certain subsets like the KRAS wild-type [population].

The younger patient population, in general, is the group enriched for fusions across multiple malignancies, such as RET, MET, ALK, ROS, and NRG1. They’re uncommon but often have high response rates [with targeted therapy] and are attractive considerations for patients [because they shift us] away from cytotoxic therapy, which is our mainstay in the GI world. These can be meaningful and rewarding for everybody when they’re identified and actionable.

Weinberg: Lack of testing is a missed opportunity for everyone, both for the patient, but also for those of us in academia to learn more about the natural history of these illnesses and the push to develop novel treatments for these specific biomarker-driven tumors. If we don’t look, we’re not going to find them.

I have a patient with metastatic CRC [mCRC] with a RET fusion, and thankfully, RET inhibitors now have an agnostic approval. He didn’t do well on cytotoxic chemotherapy, and having [RET inhibitors] is a game changer, even if he has prolonged stable disease [SD]. We’re mostly driven by response rates, but prolonged SD in many patients with advanced malignancies, especially with some treatments that are generally less toxic than cytotoxic chemotherapy, where it’s hard to maintain the same treatment intensity indefinitely, can be influential for patients.

Individually, [these fusions] are rare. There are not many RET fusion–positive CRCs or NRG1 fusion–positive pancreatic cancers. However, like in lung cancer, if you slice the pie into enough pieces, those tiny slices start to add up into a whole segment of patients with actionable, targetable alterations. That’s what we’re hoping to do in GI cancers, and we’re starting to get there.

O’Reilly: We’re now in the era of RAS therapeutics. This will be important in CRC and pancreatic cancer, and to a lesser extent in [diseases like] biliary cancers. We’ll need to know the specific allele for some of these agents, and in others, it’s less clear if that’ll be important. Understanding and targeting mechanisms of resistance in tissue and liquid biopsies together and interpreted with germline testing will be optimal for many of these clinical contexts.

Weinberg: I see many patients, especially in the CRC space, with identified KRAS mutations, and they have no reason to look beyond that. However, we do sometimes see KRAS with another actionable alteration, and we also now care what KRAS mutation it is. G12C is now targetable, and hopefully G12D and maybe G12V are not far behind. Many patients have those driver mutations, and we’re optimistic that these drugs will be safe and effective for them once they get further into the clinic.

With more early onset diagnoses in CRC, how does age factor into conversations about the importance of precision medicine and targeted therapy approaches?

Weinberg: Any patient under the age of 50 years with a diagnosis of CRC merits referral to a genetic counselor and appropriate germline testing. This group is more heavily enriched for having inherited CRC syndromes such as Lynch syndrome. We care because of that potential genetic risk.

We also care because immunotherapy has been shown to be effective in metastatic MSI-high/ dMMR CRC and in locally advanced rectal cancer. A group at MSK showed that giving single-agent immune checkpoint inhibitors for locally advanced, MSI-high rectal cancer melted tumors away, and patients didn’t need radiation, chemotherapy, or surgery. We also want to do good precision somatic testing of these tumors, because we may find more actionable alterations in this subpopulation.

O’Reilly: These trends in younger patients in CRC and GI cancers in general is interesting. It’s not completely underpinned by increased germline findings, although there’s an enrichment in that group. It may be more the tissue and somatic testing that has the greatest yield for these patients.

Weinberg: In addition to the struggle to explain this rise in young-onset CRC since the mid-1990s, other early-onset GI cancers are a big question mark we’re trying to explain. It’s not because suddenly everybody started getting diagnosed with Lynch syndrome. There’s clearly something different going on.

It does not seem to be just a genomic issue. We don’t see huge genomic differences in young-onset vs older-onset MSI-stable CRC. It may have something to do with the microbiome, [such as] a cohort exposure in the late 1970s or early 1980s, because that seems to be the age group that was affected. We still do similar biomarker testing in the younger group with the caveat that we need to do good germline testing to ensure there is no inherited genetic syndrome, which could affect the patient and their family, screening, and treatment options.

How does RNA-based testing complement DNA-based testing methods? In what situations would you turn to RNA-based testing?

O’Reilly: RNA-based testing is complementary to DNA-based testing when we’re looking for fusions. In some molecular profiling settings, we’ll have a reflex to do RNA testing. For example, pancreas cancer is the setting where we often see the rare KRAS wild-type fusion event. The algorithm there is NGS, DNA testing, and then the Archer [FusionPlex solid tumor panel] or another related approach looking for these fusions.

Weinberg: In our clinic, we tend to order [every test] at once, if there’s available tissue, because we’re concerned about not having enough tissue. Pancreas cancer is a great example because often tissue is a major issue, and when trying to get enough to do DNA-based testing and RNA-based testing, we’re quickly depleting our tissue availability. Liquid biopsies aren’t as good at finding the novel fusion alterations.

[Many clinicians] are not doing RNA-based testing, and many also aren’t aware of what testing they’re ordering. At our clinic, we work with Caris Life Sciences, but other molecular testing companies don’t do broad-based RNA whole transcriptome testing. If you don’t know what you’re ordering, you won’t know what you’re getting.

O’Reilly: This is a complex and changing area. Almost nobody can keep up with the pace. Some of these reports have significant findings. Others require discussion and curation in the clinical context and in the context of prior treatments. In community and academic settings, we haven’t provided enough resources for how to interpret these tests and utilize and access the information. We see [patients’] second and third opinions, and they may have had Caris testing, but we don’t have the reports or the raw data, [which presents] challenges.

Weinberg: Sometimes these reports are different in their appearance and construction and have different ways of communicating the findings. I like some companies’ reports more than others. For instance, I recently saw a report from a company I was not familiar with, and it took me half an hour to go through the report and figure out what was tested, what wasn’t, and what the findings were. Even for those of us who do this day in and day out, when we see something we’re not as familiar with, it can be counterintuitive and overwhelming.

What targeted agents have had the greatest influence on the GI cancer treatment landscape?

Weinberg: One is immunotherapy for MSI-high cancers across all tumor types, especially GI cancers. That’s been a paradigm shift in the [treatment] approach for those patients. Beyond that, we’re hitting many singles and doubles, but not as many home runs like the immunotherapy drugs can be for these patients.

HER2 has now has many actionable drugs. Trastuzumab [Herceptin] plus tucatinib [Tukysa] is now an approved combination for HER2-positive mCRC. In the upper GI space, we should give a fluoropyrimidine platinum doublet with trastuzumab for HER2-amplified, metastatic upper GI tumors. Other drugs like the antibody-drug conjugate [ADC] fam-trastuzumab deruxtecan-nxki [Enhertu] also appear to have robust activity in HER2 3+ or HER2-amplified [tumors]. KRAS G12C [inhibitors are] an interesting avenue that will likely be fleshed out in combination with other drugs like EGFR inhibitors in CRC, as was recently published in the New England Journal of Medicine.

[Regarding] rare fusions, drugs are being developed for NRG1 fusions, and there are 2 approved drugs forNTRK fusions. These may not [depend] as much [on] the tumor type and may lead to more agnostic approaches. [Additionally], BRAF V600E-directed agents seem effective in lung cancer, colon cancer, and maybe other tumor types as well.

O’Reilly: Another [small but] major class that spans several GI malignancies is DNA damage repair gene alterations with BRCA1 and BRCA2 and maybe more speculatively, ATM. [These can have] counterparts that [may] result in pathogenic somatic alterations that can have actionability with platinum drugs and PARP inhibitors in patients with pancreas cancer, biliary cancer, and occasionally gastric cancer and CRC.

What is your main message for your academic and community colleagues?

O’Reilly: Think broadly. Ideally, every patient in whom there’s consideration for treatment [should have] germline testing, tumor-based sequencing, and perhaps liquid biopsy. Where there isn’t tissue, we should consider liquid biopsy. Try to match this with a clinical context, because findings related to [factors like] RAS, age, tumor location, and family history can enrich for specific settings.

Weinberg: Drugs, including the antibody zolbetuximab and other antibodies, ADCs, and CAR T cells [may target CLDN18.2]. This may be a good and actionable novel target in upper GI and pancreatic cancers. We are involved in multiple studies targeting NRG1 fusions with HER2 and HER3 antibodies. Historically, we haven’t tested for these routinely, but now that we have different clinical trials open at our [clinic], we’re always looking for those patients.

When I’m seeing a patient for a second or third opinion, 1 of the first questions I ask is: What molecular profiling have they had? Have they had appropriate germline testing? Do we have tissue available for profiling? If not, should we send a liquid biopsy? That’s often what we do because we don’t want to miss the opportunity for a matched therapy, like EGFR inhibitors for [targets that predict a] lack of benefit, like RASmutations. We are always looking for clinical trials and finding the right trial for the right patient.

O’Reilly: RAS is now actionable and is not just a negative predictive biomarker. It is hopefully a positive predictive biomarker.

Disclosures: Dr Weinberg consults on advisory boards and speaks for Merus.

Dr O'Reilly disclosed research funding to institution for Genentech/Roche, BioNTech, AstraZeneca, Arcus, Elicio, Parker Institute, NIH/NCI, and Pertzye; and consulting/DSMB for Boehringer Ingelheim, BioNTech, Ipsen, Merck, Novartis, AstraZeneca, BioSapien, Astellas, Thetis, Autem, Neogene, BMS, Tempus, Fibrogen, Merus, Agios (spouse), Genentech-Roche (spouse), Eisai (spouse).