Biomarker-Guided Treatment Approaches Transform Lung and GI Cancer Management

Michael J. Pishvaian, MD, PhD

The explosion of targeted therapies is transforming the management of difficult-to-treat malignancies, including pancreatic cancer, colorectal cancer (CRC), non–small cell lung cancer (NSCLC), and gastric cancer. Advances in the development of KRAS-, BRAF-, and Claudin 18.2 (CLDN18.2)–directed therapies suggest the growing importance of molecular profiling to identify actionable mutations and guide treatment selection, and emerging data demonstrate improved response rates and survival outcomes with these more tailored treatment approaches.

During 2 recent Biomarker Consortiums, experts sat down with OncLive® to provide insights and guidance on treatment selection and sequencing based on biomarker expression in lung and gastrointestinal (GI) cancers. Topics ranged from strategies for integrating KRAS G12C inhibitors into pancreatic cancer management and ongoing development of CLDN18.2-targeted agents in gastric cancer, to the role of BRAF inhibition across several malignancies and the optimization of HER2-targeted therapies in biliary tract cancer.

The following oncologists were featured in these conversations:

• Michael J. Pishvaian, MD, PhD, director, Gastrointestinal, Developmental Therapeutics and Clinical Research Programs, Johns Hopkins Kimmel Cancer Center, associate professor of oncology, Johns Hopkins Medicine;
• Sunnie Kim, MD, associate professor, medicine-medical oncology, University of Colorado Anschutz Medical Campus;
• John L. Marshall, MD, chief, Division of Hematology and Oncology; professor, medicine and oncology; director, the Otto J. Ruesch Center for the Cure of Gastrointestinal Cancers, Georgetown Lombardi Comprehensive Cancer Center;
• Tanios S. Bekaii-Saab, MD, the David F. and Margaret T. Grohne Professor of Novel Therapeutics for Cancer Research; chair, consultant, the Division of Hematology and Medical Oncology, Mayo Clinic; co-leader, Advanced Clinical Trials and Translational Sciences Research Program at Mayo Clinic Comprehensive Cancer Center; and
• Martin F. Dietrich, MD, PhD, medical oncologist, Cancer Care Centers of Brevard

For those who missed any of the key research and discussions from these meetings, Drs Pishvaian, Kim, Marshall, Saab, and Dietrich have provided their most notable takeaways.

KRAS Inhibition Marks a New Era in Pancreatic Cancer Management

Pishvaian: In 2024 KRAS inhibitors are probably the hottest topic in pancreatic cancer with very good reason. We’ve known for decades that the KRAS gene is mutated in 95% of all pancreatic cancers and that a KRAS gene mutation is probably a critical driver of the development, growth and metastasis that comes from pancreatic cancer. It has been a holy grail to be able to target the KRAS enzyme and suppress its activity in the hopes of suppressing the growth of pancreatic cancer. Finally, after decades of trying, we’re starting to see that become a reality.

KRAS mutations exist in 90% to 95% of all pancreatic cancers, but the first drugs that showed some successes were specific to KRAS G12C mutations. [This includes] drugs like adagrasib [Krazati] and sotorasib [Lumakras], and some of the newer drugs that have come on the market just in clinical trials show very promising results, with response rates ranging from approximately 25% to upwards of 45%. These KRAS G12C inhibitors do tend to work for 6 to 7 months on average, and they tend to be well tolerated. There are GI toxicities that can occur with these patients, as well as general fatigue and malaise, but overall, they seem to work well.

The problem is that KRAS G12C is mutated in approximately 1% to 2% of all pancreatic cancers; it’s not the big chunk. Most pancreatic cancers harbor KRAS G12D mutations or other KRAS mutations such as KRAS G12V, KRAS G12R, [etc]. [Accordingly], KRAS-nonspecific drugs have been in development over the past couple of years, [as have] the more specific KRAS G12D inhibitors. The data have been slowly trickling out [for these agents]. The data that we’ve seen the most [of] have been for Revolution Medicine’s RAS(ON) inhibitor [RMC-6236]. It’s an inhibitor that will block the KRAS enzyme function in the on stage irrespective of the underlying mutation that’s causing that constitutive activation of the KRAS enzyme. This inhibitor has achieved an approximate response rate of 20% and a 67% disease control rate [DCR]. Altogether, [it produced an] 87% duration of response [(DOR) at over 14 weeks] in the overall [patient population]. The DOR can vary anywhere from 5 months to upwards of 9 months.

The data are maturing on an ongoing basis. That might not sound like a home run, and it’s not, but it certainly [has had] a significant impact in pancreatic cancer treatment. If we look at second- and third-line treatment for pancreatic cancer, [patients treated with] chemotherapy achieved a median progression-free survival [PFS] of [approximately] 4 months and an objective response rate [ORR] of 7% to 10%. These RAS(ON) inhibitors have never been compared head-to-head in a published study, but they appear to be better than the chemotherapies that we have available.

What I think is going to happen in the coming years, especially as these drugs likely get approved, is that we will start to look at the treatment of pancreatic cancer through the lens of KRAS inhibition, and these KRAS inhibitors will be used at some point in a pancreatic cancer patient’s journey, whether it’s frontline with chemotherapy, whether it’s in the maintenance setting, whether it’s second line by itself, these drugs will be used for patients with pancreatic cancer, and we’re going to be learning a lot more in the coming years how exactly they need to be combined and integrated that into the course of that patient’s therapy.

Targeting BRAF V600E Mutations in Pancreatic Cancer, CRC, and NSCLC

Pancreatic Cancer

Pishvaain: The majority of pancreatic cancers are driven by KRAS, but then there are the roughly 5% to 10% of pancreatic cancers that are so called KRAS wild type, where there’s no detectable mutation in one of the KRAS or similar genes for those patients. There are other driver alterations that can occur, and one of the driver alterations can be just downstream of the KRAS gene in the BRAF gene. BRAF mutations occur in approximately 2% to 4% of all pancreatic cancers. Specifically, the BRAF V600E mutations, [which] we know about from melanoma, lung cancer, and cholangiocarcinoma, exist in approximately 1% to 2% of all pancreatic cancers. It’s a very small population of patients, but there are indications that patients with pancreatic cancer whose tumors harbor the classic BRAF V600E mutation can respond to BRAF V600E–targeted therapy such as dabrafenib [Tafinlar] and trametinib [Mekinist] or encorafenib [Braftovi] and cetuximab [Erbitux]. So far, it’s primarily been in the context of case reports and anecdotal stories, but there are certainly patients who’ve experienced prolonged benefit from these targeted therapies. If we perform next-generation sequencing for a patient with pancreatic cancer and a BRAF V600E mutation is identified, then we should think about offering that patient, whether in the second line or beyond, treatment with a BRAF V600E–targeted therapy.

Now there are approximately 1% to 2% of other BRAF alterations that can occur [for these patients]. Unfortunately, less is known about them. Preclinically, there’s some evidence that [patients harboring these mutations] still may respond to BRAF-targeted therapies, but we don’t know necessarily how effective those are going to be across the population. When push comes to shove, if we have a patient who doesn’t have many treatment options, and there’s not a clinical trial option available either, then considering trying to treat these patients either with a RAS/MEK combination, as you would for the BRAF V600E mutations, or even a single-agent RAS inhibitor, is certainly worth exploring.

CRC

Marshall: If you have just found a BRAF V600E mutation in your patient with metastatic CRC, there are good therapeutic options for them. Encorafenib is a BRAF V600E mutation–targeted agent that, [as seen in] the phase 3 BEACON CRC trial [NCT02928224], was coupled with EGFR-targeted therapy cetuximab. These 2 [agents[ hit the pathway at the same time and were effective enough to [gain] FDA approval. [Binimetinib (Mektovi)], a MEK inhibitor, was also included in the BEACON clinical trial, but did not result in an overall survival [OS] advantage, [although it] improved response rates. Oddly enough, when you throw the third drug in, there is a reduction in some of the toxicities. However, only 2 drugs are FDA approved in the second line or beyond for BRAF V600E–mutated CRC. [Notably], panitumumab [Vectibix] may be substituted out for the cetuximab.

NSCLC

Dietrich: Treatment selection [in NSCLC] is dependent on several factors. We have 2 regimens approved in the frontline setting that are specifically tested in BRAF V600E–mutated NSCLC. Those are typically adenocarcinomas, [but we] occasionally [see this mutation in] squamous cell carcinoma as well. We don’t have dedicated data on the effect of immunotherapy in BRAF V600E–mutated NSCLC so most of the data there are retrospective. There is quite a bit of discrepancy between the results. We think of it, especially in smokers, in patients with high PD-L1 levels is better than, for example, an EGFR or ALK mutation, where we don’t expect immunotherapy to work as well [as in NSCLC] with KRAS mutations. It’s somewhere in the middle, so the actual data there are difficult to compare, because we don’t have dedicated immunotherapy-based approaches. We know this intuitively, because we don’t think of BRAF mutations as a marker associated with immunotherapy by itself, but secondary markers like smoking status and PD-L1 level.

We have very dedicated data for first-line usage of BRAF inhibitors, and we had some very compelling updated data from the [phase 2] PHAROS study [NCT03915951] presented at the 2024 ESMO Congress. The data here are showing very high response rates in the first-line setting [with encorafenib plus binimetinib]; we saw a response rate of 75%. In those patients who did respond, the median DOR was 40 months, so it is a tremendously durable regimen and gives us an insight into the adverse effect [AE] profile [by showing] that it is truly manageable. Otherwise, it wouldn’t be plausible for patients to be on treatment for so long. We see a lower level of one of the most difficult toxicities, pyrexia, which was more prevalent with dabrafenib and trametinib across different indications.

I think there are many similarities in the EGFR-mutant space, but the response rates, DOR, and the estimated PFS [with BRAF inhibitors] very much mimics what we would expect from an excellent third-generation EGFR inhibitor like osimertinib [Tagrisso]. The guideline in my clinical practice [is to] emphasize the use of targeted therapy up front. We don’t have any dedicated information for immunotherapy-based approaches, whether it’s alone or in combination with chemotherapy, so we must accept the data the way they stand right now. In fact, we were contemplating whether clinical trials should include a direct comparison of targeted therapy with encorafenib and bimentinib against chemotherapy plus immunotherapy. [However], we concluded that this is not a desirable [regimen for] the standard-of-care arm, because it would be an extrapolation from an adenocarcinoma histology with chemoimmunotherapy approaches to a very specific subset of BRAF V600E–mutant NSCLC.

Ultimately, my practice is to use targeted therapy first. The data are very compelling, [showing] deep and very durable responses. In addition to that, [they have] an excellent safety profile in the brain. Contrasting this with their efficacy in the second-line setting, we saw fewer responses in [approximately] half the patients. DORs were still respectable, but not nearly as good as they were in the first-line setting. Even though the study wasn’t evaluating first- vs later-line usage, the observation is that targeted therapy should be utilized in the most native clonal setting where we have no resistance mutations introduced using chemotherapy, so first-line usage here clearly the preferred way to go. If we see a BRAF V600E mutation on liquid biopsy, I recommend ordering targeted therapy in this treatment setting.

Zolbetuximab’s Efficacy Validates CLDN18.2 as a Key Target for Drug Development in Gastric Cancer

Kim: CLDN18.2 is a newer biomarker of interest in gastric cancer. It’s a tight junction protein that’s normally expressed in gastric mucosa cells and in states of cancer, it becomes exposed on the surface of these gastric adenocarcinoma cells, making it a more promising target. It’s overexpressed in approximately 30% to 40% of gastric/ gastroesophageal junction [GEJ] adenocarcinoma.

Recently, there were 2 global phase 3 trials that recently read out: SPOTLIGHT [NCT03504397] and GLOW [NCT03653507]. [These] were concurrent trials that evaluated the addition of zolbetuximab-clzb [Vyloy], which is a monoclonal antibody targeted against CLDN18.2 in combination with a fluoropyrimidine and a platinum combination. CLDN18.2 intensity had to be 2+/3+ in 75% or more of cancer cells.

SPOTLIGHT looked at the addition of zolbetuximab to mFOLFOX vs mFOLFOX alone and GLOW evaluated the addition of zolbetuximab to capecitabine [Xeloda] and oxaliplatin [CAPOX] compared with CAPOX alone. The reason why these 2 concurrent phase 3 trials [were conducted was] because SPOTLIGHT was mostly looking at the patients in North America and Europe. [Conversely], GLOW was primarily [conducted in] East Asian countries, and we know from many prior studies that sometimes the prognosis and outcomes can be different in the East Asian patient population vs the North American/European patient population. What we found, though, in both studies was that the OS benefit was approximately 3 months compared with the standard of care. Unfortunately, there was no change in ORR with the addition of zolbetuximab.

In terms of AEs, the most important ones to know are nausea and vomiting. Unfortunately, it happens in the chair. [With] a lot of our chemotherapy regimens, the patients experience nausea or vomiting after they leave at their home. In this case, it can happen immediately. The way we treat that is by decreasing the infusion rate and maximizing the antiemetic regimen. There’s also a suggestion that the nausea or vomiting is worse with the first cycle and improves significantly with subsequent cycles.

FDA approval [was granted to] zolbetuximab plus doublet chemotherapy for first-line, CLDN18.2-overexpressing gastric cancer [in October 2024]. In terms of the future of CLDN18.2 there are many drugs that are undergoing clinical investigation right now. The way we’re trying to target CLDN18.2 goes beyond monoclonal antibodies. We’re looking at antibody-drug conjugates, bispecific antibodies, and CAR T-cell therapy. We’re finding that with these other drugs, we may not need the CLDN18.2 expression to be as high as it was in the SPOTLIGHT and GLOW trials.

Kim and Marshall Provide Guidance for Biomarker-Based Gastric Cancer and CRC Management

Gastric Cancer

Kim: We are fortunate in the gastric cancer field to be seeing an increased number of biomarkers that we can target. It’s exciting, but it also raises a lot of questions about which biomarker we should prioritize. This is essentially what we call a data-free zone, where we, in some cases, don’t know what’s the best way to approach it. I’ll give you some of my recommendations.

[If we] have patients who are PD-L1–, CLDN18.2-, and HER2-positive, which biomarker should be prioritized? This is an easier situation, because the trials looking at PD-L1– and CLDN18.2-positive tumors specifically excluded patients who have HER2-positive gastric cancer. In this situation, I would prioritize targeting HER2. We know that when trastuzumab [(Herceptin) with/without] pembrolizumab [Keytruda] is combined with doublet chemotherapy, there is that chance for a durable response. We have multiple drugs that can target HER2, namely trastuzumab and fam-trastuzumab deruxtecan-nxki [Enhertu; T-DXd]. We have the most positive data with HER2-positive tumors, and I would prioritize that over the other biomarkers.

The other situation we see is someone who has HER2-negative, PD-L1–positive and CLDN18.2-positive gastric cancer. [Questions about how to approach treatment for these patients] are probably the most common [ones] I get. Here it depends on high how high the PD-L1 score is. If we have a patient with a PD-L1 score of 10 or higher and a CLDN18.2-positive tumor, I would go with a PD-1 inhibitor–based approach, because we see an improvement not only in median OS compared with chemotherapy alone, but in ORR. Now that we have approximately 4 years of follow-up, we find that 10% of these patients experience a durable response. For me, if the PD-L1 score is 10 or higher, I would go with chemotherapy plus a PD-1 inhibitor.

For the patient who is CLDN18.2-positive but has a lower PD-L1 score between 1 to 4, I would aim for giving that patient doublet chemotherapy and zolbetuximab because I think it’s questionable whether a gastric cancer with a PD-L1 score that’s on the lower side is going to be able to derive benefit from the addition of a PD-1 inhibitor.

We’re looking at some newer biomarkers coming up. FGFR2B looks quite promising, in addition to doublet chemotherapy plus an immune checkpoint inhibitor. If that becomes FDA approved, then we’ll be trying to prioritize what biomarker to target first. In the field, we are developing clinical trials so we can understand the proper sequencing [of available regimens]. [This] can give patients the best shot at having all their biomarkers targeted during treatment for their disease.

CRC

Marshall: [In this next scenario], you just opened your molecular profiling report and [your patient] has HER2-positive metastatic CRC. You’re glad you did the test, because you have very good therapeutic options for this patient. The primary one is based on the [phase 2] MOUNTAINEER trial [NCT03043313] in HER2-positive CRC [evaluating] the combination of tucatinib [Tukysa] and trastuzumab. [The combination produced] a dynamite 38% ORR [and 8.1-month] median PFS for those patients who were benefiting. [This is] a very strong benefit with a tolerable AE profile. I want to reiterate that those HER2-positive patients are frequently found in those patients you also want to be giving an EGFR[-targeted] drug to because they are left-sided, they are RAS wild-type and BRAF wild-type. You have to make sure [to identify] HER2-positivity in that patient subset because of this therapeutic option.

FGFR2 Inhibitors Show Durable Responses in Cholangiocarcinoma

Bekaii-Saab: [Between] 5% to 10% of patients with intrahepatic cholangiocarcinoma harbor an FGFR2 fusion. There are 2 agents, previously 3 agents, FDA approved for treating patients with FGFR2 fusions in biliary tract cancer. Currently, the ones that remain on the market are pemigatinib [Pemazyre] and futibatinib [Lytgobi]. There was a third agent called infigratinib [Truseltiq], which unfortunately was pulled [by the FDA]. However, my understanding is it will be approved in a non-cancer indication.

These FGFR inhibitors induce some meaningful responses in many of these patients. In fact, approximately 40% of patients will have a response, and many of these responses tend to be quite durable. Many of my patients have gone on to [experience] many years of response. These agents do come with some toxicities, primarily skin, nail and hair toxicities, but also phosphate abnormalities and ocular changes. They [are associated] with quite a few toxicities that we need to keep an eye on [and potentially] refer patients to the specialists such as the pharmacologist or dermatologist, as needed.

Interestingly, FGFR2 fusions are primarily found not only in intrahepatic cholangiocarcinoma, but in younger patients and specifically in female patients. Therefore, younger females with intrahepatic cholangiocarcinoma are more likely to harbor an FGFR2 fusion. Now attempts to bring this class of agents into the first line have led to a couple of studies. One [evaluating] infigratinib has stopped accruing after [the agent] was pulled from the market. Another one [was investigating] pemigatinib vs gemcitabine and cisplatin.

Optimizing Treatment Sequencing in HER2-Overexpressed Biliary Tract Cancer

Bekaii-Saab: Agents for HER2-overexpressing biliary tract cancer are increasing given the interest and given the high likelihood for response. We have agents such as trastuzumab, T-DXd, pertuzumab [Perjeta], tucatinib, and a bispecific antibody called zanidatamab-hrii [Zihera]. When you look at the cumulative data, the 3 agents or combinations that stand out right now are trastuzumab and tucatinib, T-DXd, and zanidatamab. When we look at the response rates, they all have a response rate [around] 48%. These responses tend to be durable, although with zanidatamab we can see that the durability of responses, at least historically, has been the longest reported.

Today in actual practice, given what we have on hand, I would start a patient on tucatinib plus trastuzumab following progression on chemotherapy in the second line. That would be followed upon progression with T-DXd. The reasoning for this is that first I want to hit the receptor. Once the tumor develops resistance to HER2 blockade, then I want to use that receptor as a docking station for T-DXd, which essentially serves as targeted chemotherapy. It doesn’t care as much about the biology of the receptor as much as it cares about its presence. T-DXd delivers the [payload], deruxtecan, into the tumor and achieves a cancer kill. The studies that looked at T-DXd looked at it both in patients who didn’t receive prior HER2-targeted therapies vs those who did. When we look at responses and outcomes for patients who either received prior anti-HER2 therapies and those that didn’t, you see [similar] response rates.

The toxicities [associated with] T-DXd can be limiting. There are quite a few chemotherapy-related toxicities [as well as] interstitial lung disease, which can be quite problematic, for approximately 10% of the patients, including the possibility of death. [Now that] zanidatamab [has been] approved, [the question is] where we would place it in that sequence.

c-Met–Targeted Therapies Should Be Prioritized for Frontline Management of NSCLC Harboring MET Exon 14 Skipping Mutations

Dietrich: The frontline [treatment] of [patients with] MET exon 14 skipping mutations isn’t sorted. It’s similar to some of the other oncogenes that we are treating based on side-by-side clinical trial comparisons. We see nice responses in the first-line setting, and many of these responses are durable. Again, we don’t have any dedicated data [evaluating] the effect of immunotherapy in [NSCLC harboring MET exon 14 skipping mutations]. This can be misleading, because we have crosstalk between c-Met and PD-L1 expression. A lot of patients with MET exon 14 skipping mutations have very high PD-L1 levels, but their prognostic outlook is not [aligning] with our expectations for these high and very-high levels of PD-L1 expression. It is a bit of a diagnostic trap that we are falling into.

In my practice, if we have an oncogenic driver—and this is in line with National Comprehensive Cancer Network guidelines— we prioritize targeted therapy treatment. We have 2 agents approved [to target MET exon 14 skipping mutations in NSCLC]: tepotinib [Tepmetko] and capmatinib [Tabrecta]. [These] 2 agents have no direct side-by-side comparisons. We have a very large dataset in the [phase 2] VISION trial [NCT02864992] investigating tepotinib, [which is] the largest trial that we see in the space, and we see [data from the phase 2], GEOMETRY [mono-1 trial (NCT02414139) evaluating] capmatinib. Both demonstrated high levels of efficacy in the body and the brain [with these respective agents], and the response rates were very high. There [are some] logistical considerations; tepotinib is a once-daily drug, [while] capmatinib has to be taken twice daily, which could influence care delivery. The response rates are around 60% to 70% for patients in the frontline setting and may be a little bit lower in the second-line setting. The DOR is variable. There are some very broad observations which are more difficult to quantify. In these multicohort studies, we expect that approximately 50% of patients have not progressed at 12 months. The median [PFS] here is harder to estimate, but survival [was greatly impacted] for patients who received treatment with the c-Met–targeted agent up front.

In my practice, it’s the most important one. We don’t know a lot about the post-progression mechanisms for c-Met. We certainly don’t have any therapeutic options currently approved on label. Antibodies may be of consideration, but the escape mechanisms have quite a bit of breadth here in presentation, so we would [most likely] use a biomarker-guided therapy in the second-line setting.

For more expert soundbites from the GI and Lung Cancer Biomarker Consortiums, click here.