Immunotherapy Combination Selection Is Driven by PD-L1 in Gastroesophageal Cancer

Andrew H. Ko, MD

The addition of nivolumab (Opdivo) or pembrolizumab (Keytruda) to a chemotherapy backbone has become part of the frontline standard of care for patients with gastroesophageal cancer. Deciding on the right combination for individual patients depends on their PD-L1 status, since different agents and combinations yield varying benefits, according to Andrew H. Ko, MD.

“I would emphasize that PD-L1 combined positive score [CPS] should be assessed prior to [deciding to utilize] a combination strategy. But in those individuals [with high PD-L1 CPS], there is significant benefit with the combined approach,” Ko said in an interview with OncLive^® following an Institutional Perspectives in Cancer (IPC) webinar on gastrointestinal (GI) cancer.

In the interview, Ko, who was the chair of the IPC event, discussed the focus of each presentation from the meeting: the multitude of treatment options for patients with gastroesophageal cancer; emerging combinations in hepatocellular carcinoma (HCC); and how HER2 expression can drive therapy decisions in colorectal cancer (CRC). Ko is a professor of clinical medicine in the Division of Hematology/Oncology at the University of California – San Francisco (UCSF) Helen Diller Family Comprehensive Cancer Center.

OncLive^®: At the IPC meeting, you spoke on the role of immunotherapy in gastroesophageal cancer. How can PD-L1 expression factor into the choice of therapy for these patients?

Ko: The emerging standard for treatment of metastatic gastric and esophageal cancers involves a combination of chemotherapy plus an immune checkpoint inhibitor [ICI]. I emphasized that this should be guided by assessment of the PD-L1 expression as measured by the CPS of tumors. If you dissect the data from both the pivotal [phase 3] CheckMate 649 [NCT02872116] and KEYNOTE-590 [NCT03189719] studies for adenocarcinomas, the benefit of adding an ICI, whether it be nivolumab or pembrolizumab, appears to be driven by the subset of patients who have high PD-L1 scores, as measured by CPS. For nivolumab, that would be a CPS [of at least] 5, and for pembrolizumab, that would be a CPS [of at least] 10.

For esophageal squamous cell cancer [ESCC], an alternative for frontline treatment can involve a chemotherapy-free regimen with the combination of ipilimumab [Yervoy] and nivolumab. In the [phase 3] CheckMate 648 study [NCT03143153], that group of patients [with ESCC] treated with this dual immune checkpoint inhibitor strategy showed improved survival compared with chemotherapy by itself. Thus, this represents another option besides chemotherapy plus nivolumab for ESCC. Importantly, I would not consider this immunotherapy-only strategy to hold true for esophageal/gastric adenocarcinomas. The predictive value of PD-L1 as a biomarker for ESCC is not quite as strong as it is for adenocarcinomas.

Meanwhile, the [phase 3] CheckMate 577 study [NCT02743494] looked at adjuvant nivolumab for patients with resected esophageal and gastroesophageal junction cancer, both squamous cell and adenocarcinomas, following neoadjuvant chemoradiation and surgery. Those who have less than a pathologic complete response [CR] should be getting 1 year of adjuvant nivolumab as the gold standard of treatment. That is irrespective of PD-L1 expression.

Given that the threshold for nivolumab is a CPS of 5 or greater vs 10 or greater for pembrolizumab, are more providers opting for nivolumab to catch a greater proportion of patients?

That is true for my practice, although either option is entirely appropriate. However, it is also worth noting for the subset of patients [with] HER2-positive disease, we have the [phase 3] KEYNOTE-811 trial [NCT03615326], which looked specifically at chemotherapy plus pembrolizumab plus trastuzumab [Herceptin]. That might be the group that I would preferentially use pembrolizumab in.

R. Kate (Katie) Kelley, MD, of the UCSF Helen Diller Family Comprehensive Cancer Center, discussed updates in HCC, covering the phase 3 IMbrave150 (NCT03434379), COSMIC-312 (NCT03755791), and HIMALAYA (NCT03298451) trials. How did Dr Kelley reconcile these 3 datasets and how they affect the current standard of care?

Atezolizumab [Tecentriq] plus bevacizumab [Avastin] remains the frontline standard for patients [with HCC]. Dr Kelly tried to counsel against making too many cross-study comparisons, but the most robust survival data are with the combination of atezolizumab [plus] bevacizumab. There is, however, going to be a substantial proportion of patients who will not be candidates for bevacizumab for one reason or another, with the most common being significant esophageal varices and the associated bleeding risk.

It does appear that now other options are [emerging], including tremelimumab [plus] durvalumab [Imfinzi], which will be a combination that winds up being a popular alternative for those patients who have a contraindication to receiving bevacizumab. It is important to emphasize the way that tremelimumab/durvalumab was given in [HIMALAYA]—by front-loading the dose of [tremelimumab]. There are also encouraging phase 3 data with atezolizumab plus cabozantinib [Cabometyx], although maybe not as strong [as with tremelimumab and durvalumab].

There’ll be a lot of room for further study in terms of sequencing these various regimens in the frontline and second-line [settings]. Patient selection for these respective regimens will be important to tease out. While we do not have a definitive right or best single approach yet, it’s amazing how many positive readouts we’ve had in HCC over the past several years.

Chloe Atreya, MD, PhD, of UCSF Helen Diller Family Comprehensive Cancer Center, discussed BRAF-mutated and HER2-positive CRC. How could fam-trastuzumab deruxtecan-nxki (Enhertu) factor into the treatment landscape for patients with HER2-positive CRC?

We are learning that it’s critical to perform molecular testing, including next-generation sequencing, in patients with CRC to look for multiple targets, including HER2 expression, RAS/RAF mutations, and KRAS G12C mutations. Patients with BRAF mutations, particularly those with the canonical BRAF V600E mutation, will be candidates for the BEACON regimen consisting of encorafenib [Braftovi] together with cetuximab [Erbitux]. This has emerged as an appropriate second-line standard for that subgroup of patients, who have worse disease biology overall. Dr Atreya also shared some nice data with regard to the subset of patients with HER2 overexpression and the possibility of responding to the antibody-drug conjugate [ADC] trastuzumab deruxtecan. This serves as a good reminder in terms of how we’re able to identify these molecular subtypes of CRC in which there will be options on top of or beyond our standard chemotherapies.

Based on the data that we’ve seen with the BEACON CRC regimen, are you recommending that previously untreated patients with metastatic CRC enroll, if eligible, in the phase 3 BREAKWATER trial (NCT04607421), which is exploring encorafenib plus cetuximab with or without chemotherapy?

Yes, mature trial results looking at different contexts in which to use this regimen, whether it be in the frontline or in the adjuvant setting, will be eagerly awaited. Those readouts would potentially be practice changing.

How do you predict that trastuzumab deruxtecan will ultimately stack up against some of the combination and triplet regimens that we’re seeing with trastuzumab in CRC?

We’re already seeing robust results [with trastuzumab deruxtecan] in HER2-low breast cancer, and we’ve seen some data recently come out [with that agent] in cholangiocarcinoma as well. Teasing out levels of HER2 expression will be a key part in understanding how this ADC may play a role in these different tumor types, including CRC. Everyone rightfully has an eye out for the interstitial lung disease that can occur with the agent, but I suspect that [trastuzumab deruxtecan] will be a major player for HER2-positive CRC in the future.

Alan P. Venook, MD, of UCSF Helen Diller Family Comprehensive Cancer Center, gave a presentation on the shifting paradigms for neoadjuvant treatment for rectal cancer and the increasing neoadjuvant use of immunotherapy for microsatellite instability–high disease. How has the availability of neoadjuvant immunotherapy affected traditional paradigms?

Even with the small numbers, the key messages from [the phase 2] trial [NCT04165772] are the 100% clinical CR rate and the [apparent] absence of any substantial toxicity [with single-agent dostarlimab-gxly (Jemperli)]. As we’re moving more in the rectal cancer treatment paradigm [toward] trying to identify patients who will be suitable for nonoperative management, those who will not require a big rectal cancer surgery, it is—even with the small numbers—a practice-changing paradigm for those patients with MSI-high tumors. [Given that neither] surgery, radiation treatment, or chemotherapy [would be required], it’s no wonder that patients are so excited about [the results]. [However, we have to recognize that the criteria for this trial] apply to 5% or less of patients. Of note, the study was done with dostarlimab, as opposed to the PD-1 antibodies we more commonly use. I’m not sure I would expect different results with other antibodies.

[Because of the mainstream media coverage of this trial], it is important to contextualize [the dostarlimab data]. All of us have patients asking about this approach for rectal cancer and even other disease types. We need to put [these data] in context and help patients understand that while these are terrific data, they are not broadly applicable to MSS tumors or cancers other than rectal cancer at this point.

In general, we are moving more toward a paradigm of total neoadjuvant therapy for rectal cancer. Additionally, we are seeing promising data now with longer-term follow-up [from studies such as the phase 2 OPRA trial (NCT02008656)], in terms of nonoperative management being suitable for patients who demonstrate a robust response to induction chemotherapy and radiation treatments on the front end. Specifically, selected patients who have achieved a clinical CR may be able to avoid a major operation as long as they undergo intensive surveillance and follow-up, with results looking quite good after several years.

Would you like to highlight any further research efforts at UCSF?

We are very interested in pushing the envelope in pancreatic cancer. UCSF is part of several important national initiatives with what I’ll call platform trial designs exploring multiple new experimental agents in succession. One is the Precision Promise platform [trial (NCT04229004)] with the Pancreatic Cancer Action Network, and another is the REVOLUTION trial (NCT04787991)], supported by the Parker Institute for Cancer Immunotherapy, both looking at novel therapeutic strategies for metastatic pancreatic cancer.

More broadly, we are also interested in cellular therapies for solid tumors, a field that’s still very much in its infancy. But we now have a chimeric antigen receptor [CAR] T-cell study, [a phase 1/2 trial (NCT04404595)], open for gastric cancer, and we will hopefully activate one soon specifically for CRC. We’re [also] trying to develop similar trials in the years to come for [patients with] pancreatic cancer. Patient selection for these CAR T trials will be critical because [they are] intensive therapies. If we can [get] some traction with this very different way of trying to treat some of these GI tumors, it will be very exciting.