Oncologists Weigh Real-World Nuances in First-Line Gastric Cancer

During a live Case-Based Roundtable event, Ronan J. Kelly, MD, MBA, FASCO, Chief of Oncology at Baylor Charles A. Sammons Cancer Center in Dallas, Texas, and participants discussed the practical differences and surprising similarities among 3 approved PD-1 inhibitors in the first-line treatment of HER2-negative metastatic gastric and gastroesophageal junction cancer, and how the evolving dosing and administration are changing practice when using immune checkpoint inhibitors.

This is the second of 2 parts. Read part 1.

Register today to join a Case-Based Roundtable near you.

CASE SUMMARY

• A 60-year-old man with abrupt 10-pound weight loss, dyspepsia, bloating after meals, and loss of appetite, and all symptoms worsening over the past 3 months
• Family history: father deceased age 50 (gastric cancer)​
• Medical history: overweight, hypertension​
• Esophagogastroduodenoscopy showed 2-cm protruding mass in the body of the stomach, without ulceration​
• Biopsy revealed poorly differentiated adenocarcinoma
• Molecular testing: HER2/neu (-); mismatch repair proficient/microsatellite stable; PD-L1 CPS: 5; Claudin18.2 < 75%​
• CT scan of abdomen and chest revealed a gastric polypoid mass and thickening; no ascites; 2 lesions in the left lung lower lobe​
• Lung biopsy confirmed metastatic adenocarcinoma consistent with gastric primary ​
• Diagnosis: stage IV gastric adenocarcinoma

Before reviewing any trial data, the group's treatment preferences were nearly split: 45.5% of participants favored chemotherapy plus nivolumab (Opdivo), 45.5% chose chemotherapy plus pembrolizumab (Keytruda), and 1 participant selected tislelizumab (Tevimbra). Kelly noted that no single answer was wrong; the split reflected where the field stands.

DISCUSSION QUESTION

• What factors influence your decision regarding which immune checkpoint inhibitor to use with first-line chemotherapy?

Kelly anchored the data review with CheckMate 649 (NCT02872116), the trial that broke a longstanding barrier in gastric oncology. Before this study, he noted, no regimen had pushed median overall survival past 12 months in unselected HER2-negative metastatic gastric cancer. At 5 years of follow-up, roughly 16% of patients in the nivolumab arm remained alive, a figure Kelly acknowledged was modest but meaningful given the prior natural history.¹

KEYNOTE-859 (NCT03675737) followed a similar design but enrolled at a broader biomarker threshold: PD-L1 CPS (combined positive score) of ≥1 compared with CPS ≥5 in CheckMate 649. He said CPS ≥1 captures approximately 80% of the gastric cancer population. Kelly noted that comparing efficacy numbers across CheckMate 649 and KEYNOTE-859 is inherently an apples-to-oranges exercise because of the different enrichment thresholds. The nivolumab data reflect a more biomarker-selected population, but the overall hazard ratio profile is strikingly similar.^1,2

Tislelizumab was evaluated in RATIONALE-305 (NCT03777657). Kelly flagged a notable methodological difference: this trial used an alternative to CPS called tumor area positivity (TAP), which estimates the proportion of the tumor area occupied by PD-L1–staining cells rather than counting individual positive cells. He emphasized that concordance between CPS and TAP is high in practice, so clinicians reading the results need not treat the populations as categorically different. Kelly observed that the safety profile of tislelizumab‚ including immune-mediated adverse events occurring in 7.6% at grade 3 or higher‚ was similar to what was seen with nivolumab and pembrolizumab.^1-3

DISCUSSION QUESTIONS

• How have you incorporated different dosing schedules and methods of administration?
• How does the CPS score influence your treatment selection?​

An extended discussion emerged around dosing and administration of these agents. Kelly asked whether the longer dosing intervals now available for both nivolumab and pembrolizumab changed the toxicity experience. Ahmed Gulrayz, MD, reported that patients switched from 3-week to 6-week pembrolizumab experienced more adverse event and asked to return to the shorter schedule. Deepak Bhamidipati, MD, said the schedule seems to not matter generally, but he is comfortable initiating nivolumab at the ever-4-weeks dose.

Bipin Amin, MD, said that when patients came in less frequently during the COVID-19 pandemic, he started using the 4-week dose for nivolumab and 6-week dose for pembrolizumab without problems. Rashad Khan, MD, described a patient who developed more diarrhea on once-monthly subcutaneous nivolumab and transitioned back to biweekly intravenous dosing, and added that for patients with esophageal cancer, he actually prefers the 2-week interval regardless, because it provides more frequent opportunities to monitor nutrition and hydration.

The subcutaneous nivolumab formulation also drew comment. Karthik Anand, MD reported that his hospital had made subcutaneous the default, requiring active effort to order the intravenous formulation. Rozina Chowdhery, MD, said her academic affiliate had done the same. Gulrayz said that although subcutaneous dosing is quicker, his institution does not prefer one or the other. Kelly noted that institutional adoption of the subcutaneous immunotherapies is likely to accelerate as infusion chair time becomes a constraint, though formal comparative toxicity data across schedules and formulations remain limited.

The FDA ODAC Decision and Its Practical Implications

Kelly recounted a pivotal September 2024 FDA Oncologic Drugs Advisory Committee meeting in which stakeholders including the 3 drug manufacturers assembled to address a key question: should PD-1 inhibitors continue to be prescribed in patients with CPS less than 1, approximately 20% of the gastric cancer population? The committee voted unanimously to remove that indication across all 3 agents. Kelly said he had testified as an expert witness and favored retaining the all-comers approval, but the committee found insufficient evidence of benefit in immunologically cold tumors. The result is that CPS ≥1 is now the operative threshold.

Choosing Among the Three Agents

After walking through all 3 datasets, Kelly asked participants whether they saw meaningful differences between the regimens. Jack Erter, MD, said he uses pembrolizumab in practice largely for formulary reasons and regards the 3 agents as broadly interchangeable in this context, given the similarity of their mechanisms, efficacy profiles, and adverse event rates. Anand agreed, noting that in his practice the outcomes track the trial data, and that the dominant treatment-related concern remains peripheral neuropathy from oxaliplatin rather than immune-related adverse events from the checkpoint inhibitor.

Kelly added that the 1 area where tislelizumab is actively being differentiated is peritoneal metastasis, a subset in which RATIONALE-305 showed a signal, and which is clinically significant given that peritoneal disease affects up to half of all gastric cancer patients and carries a particularly poor prognosis.³ He cautioned, however, that this comparison is cross-trial and should be interpreted carefully. The post-discussion vote reflected modest redistribution rather than a clear consensus shift.

Kelly noted that the next evolution in this space will not be a better PD-1 inhibitor in isolation, but rather combination strategies. He flagged the HERIZON-GEA-01 trial (NCT05152147), which evaluated zanidatamab (Ziihera), tislelizumab, and chemotherapy in HER2-positive gastric cancer and is being reviewed by the FDA with a target date of August 25, 2026. In the HER2-negative space, the question is whether adding Claudin 18.2 targeting to the checkpoint inhibitor backbone will produce additive benefit‚ a question the ongoing LUCERNA trial (NCT06901531) is designed to answer.

Register today to join a Case-Based Roundtable near you.

_{DISCLOSURES: Kelly previously reported advisory board/consulting fees from Astellas, AstraZeneca, Bristol Myers Squibb, Daiichi Sankyo, Eisai, Eli Lilly, EMD Serono, Exact Sciences, Grail, Ipsen, Merck, Novartis, Novocure, Phillips, Takeda, Toray.}

REFERENCES

1. Janjigian YY, Kawazoe A, Bai Y, et al. Pembrolizumab plus trastuzumab and chemotherapy for HER2+ metastatic gastric or gastroesophageal junction (mG/GEJ) adenocarcinoma: Survival results from the phase III, randomized, double-blind, placebo-controlled KEYNOTE-811 study. Ann Oncol. 2023;34(suppl 2):S851-S852. doi:10.1016/j.annonc.2023.09.1424

2. Rha SY, Wyrwicz L, Yanez Weber PE, et al. KEYNOTE-859: 4.5-year median follow-up of pembrolizumab plus chemotherapy for previously untreated advanced HER2-negative gastric or gastroesophageal junction (G/GEJ) adenocarcinoma. J Clin Oncol. 2025;43(suppl 16):4036. doi:10.1200/JCO.2025.43.16_suppl.40363.

3. Cruz-Correa M, Oh DY, Kato K, et al. Tislelizumab + chemotherapy in gastric cancer: long-term RATIONALE-305 randomized trial follow-up. Adv Ther. 2026;43(1):165-183. doi:10.1007/s12325-025-03415-04.