Gut Dysbiosis as a Potential Guide for Immunotherapy (Dis)Continuation After 2 Years in NSCLC: A Brief Report

Although most phase III pivotal trials have set the duration of immune checkpoint blockers (ICB) for advanced NSCLC at 2 years, the criteria for safely discontinuing ICB remain undefined. Growing evidence links ICB efficacy to gut microbiota, positioning gut microbial taxonomic profiling as a promising biomarker to guide treatment decisions. We performed a retrospective analysis exploring clinical outcomes and the utility of multiomic decision-making tools in patients with NSCLC at Gustave Roussy who completed 24 months of ICB-based therapy without disease progression (PD).

Patients receiving ICB between July 2016 and January 2023 were identified from the ONCOBIOTICS (NCT04567446) and STING (NCT04932525) study datasets. We selected those who reached 24 months of treatment without disease progression. Clinical characteristics and multiomic assessments, including gut microbiota profiling (TOPOSCORE by whole-genome sequencing), positron emission tomography-18-fluorodeoxyglucose imaging, and circulating tumor DNA, collected at 24 months, were analyzed. Key outcomes included overall survival (OS), progression-free survival (PFS), and PFS rates at 24 months after the completion of 2 years of ICB, stratified by molecular, metabolic, and microbial signatures.

Out of 123 patients treated for at least 18 months, 35 completed 24 months, with 31 eligible for the analysis. Of these, 68% continued ICB, whereas 32% discontinued therapy at the physician's decision. Clinical characteristics were similar across groups. After a median follow-up of 59.1 months, OS and PFS did not differ significantly between those who discontinued and those who continued treatment (OS p = 0.9012). Among all multiomic tools, gut microbiota composition exhibited a trend (though not statistically significant) association with PFS rates at 24 months after the completion of 2 years of ICB. Patients with a favorable microbiota profile had a higher rate of sustained response at 24 months compared with those with dysbiotic signatures (81% versus 44%, respectively, p = 0.0870).

Discontinuing ICB after 24 months did not negatively impact OS in our real-world cohort. Although limited by the small sample size, these findings support the potential of gut microbiota profiling as a promising tool to guide ICB duration. Integrating a translational multiomic algorithm, in particular microbial signals, may help personalize treatment strategies and safely shorten immunotherapy courses.