Clinical characterization of FGFR2b expression in patients with advanced gastric or gastroesophageal junction adenocarcinoma.

Background: Fibroblast growth factor receptor 2b (FGFR2b) is a novel protein biomarker expressed in gastric and gastroesophageal junction tumors (GC/GEJC). Phase III trials are evaluating the efficacy and safety of FGFR2b-targeting therapies. However, features of FGFR2b-expressing tumors and co-occurrence of FGFR2b expression with currently actionable biomarkers in gastric cancer remain unclear.

Materials And Methods: We carried out a single-institution retrospective cohort study of patients who initiated systemic therapy for GC/GEJC to evaluate features of FGFR2b-positive tumors, first-line (1L) treatments received, and co-occurrence with select actionable biomarkers [human epidermal growth factor receptor 2 (HER2), mismatch repair, programmed death-ligand 1 (PD-L1), claudin-18 isoform 2]. A sample was deemed FGFR2b-positive when any (>0%) tumor cells exhibiting moderate (2+) or strong (3+) membrane staining were detected ('any 2+/3+'). Other biomarkers were assessed based on current clinical guidelines. FGFR2b-stratified real-world overall survival (OS) was estimated using the Kaplan-Meier and Cox proportional regression models.

Results: Of 547 GC/GEJC patients identified, 492 (89.9%) met inclusion/exclusion criteria, had evaluable FGFR2b staining, and had complete clinical data. Estimated prevalence of FGFR2b any 2+/3+ was 15.4% [95% confidence interval (CI) 12.4% to 19.0%] in the full cohort of patients and 29.8% (95% CI 22.0% to 38.7%) in patients with samples collected within 1.5 years of study initiation. The majority (53.9%; 95% CI 42.1% to 65.5%) of FGFR2b any 2+/3+ tumor specimens were negative for other assessed biomarkers at a PD-L1 cut-off of combined positive score ≥5. In HER2-negative patients treated in 1L with chemotherapy alone, median OS was 11.5 months (95% CI 10.0-16.3 months) and 15.3 months (95% CI 13.0-16.8 months) for FGFR2b any 2+/3+ and FGFR2b 0/1+, respectively. There was no association between FGFR2b overexpression level and OS [adjusted hazard ratio (HR) 1.14, 95% CI 0.84-1.55].

Conclusions: This study revealed limited overlap of FGFR2b overexpression with currently actionable biomarkers, suggesting FGFR2b is a novel biomarker that identifies a distinct GC/GEJC patient population who may benefit most from an FGFR2b-targeting therapy.