Prognostic impact of CEACAM5 in nonsquamous non-small cell lung cancer: a critical reappraisal driven by cut-off optimization

Background: Carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5) has emerged as a promising therapeutic target for antibody-drug conjugates (ADCs) in nonsquamous non-small cell lung cancer (NSCLC). However, the landscape of CEACAM5 protein expression and its association with clinicopathological features in nonsquamous NSCLC remain poorly characterized. This study represents the first comprehensive investigation to profile CEACAM5 protein expression in this specific population using a clinical-grade immunohistochemistry (IHC) assay.

Methods: We retrospectively analyzed 218 patients with resected nonsquamous NSCLC. A standardized IHC assay derived from a clinical trial was employed, utilizing a proprietary monoclonal antibody (clone 769) specifically developed for clinical application. We systematically evaluated the relationship between CEACAM5 expression and tumor characteristics using both a conventional therapeutic threshold (≥50%) and a sensitive data-driven cut-off (>0%) to explore the full spectrum of antigen expression.

Results: The assay demonstrated excellent reproducibility. We found that CEACAM5 expression was significantly associated with markers of tumor aggressiveness, including lymphovascular and pleural invasion. Notably, the sensitive >0% cut-off revealed broader biological correlations than the restrictive ≥50% threshold. While CEACAM5 was not an independent prognostic factor in the overall cohort, a stage-specific, hypothesis-generating analysis—supported by external genomic validation—suggested that positive CEACAM5 expression may be associated with a trend towards poorer disease-free survival specifically in early-stage (stage I–II) patients, although this did not reach statistical significance (P=0.06).

Conclusions: This study provides the first detailed characterization of CEACAM5 protein expression in nonsquamous NSCLC, establishing it as a biomarker linked to aggressive disease phenotypes. Our findings suggest that adopting a sensitive detection threshold (>0%) may better capture the population of patients with biologically active CEACAM5, thereby potentially refining candidate selection for targeted therapies and risk stratification in early-stage disease, a hypothesis that warrants prospective validation.