# PD‐L1 Expression and Histopathological Features in EGFR‐Mutated Non‐Small Cell Lung Cancer: Implications for Immune Checkpoint Inhibitors After EGFR‐Tyrosine Kinase Inhibitors Resistance

**Authors:** Toshiyuki Sumi, Taiki Ishigooka, Keigo Matsuura, Takumi Ikeda, Yuichi Yamada, Kotomi Arioka, Hirofumi Chiba

PMC · DOI: 10.1111/1759-7714.70252 · Thoracic Cancer · 2026-02-09

## TL;DR

This study shows that high PD-L1 levels in EGFR-mutated lung cancer are linked to aggressive features and better outcomes with immune therapy after resistance to initial treatment.

## Contribution

The study introduces a time-varying survival analysis to show how PD-L1 status modifies ICI efficacy after EGFR-TKI resistance in NSCLC.

## Key findings

- High PD-L1 expression correlates with solid histology and larger tumors in EGFR-mutated NSCLC.
- ICI treatment after EGFR-TKI resistance shows better survival outcomes in PD-L1-high patients.
- PD-L1 status significantly interacts with ICI efficacy in post-progression survival analysis.

## Abstract

Epidermal growth factor receptor‐mutated (EGFRm) non‐small cell lung cancer (NSCLC) responds well to EGFR tyrosine kinase inhibitors (EGFR‐TKIs), yet optimal therapy after resistance remains uncertain. Although immune checkpoint inhibitors (ICIs) show limited overall efficacy, heterogeneity in response by programmed cell death ligand 1 (PD‐L1) expression exists.

We retrospectively evaluated 90 patients with advanced/recurrent EGFRm NSCLC treated with first‐line EGFR‐TKIs (October 2018–October 2023). Clinicopathological and radiologic features were compared by PD‐L1 tumor proportion score (< 50% vs. ≥ 50%). The primary analysis evaluated overall survival from first progression (post‐progression OS/PPS) using a time‐varying Cox model with start–stop intervals, modeling ICI as a time‐dependent exposure ICI(t) and testing the ICI(t) × PD‐L1 interaction.

The PD‐L1 ≥ 50% group (n = 26) more often had solid histology (62% vs. 8%), solid nodules on computed tomography (96% vs. 52%), and larger tumors (median 42.0 vs. 27.5 mm). Among 68 patients evaluable from t
0, the ICI(t) × PD‐L1 interaction was significant (Wald p = 0.015; likelihood‐ratio p = 0.036). Stratum‐specific adjusted ICI effects suggested a detrimental association in PD‐L1 < 50% (hazard ratio [HR]: 3.11, 95% confidence interval [CI]: 0.94–10.30) but favorable association in PD‐L1 ≥ 50% (HR: 0.457, 95% CI: 0.138–1.512). In exploratory analyses of 22 ICI‐treated patients, PD‐L1 ≥ 50% showed higher response rates (64% vs. 9%) and longer time to treatment failure (3.4 vs. 1.4 months).

High PD‐L1 expression in EGFRm NSCLC is associated with more aggressive morphologic features and modifies the association between post‐progression ICI and survival. These findings support PD‐L1‐informed selection after EGFR‐TKI failure, while prospective confirmation is needed.

This retrospective study elucidates the clinical significance of PD‐L1 expression in EGFR‐mutated non‐small cell lung cancer. High PD‐L1 expression (TPS ≥ 50%) was significantly associated with solid histological patterns and radiological solid nodules. Furthermore, a time‐varying survival analysis revealed that PD‐L1 status modifies the efficacy of immune checkpoint inhibitors (ICIs) administered after EGFR‐TKI resistance. The findings suggest that patients with high PD‐L1 expression may derive survival benefits from post‐progression ICI therapy compared to the PD‐L1‐low population.

## Linked entities

- **Genes:** EGFR (epidermal growth factor receptor) [NCBI Gene 1956]
- **Proteins:** CD274 (CD274 molecule)
- **Diseases:** non-small cell lung cancer (MONDO:0005233)

## Full-text entities

- **Genes:** CD274 (CD274 molecule) [NCBI Gene 29126] {aka ADMIO5, B7-H, B7H1, PD-L1, PDCD1L1, PDCD1LG1}, EGFR (epidermal growth factor receptor) [NCBI Gene 1956] {aka ERBB, ERBB1, ERRP, HER1, NISBD2, NNCIS}
- **Diseases:** NSCLC (MESH:D002289), tumor (MESH:D009369), PPS (MESH:C562509)
- **Species:** Homo sapiens (human, species) [taxon 9606]
- **Mutations:** start-stop

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12885618/full.md

## Figures

4 figures with captions in the complete paper: https://tomesphere.com/paper/PMC12885618/full.md

## References

29 references — full list in the complete paper: https://tomesphere.com/paper/PMC12885618/full.md

---
Source: https://tomesphere.com/paper/PMC12885618