相馬 智英

BACKGROUND: Patients with thoracic malignancy and interstitial pneumonia (IP) are often excluded from clinical trials, consequently lacking quantitative evidence of poorer prognosis and lower programmed death-ligand 1 (PD-L1) testing rates. METHODS: We evaluated the real-world impact of comorbid IP on biomarker adoption and survival in thoracic malignancy patients receiving first-line systemic therapy at a tertiary teaching hospital between 2016 and 2023. RESULTS: Among 1247 patients, 98 (7.5%) had comorbid IP. Multigene testing rates in IP patients were similar to those in non-IP patients. Only three actionable genomic alterations were found in the IP group, highlighting PD-L1 testing as the key element. PD-L1 testing was underutilized in the IP group (63.3%) compared with the non-IP group (75.1%). Immune checkpoint inhibitor (ICI) therapy was utilized in 12.2% of IP versus 29.3% in non-IP, despite comparable clinical situations. Comorbid IP predicted worse survival (hazard ratio: 1.789; 95% confidence interval: 1.373-2.331; p < 0.001). Although survival significantly improved in non-IP after 2020, no benefit was observed in IP. A multivariable model incorporating an IP × Period interaction confirmed comorbid IP remained a negative prognostic factor, highlighting recent advances have not bridged the survival disparity for this high-risk group. CONCLUSIONS: Despite recent progress, patients with comorbid IP experience limited clinical benefit, characterized by lower rates of PD-L1 testing, restricted use of immune checkpoint inhibitors, and absence of post-2020 survival gains. This large-scale and quantitative evidence demonstrates persistent disparities and their prognostic significance, reflecting the limited applicability of current immunotherapy-based strategies in this high-risk population.