Eosinophil as a biomarker for diagnosis, prediction, and prognosis evaluation of severe checkpoint inhibitor pneumonitis

Introduction: Checkpoint inhibitor pneumonitis (CIP) is a common serious adverse event caused by immune checkpoint inhibitors (ICIs), and severe CIP can be life-threatening. We aimed to investigate the role of peripheral blood cells in diagnosis, prediction, and prognosis evaluation for all and severe CIP. Materials and methods: Patients with lung cancer receiving ICIs were enrolled in this retrospective study. Baseline was defined as the time of ICI initiation, endpoint was defined as the time of clinical diagnosis of CIP or the last ICI treatment, and follow-up point was defined as 1 week after CIP. Eosinophil percentages at baseline, endpoint, and follow-up point were shortened to “E_bas”, “E_end and “E_fol”, respectively. Results: Among 430 patients included, the incidence of CIP was 15.6%, and severe CIP was 3.7%. The E_end/E_bas value was lower in patients with CIP (p = 0.001), especially severe CIP (p = 0.036). Receiver operating characteristic curves revealed that E_end/E_bas could serve as a biomarker to diagnose CIP (p = 0.004) and severe CIP (p < 0.001). For severe CIP, the eosinophil percentage declined before the symptoms appeared and CT diagnosis. The eosinophil percentage significantly elevated at the follow-up point in the recovery group but not in the non-recovery group. The CIP patients with E_fol/E_bas ≥1.0 had significantly prolonged overall survival (p = 0.024) and after-CIP survival (AS) (p = 0.043). The same results were found in severe CIP but without a statistical difference. Conclusions: Eosinophil percentage was associated with the diagnosis, prediction, and prognosis of CIP and severe CIP.