Predictive value of serum tumor markers (carcinoembryonic antigen, neuron-specific enolase, and squamous cell carcinoma antigen in non-small cell lung cancer patients treated with programmed cell death protein 1 inhibitors.

Background: Adverse reactions (ARs) may occur in patients with advanced non-small cell lung cancer (ANSCLC) undergoing treatment with programmed cell death protein 1 (PD-1) inhibitors (PD-1Is). Establishing a risk assessment model can facilitate personalized treatment.

Methods: Clinical data were collected from 215 ANSCLC patients treated with PD-1Is. Patients who experienced ARs were classified as the observation group (OG, 92 cases), while those who did not experience ARs were classified as the control group (CG, 123 cases). A multivariable logistic regression (LR) model was employed to analyze independent risk factors (RFs) associated with ARs, and R Studio software was utilized to create a nomogram predictive model.

Results: The concordance index for the nomogram predictive model for ARs in ANSCLC patients treated with PD-1Is was 0.911. The threshold for predicting ARs using the nomogram was more significant than 0.25, providing a clinical net benefit superior to individual indicators such as smoking, tumour-node-metastasis (TNM) staging, neutrophil-to-lymphocyte ratio (NLR), systemic immune-inflammation index (SII), and prognostic nutritional index (PNI). The proportion of smokers in the OG was markedly superior to that in the CG (P<0.05).

Conclusions: Smoking, TNM staging, and peripheral blood indicators such as NLR, SII, and PNI are independent RFs for the occurrence of ARs. The constructed nomogram predictive model demonstrates greater clinical utility than individual indicators, enhancing the accuracy of AR predictions.