Clinical predictors of treatment efficacy and a prognostic nomogram in patients with lung adenocarcinoma receiving immune checkpoint inhibitors: a retrospective study

BACKGROUND: At present, there is no accurate biomarker for immune checkpoint inhibitors (ICIs). Since the efficacy of ICIs is associated with a variety of indicators, establishing a model to predict its efficacy is more clinically significant and in line with clinical needs. METHODS: We collected and retrospectively analyzed the relationship between immunotherapy efficacy and clinicopathologic features in lung adenocarcinoma patients treated with ICIs. Progression-free survival (PFS) and overall survival (OS) were analyzed. Univariate and multivariate Cox proportional hazards regression analyses were conducted to identify prognostic factors associated with PFS. Besides, a clinical prediction model was established based on the results of the multivariate Cox regression analyses to predict PFS. RESULTS: A total of 201 lung adenocarcinoma patients treated with ICIs were assessed. Univariate analysis showed that male gender [hazard ratio (HR) =0.521, 95% confidence interval (CI): 0.356–0.761, P=0.001], smoking (HR =0.595, 95% CI: 0.420–0.843, P=0.003), epidermal growth factor receptor (EGFR) wild type (HR =2.766, 95% CI: 1.719–4.452, P<0.001), Kirsten rat sarcoma viral oncogene homolog (KRAS) mutation (HR =0.449, 95% CI: 0.271–0.743, P=0.001), positive programmed death ligand 1 (PD-L1) expression (HR =0.527, 95% CI: 0.336–0.825, P=0.004), early tumor node metastasis (TNM) stage (HR =0.581, 95% CI: 0.344–0.983, P=0.039), no liver metastasis (HR =1.801, 95% CI: 1.046–3.102, P=0.031), ICIs combined with chemotherapy (HR =0.560, 95% CI: 0.384–0.815, P=0.002), having immune-related adverse effects (HR =0.354, 95% CI: 0.228–0.511, P<0.001) and first-line immunotherapy (HR =0.596, 95% CI: 0.420–0.845, P=0.003) were significantly associated with better PFS in patients with lung adenocarcinoma receiving immunotherapy. Multivariate analysis showed that smoking status, KRAS mutation, PD-L1 expression, line of immunotherapy and immune-related adverse effects were independent prognostic factors affecting PFS. A clinical prediction model was established to predict the PFS of lung adenocarcinoma patients treated with ICIs. The model showed good predictive ability via C-index, calibration curve and receiver operating characteristic (ROC) curve validation. CONCLUSIONS: The clinical prediction model developed in this study can be used to some extent to predict PFS after immunotherapy in lung adenocarcinoma patients. However, the model still needs to be validated in studies with large sample size.