Abstract
Objective
This study aimed to investigate the value of a novel composite index, the ALP score (calculated as the product of lymphocyte count and serum albumin), in predicting the risk of early progression to chemotherapy combined with immune checkpoint inhibitors (ICIs) in patients with driver mutation-negative advanced non-small cell lung cancer (NSCLC).
Methods
Clinical data from 433 patients with advanced NSCLC who received first-line chemoimmunotherapy were retrospectively collected. Independent risk factors for early progression (defined as progressive disease, PD) were identified using univariate and multivariate logistic regression analyses. The predictive power of the ALP score for PD was evaluated, and its nonlinear relationship with risk was explored using restricted cubic splines (RCS). Ten machine learning models and the SHAP interpretability technique were further employed to validate the importance of the ALP score.
Results
Multivariate analysis identified the absence of liver metastasis and the absence of bone metastasis as independent protective factors against early progression, whereas lower lymphocyte count, lower albumin level, and lower Body Mass Index were identified as independent risk factors. The ALP score predicted PD with an AUC of 0.643, outperforming its individual components. Critically, multivariable analyses confirmed the ALP score as an independent predictor for both endpoints: patients in the highest score tertile had significantly lower risks of early progression (OR = 0.33, P = 0.001) and disease progression or death (HR = 0.65, P = 0.002) compared to the lowest tertile. RCS analysis revealed a significant nonlinear inverse relationship between the ALP score and PD risk (P
Conclusion
The ALP score, a composite index derived from routine laboratory parameters, serves as an accessible and practical prognostic marker associated with early progression risk and progression-free survival in patients with advanced NSCLC. Reflecting baseline systemic immune and nutritional readiness, it offers a practical and low-cost tool for clinical risk stratification in patients receiving chemo-immunotherapy.