Related Experiment Video
Updated: Apr 23, 2026

Predictive Immune Modeling of Solid Tumors
Published on: February 25, 2020
Inflammatory biomarkers refine progression risk stratification in NSCLC patients with stable disease
Kleinberger M1, Laengle S2, Berger Jm1
1Department of Medicine I, Division of Oncology, Medical University of Vienna, Austria; Christian Doppler Laboratory for Personalized Immunotherapy, Department of Medicine I, Medical University of Vienna, Austria.
Introduction:
Early risk stratification of non-small cell lung cancer (NSCLC) patients with stable disease (SD) at first restaging is particularly challenging. We explored the prognostic value of clinical and inflammatory markers in this population.
Methods And Material:
We analysed a real-world cohort of prospectively enrolled advanced NSCLC patients undergoing systemic intravenous anticancer treatment in a palliative intent at the Medical University of Vienna between 2019 and 2024. Inflammatory blood markers were measured at baseline and first restaging, with blinded radiologic assessment. Uni- and multivariable logistic regression models evaluated associations with durable clinical benefit (DCB).
Results:
Eighty NSCLC patients with SD at first restaging were included (median age 65 years, 50% female). Of those, 41 (51.3%) achieved DCB. Baseline characteristics were largely comparable. Patients with DCB had lower baseline neutrophil-to-lymphocyte and lymphocyte-to-leukocyte ratios. At first follow-up, CRP was lower and albumin higher in patients with DCB. In univariable analysis, lower follow-up albumin and higher LDH were associated with reduced odds of DCB. In multivariable models, PD-L1 positivity and follow-up albumin remained associated with DCB. The combined clinical-inflammatory model showed the highest apparent discriminative performance (AUC 0.766), compared to clinical-only (AUC 0.657) and inflammatory-only models (AUC 0.727), although the incremental improvement was modest.
Discussion:
In patients with advanced NSCLC and SD at first restaging, inflammatory biomarkers were associated with additional discriminative information beyond clinical characteristics alone. A combined clinical-inflammatory model showed numerically higher discriminative performance; however, the improvement was modest.

