Abstract
BACKGROUND: In the treatment of non-small cell lung cancer (NSCLC), immune checkpoint inhibitors (ICIs) have markedly improved patient survival, yet some patients do not benefit. The existing prognostic factors are limited, highlighting the development of reliable and convenient predictive indicators. METHODS: A retrospective analysis was performed on 219 NSCLC patients treated with ICIs from June 2019 to January 2024. The nutritional risk screening (NRS 2002) and the neutrophil-to-lymphocyte ratio (NLR) were employed to evaluate the patients' nutritional status and inflammatory response, aiming to investigate the correlation between these markers and treatment outcomes. RESULTS: The median follow-up duration for the overall population was 29 (IQR: 25.96-32.04) months. The analysis showed that the median progression-free survival (mPFS) and median overall survival (mOS) in the high nutritional risk group (NRS2002 ≥ 3, accounting for 23.74%) were significantly lower than those in the low nutritional risk group (NRS2002 < 3, accounting for 76.26%) (mPFS: 2.5 vs 16 months; mOS: 8 vs 16 months, both P < 0.001). Similarly, patients with high NLR values (> 4.92) had significantly shorter OS and PFS than those with low NLR (≤ 4.92) values (mOS: 7 vs 18 months; mPFS: 3 vs 17 months, both P < 0.001). Multivariate Cox analysis revealed that a high NRS 2002 score (HR = 2.76, 95% CI 1.68-4.54, P < 0.001) and high NLR (HR = 2.77, 95% CI 1.65-4.64, P < 0.001) were independent predictors of poor prognosis. Risk stratification was performed using a combined scoring system of NRS 2002 and NLR (0 points-low risk, 1 point-moderate risk, 2 points-high risk), and it was found that as the risk score increased, OS and PFS significantly decreased (mOS: 8 [2.61-13.39] vs 16 [13.04-18.96] vs NA [NA-NA] months; mPFS: 2.5 [0.99-4.02] vs 8.5 [5.47-11.53] vs 16 [11.41-20.59] months, respectively, both P < 0.001). The utility of the combined NLR and NRS2002 scoring model was assessed using a time-dependent receiver operating characteristic (ROC) curve, with results indicating that at 12 months, the AUC value of the combined scoring model was 0.81 (CI 0.72-0.90). At 24 and 36 months, the AUC values were 0.73 (CI 0.66-0.80) and 0.70 (CI 0.64-0.76), respectively. Moreover, the nomogram model exhibited high predictive accuracy in predicting survival prognosis, with AUC values of 0.84 (CI 0.77-0.91), 0.85 (CI 0.79-0.91), and 0.78 (CI 0.69-0.88) at 12, 24, and 36 months, respectively. CONCLUSION: The combined NRS 2002 and NLR scoring can serve as an effective prognostic tool for NSCLC patients receiving ICIs treatment. This scoring system helps clinicians more accurately identify patients who will benefit from immunotherapy, thereby facilitating more personalized treatment plans. Further validation of this scoring system's applicability and reliability is warranted in future multicenter, large-sample prospective studies.