The prognostic value of the Naples prognostic score for patients with non-small-cell lung cancer

The Naples prognostic score (NPS) is an effective inflammatory and nutritional scoring system widely applied as a prognostic factor in various cancers. We aimed to analyze the prognostic value of the NPS in patients diagnosed with non-small-cell lung cancer (NSCLC). We prospectively collected 395 patients diagnosed with NSCLC between January 2016 and December 2018 in two university-affiliated hospitals. Patients were divided into three groups according to their pretreatment NPS (Group 0: NPS = 0; Group 1: NPS = 1–2; Group 2: NPS = 3–4). Kaplan–Meier survival curves indicated that patients with higher NPS had a poorer overall survival (OS) and progress-free survival (PFS) (both P < 0.05). NPS was further confirmed as an independent prognostic factors of OS and PFS by multivariable survival analysis (both P < 0.05). Furthermore, stratifying by TNM stage, NPS also has significant predictive performance for OS and PFS in both early (I–IIIA) and advanced (IIIB–IV) stage NSCLC (all P < 0.05). The time-dependent receiver operating characteristic curve analysis demonstrated that NPS was more superior to other prognostic factors in predicting OS and PFS. In conclusion, NPS may serve as an effective indicator to predict OS and PFS in NSCLC patients regardless of TNM stage.

. Calculation of the NPS. LMR lymphocyte-to-monocyte ratio, NLR neutrophil-to-lymphocyte ratio, NPS Naples prognostic score. www.nature.com/scientificreports/ surgery, while we also recognized that these analyses are subject to limited statistical power. Interactions were evaluated using the Wald test. Time-dependent receiver operating characteristic curves for the prognostic values of NPS, PNI, CONUT and SIS were estimated using the R package survivalROC. All analyses were performed by IBM SPSS 25.0 and R software version 3.6.3. Two-sided P value < 0.05 was considered statistically significant.
OS and PFS based on the NPS. Of the 395 patients, the mean OS time were 37.1, 33.3 and 26.7 months in group 0, 1 and 2, respectively. Three-year OS rates of group 0, 1 and 2 were 73.9%, 58.5% and 38.9% (P < 0.001, log-rank test), respectively. Additionally, the mean PFS time were 34.0, 26.1 and 20.2 months in group 0, 1 and 2, respectively. Three-year PFS rates of group 0, 1 and 2 were 69.6%, 44.9% and 26.5% (P < 0.001, log-rank test), respectively. The Kaplan-Meier analysis showed that both OS and PFS of patients in group 0 was significantly higher than those of patients in group 1 and group 2 (Fig. 1A,B). We further divided the whole group into stage I-IIIA and stage IIIB-IV subgroups, significant difference was also observed in both OS and PFS based on the NPS regardless of the stage subgroups ( Fig. 1C-F).

Univariable and multivariable analyses of OS and PFS.
We assessed the NPS and other possible prognostic factors (age, gender, smoking, drinking, COPD, family history of cancer, BMI, TNM stage, histology, lesion, laterality, surgery, chemotherapy, radiotherapy, targeted therapy, CEA, NSE, albumin, total cholesterol, NLR, LMR, PNI, CONUT and SIS), to predict the OS and PFS of NSCLC patients in present study.
Prognostic value of NPS. We compared the prognostic value of the NPS with other scoring systems (PNI, CONUT and SIS). The time-dependent ROC curves showed that NPS obtained higher AUCs in dynamic trends compared with other variables within the follow-up time. The AUC of the NPS, PNI, CONUT, and SIS for predicting 3-year OS were 0.703, 0.606, 0.575 and 0.596, while the corresponding AUCs for predicting three-year PFS were 0.681, 0.597, 0.558 and 0.600, indicating that NPS was superior to other scoring systems for predicting long-term survival (Fig. 2). We further compared their prognostic value in different stage subgroups, and NPS remained a higher prognostic performance in both stage I-IIIA and stage IIIB-IV patients (see Supplementary  Fig. S1

Discussion
In the present study, we evaluated the potential importance of NPS as a pre-treatment prognostic indicator for OS and PFS among patients with NSCLC. We identified that NPS had discrimination for long-term survival in NSCLC patients, as high NPS was associated with poor OS and PFS regardless of the patients' TNM stage. Furthermore, NPS showed more superior prognostic value than previous scoring systems (PNI, CONUT, and SIS) for the prediction of long-term OS and PFS. However, the correlation between NPS and survival in NSCLC may be influenced by smoking status. Inflammation and nutritional status are closely related to the occurrence and development of cancers 8,[22][23][24][25] . And an increasing number of studies have explored various inflammatory and nutritional prognostic indicators and scoring systems. Inflammatory indicators such as NLR, LMR and PLR [26][27][28][29] , and nutritional indicators such as albumin and total cholesterol have been used to estimate the prognosis of various types of cancers, including lung cancer 12,13 . However, a single marker of inflammation or nutrition cannot fully represent the general status of cancer patients. Furthermore, several studies have also suggested that albumin, total cholesterol, NLR and LMR hold no prognostic value for certain groups of cancer patients [30][31][32][33] . In our study, none of the four parameters could be considered as an independent prognostic factor for OS, and only total cholesterol had significant prognostic value for PFS in NSCLC, supporting that the single marker with a limited predictive ability cannot be applied to a wide variety of malignancies. Thus, in recent years, various scoring system integrating one or more above parameters, such as PNI, CONUT and SIS [14][15][16] , are used to the prognostic assessment of patients with NSCLC. However, their performances of prognosis are still uncertain.
NPS, which is a novel scoring system based on serum albumin, total cholesterol, NLR and LMR, may more comprehensively reflect the host inflammatory and nutritional status. Recent studies have suggested that an elevated level of NPS is related with poor prognosis in patients with colorectal cancer, endometrial cancer or pancreatic cancer [17][18][19] . And a study reported the NPS was of great prognostic significance in stage I-II NSCLC patients who underwent completely VATS lobectomy 34 . However, the association between the NPS and advanced NSCLC patients remains unclear. Therefore, we focused on the prognostic value of NPS in NSCLC patients with www.nature.com/scientificreports/ different TNM stages. To our knowledge, this is the first study to completely investigate the association between pretreatment NPS and prognosis in NSCLC patients. Our multivariable analysis suggested that NPS was an independent prognostic factor for OS and PFS in NSCLC patients. And compared with a single indicator (including albumin, total cholesterol, NLR and LMR) and previous scoring system (including PNI, CONUT and SIS), the NPS turned to be more reliable for the prognosis of NSCLC. Considering that the TNM staging system is the main reference of the prognostic assessment in NSCLC patients, and that patients at the same TNM stage may still have different clinical outcomes, other factors are needed to further identify the prognosis for patients in the same TNM stage. We further performed a subgroup analysis of NPS and NSCLC survival according to different TNM stages and showed that NPS was an independent prognostic factor for OS and PFS in early (I-IIIA) stage patients. This was consistent with findings from the study by Li et al. 34 . Besides, our findings suggested that NPS could be also applied to predict the prognosis of advanced (IIIB-IV) stage NSCLC.
In addition, our results demonstrated that the effect of NPS on prognosis differed according to smoking status, particularly among former/current smokers. Several studies have indicated that tobacco smoke activates epithelial cells and macrophages and induces pulmonary inflammation and secretion of inflammatory cytokine and chemokine, which promotes lung cancer cell proliferation and migration 35,36 . Besides, a recent research suggested that previous or current waterpipe smoking was associated with changes in lipid metabolism, including elevated LDL-C and total cholesterol levels 37 , indicating that smoking may disturb the balance of nutrition and immune response. Further studies are needed to evaluate the association between smoking and inflammatory Table 3. Univariable and multivariable analysis of overall survival in NSCLC patients. AC adenocarcinoma, BMI body mass index, CEA carcinoembryonic antigen, CI confidence interval, CONUT controlling nutritional status, COPD chronic obstructive pulmonary emphysema, HR hazard ratio, LMR lymphocyte-to-monocyte ratio, NLR neutrophil-to-lymphocyte ratio, NPS Naples prognostic score, NSE neuron-specific enolase, PNI prognostic nutritional index, SCC squamous cell carcinoma, SIS system inflammation score. www.nature.com/scientificreports/ and nutritional status and how their potential interaction may affect the survival of patients with smokingrelated cancers. The mechanism by which high NPS contributes to a poor prognosis in patients with NSCLC is unclear. Neutrophils are recruited to tumor sites via chemokines and cytokines, and they subsequently release factors that remodel the extracellular matrix in the tumor microenvironment or act directly on tumor cells to enhance tumor proliferation and invasion 38 . Monocytes can differentiate into tumor-associated macrophages that increase angiogenesis, enhance tumor cell mobility and invasiveness, and therefore accelerate tumor cell intravasation and systemic tumor cell dissemination 39 . Unlike neutrophils and monocytes, lymphocytes have a central role in antitumor immunity by inducing cytotoxic cell death and inhibiting tumor cell proliferation and migration 40 . And lymphopenia may contribute to aggressive tumor biology, cancer progression, and a poor prognosis 41 . Albumin, a serum protein with greatest abundance, is responsible for maintaining colloid osmotic pressure and may influence microvascular integrity and aspects of the inflammatory pathway. Thus, the reduction of serum albumin is a sign of both inflammation and malnutrition 42 . Cholesterol is an important lipid for maintaining cellular homeostasis. Low cholesterol may impair function of the immune system, increase susceptibility to oxidative stress, induce production of cancer-related inflammation factors 43 . A recent study also suggests that a high serum cholesterol level reinforces the anti-tumor ability of NK cells and thus protects against cancer progression 44 . Thus, an elevated Table 4. Univariable and multivariable analysis of progression-free survival in NSCLC patients. AC adenocarcinoma, BMI body mass index, CEA carcinoembryonic antigen, CI confidence interval, CONUT controlling nutritional status, COPD chronic obstructive pulmonary emphysema, HR hazard ratio, LMR lymphocyte-to-monocyte ratio, NLR neutrophil-to-lymphocyte ratio, NPS Naples prognostic score, NSE neuron-specific enolase, PNI prognostic nutritional index, SCC squamous cell carcinoma, SIS system inflammation score. www.nature.com/scientificreports/ NPS, which is characterized by low levels of albumin, total cholesterol and NLR while a high level of LMR, is likely a reflector of more serious inflammation and a poorer nutrition status in patients. This study has several strengths, including the prospective design to minimize the potential for reverse causality. Furthermore, to the best of our knowledge, this is the first study to explore of the relationship between pretreatment NPS and NSCLC survival. Nevertheless, several limitations should be acknowledged in current study. First, the sample size was relatively small. There were only 46 patients in NPS Group 0, resulting in limited or no outcome events in several subgroups. Thus, the findings of the present study may need to be generalized with caution, and further studies with larger sample sizes are needed to precisely validate these results. Second, there was a lack of complete data collections regarding functional status (e.g., Karnofsky and ECOG performance status scoring) and gene mutation (e.g., EGFR, KRAS and ALK) that may influence tumor progression. Therefore, the influence of above potential confounders or effect modifier could not be excluded and studies integrating these indicators could contribute to further understanding of the prognostic factors for NSCLC patients.

Conclusion
Our findings showed that a higher NPS was substantially associated with poorer survival among NSCLC patients. Further studies are needed to validate our findings and to further explore the potential role of smoking in inflammation-and nutrition-mediated prognosis among NSCLC patients. Table 5. Subgroup analysis of overall survival and progression-free survival in NSCLC patients. *Adjusted for age, gender, smoking, drinking, COPD, family history of cancer, BMI, TNM stage, histology, lesion, laterality, surgery, chemotherapy, radiotherapy, targeted therapy, CEA, NSE, albumin, total cholesterol, NLR, LMR, PNI, COUNT and SIS and the corresponding variable was removed from the models when it was a stratified factor. AC adenocarcinoma, BMI body mass index, CEA carcinoembryonic antigen, NSE neuron-specific enolase, SCC squamous cell carcinoma.

Data availability
The data that support the findings of this study are available on request from the corresponding authors. The data are not publicly available due to privacy or ethical restrictions. www.nature.com/scientificreports/