The predictive value of modified-Naples prognostic score for patients with locally advanced non-small cell lung cancer undergoing surgery after neoadjuvant chemotherapy.

Introduction
In cancer statistics, lung cancer has the highest mortality rate. Non-small cell lung cancer (NSCLC) accounts for up to 85% of all lung cancers (1, 2). Early-stage NSCLC, due to the lack of obvious symptoms, misses the best treatment time and is often discovered at a locally advanced or even metastatic stage (3, 4). Patients with locally advanced NSCLC often exhibit significant tumor heterogeneity. This heterogeneity is associated with an unfavorable prognosis and a dismal five-year survival rate (5, 6). According to current lung cancer treatment guidelines, locally advanced NSCLC patients could accept surgical treatment after neoadjuvant therapy (7, 8). Therefore, there is a pressing need to stratify advanced NSCLC patients to identify robust prognostic biomarkers. Previous studies have demonstrated that single inflammatory or nutritional markers have predictive value for prognosis. However, single markers often have limitations and instability. Therefore, a comprehensive predictive index combining multiple markers may have higher predictive value (9, 10). Against this backdrop, the Naples Prognostic Score (NPS), which incorporates systemic inflammation and nutritional status, was proposed (11, 12).
Currently, NPS has been confirmed to hold independent prognostic value for cancers affecting the lung, rectum, endometrium, breast, esophagus, and pancreas (13–18). NPS has emerged as an independent and practical prognostic tool in NSCLC (19). In our previous study, NPS was further validated in a cohort of locally advanced NSCLC patients who received surgical resection after neoadjuvant chemotherapy, confirming that it served as an independent predictor for both progression in free survival (PFS) and overall survival (OS) (20). The modified Naples Prognostic Score (mNPS) was composed of the optimal cut-points of NPS reconfigured using X-tile software (21). Current studies have indicated that mNPS re-established based on relevant research data is valuable for the prognosis of oral squamous cell carcinoma and colon cancer (22, 23). Presently, there is an absence of robust evidence regarding the prognostic role of mNPS following surgical resection for locally advanced NSCLC.
Hence, our work sought to delineate mNPS and clinicopathological variables and to determine its prognostic value in locally advanced NSCLC. Furthermore, the mNPS was compared with the NPS to evaluate whether it has a higher predictive value.

Data and methods

Clinical data
The 126 patients included in this study were all from the Second Affiliated Hospital of Shandong Medical University (September 2012 - April 2019). Inclusion criteria: (1) Patients were over 18 years old; (2) Patients were diagnosed with NSCLC by histopathology; (3) Patients had a Karnofsky Performance Status (KPS) score of 80–100; (4) Patients had not received anti-tumor treatment before admission; (5) Peripheral blood tests had been completed one week before treatment, including albumin, cholesterol, neutrophils, monocytes, lymphocytes, and tumor markers; (6) Patients did not have any major internal medical conditions and must met the indications for chemotherapy and surgery as confirmed jointly by oncologists and thoracic surgeons; (7) The patients and their families agreed to the surgery and preoperative neoadjuvant chemotherapy, and already signed informed consent forms. Exclusion criteria: (1) Concurrent malignant tumors; (2) Previous use of antibiotics or nonsteroidal anti-inflammatory drugs; (3) Presence of chronic or active infectious diseases.

Data collection
We have collected the following information: gender, age, smoking status, pathological type, tumor location, clinical stage, degree of differentiation, KPS score, CEA level, chemotherapy regimen, surgical procedure, postoperative complications, PFS, OS, as well as albumin, cholesterol, neutrophil, monocyte counts, and lymphocyte was measured one week before neoadjuvant chemotherapy. There were no missing data for any of the patients included in the analysis.

Treatment methods
According to current lung Cancer treatment guidelines, treatment plans were developed for all patients receiving chemotherapy through multidisciplinary discussion and evaluation by senior oncologists. Specific chemotherapy regimens included: pemetrexed combined with platinum-based drugs, paclitaxel combined with platinum-based drugs, gemcitabine combined with platinum-based drugs, and pemetrexed or paclitaxel combined with platinum-based chemotherapy. Specific dosages were determined based on the patient’s tolerance, tumor status, and the dosage and duration of chemotherapy drugs.
Patients rested for two weeks after chemotherapy. If there were no contraindications, surgery could then proceed; if contraindications existed, the rest period was extended until surgical indications were met. The specific surgical approach depended on the patient’s lesion characteristics; thoracoscopic surgery or open thoracotomy was selected. The choice between the two surgical approaches was made after a comprehensive evaluation by at least three thoracic surgeons to ensure surgical safety and minimize patient injury.

Follow-up
All enrolled patients underwent regular follow-up. Follow-up was primarily conducted via telephone or outpatient examinations. Follow-up assessments were scheduled at three-month intervals during the initial three-year period, transitioning to six-month intervals for the subsequent six years. Scheduled follow-up included physical examinations, imaging studies, and laboratory tests.
OS was measured from the date of initial treatment to the date of all-cause death or the last known follow-up. PFS was assessed as the time from treatment initiation to the first evidence of disease progression, including local recurrence, metastatic spread, as well as death.

Construction of mNPS
MNPS was calculated based on cholesterol, albumin, NLR, and LMR. Data analysis was conducted with X-tile 3.6.1 (Yale University, New Haven, CT, USA), aiming at establishing cohort-specific optimal cutoffs for cholesterol, albumin, NLR, and LMR. (Figure 1) The modified Naples prognostic score was constructed as follows: albumin concentration >41.8 g/L = 0 points, ≤41.8 g/L = 1 point; cholesterol concentration >137.5 mg/dL = 0 points, ≤137.5 mg/dL = 1 point; NLR ≤2.1 = 0 points, >2.1 = 1 point; LMR >4.8 = 0 points, ≤4.8 = 1 point. The total score of the four indicators was the mNPS. Group 0 had an mNPS score of 0, Group 1 had scores of either1 or 2, and Group 2 had scores of 3 or 4.

Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics, Version 27.0 (SPSS Inc., Chicago, IL). The predictive accuracy of the mNPS, albumin, cholesterol, NLR, and LMR was evaluated using receiver operating characteristic (ROC) methodology. Survival outcomes across Groups were compared using the Kaplan-Meier method with comparison by the log-rank test. Univariate and multivariate Cox proportional hazards models were employed to identify candidate and delineate independent prognostic factors, respectively, with results expressed as hazard ratios (HR) and corresponding 95% confidence intervals (95% CI). Proportional hazards assumptions for all Cox regression models were verified using Schoenfeld residuals.

Results

Clinical characteristics
126 patients were included through strict inclusion and exclusion criteria. All patients received neoadjuvant chemotherapy prior to surgery and underwent regular follow-up (Figure 2). 66 patients (52.4%) were female, and 60 patients (47.6%) were male. 67 patients (53.2%) were aged <60 years, and 59 patients (46.8%) were aged ≥60 years. 52.4% of patients smoked. 82 patients (65.1%) were clinically stage IIIA, and 44 patients (34.9%) were stage IIIB. The majority of patients (98, 77.8%) recorded a KPS score of 100, with the remaining 28 (22.2%) scoring between 80 and 90. Of the 72 patients (57.1%), the pathological type was adenocarcinoma, and 54 patients (42.9%) were squamous cell carcinoma. 56 patients (44.4%) underwent open thoracotomy, and 70 patients (55.6%) underwent VAST surgery. 59 patients (46.8%) received neoadjuvant chemotherapy with paclitaxel+ platinum-based drugs, 30 patients (23.8%) received pemetrexed+ platinum-based drugs, and 37 patients (29.4%) received gemcitabine+ platinum-based drugs. 62 patients (49.2%) experienced postoperative complications.

Relationship between mNPS and clinicopathological features
Based on the mNPS criteria, Group 0 (0 points) comprised 20 patients (15.9%); Group 1 (1–2 points) comprised 85 patients (67.5%); and Group 2 (3–4 points) comprised 21 patients (16.7%). MNPS showed no statistically significant association with any clinicopathological characteristic (all P > 0.05) (Table 1).

Survival difference analysis
The predictive performance of mNPS, albumin, cholesterol, LMR, and NLR was assessed via ROC curve analysis. Table 2 lists the area under the curve (AUC) for mNPS, albumin, cholesterol, NLR, and LMR. The results indicate that mNPS had greater predictive value for OS (AUC = 0.640, P = 0.007) and PFS (AUC = 0.623, P = 0.024) than albumin, cholesterol, NLR, and LMR. The ROC curves clearly demonstrated the predictive ability of mNPS (Figure 3). Figure 4 shows the OS and PFS survival curves for the three Groups. For all enrolled patients, the OS and PFS of Group 0 were superior to those of Groups 1 and 2 (OS: Group 1 vs Group 0, P = 0.038; Group 2 vs Group 0, P = 0.001; PFS: Group 1 vs Group 0, P = 0.005; Group 2 vs Group 0, P = 0.004).

Single-factor and multi-factor analysis
Table 3 presented the results of the univariate Cox proportional hazards regression analysis for clinicopathological features, mNPS, and their association with OS and PFS. Univariate analysis showed that surgical method (P = 0.042), clinical stage (P = 0.004), and KPS score (P = 0.003) were significantly associated with OS. Conversely, smoking history (P = 0.744), gender (P = 0.506), age (P = 0.460), tumor location (P = 0.419), pathological type (P = 0.969), differentiation degree (P = 0.323), CEA level (P = 0.806), and mNPS (Group 1 vs Group 0, P = 0.051; Group 2 vs Group 0, P = 0.001) were not statistically associated with OS. KPS score (P = 0.002), clinical stage (P = 0.005), and mNPS (Group 1 vs Group 0, P = 0.010; Group 2 vs Group 0, P = 0.002) were significantly associated with PFS. Other clinicopathological factors were not statistically significant with PFS. Indicators significant in the univariate analysis of PFS were included in the multivariate analysis. The multivariate analysis results showed that mNPS score (Group 1 vs Group 0, P = 0.038; Group 2 vs Group 0, P = 0.008), clinical stage (P = 0.022), and KPS score (P = 0.017) were independent prognostic factors for PFS (Table 4).

Discussion
The tumor microenvironment, as one of the hallmarks of cancer, is the internal environment for the generation and survival of tumor cells, which includes immune and inflammatory cells (24–26). Tumors weaken anti-tumor immune responses through the tumor microenvironment, thereby maintaining proliferation, evading apoptosis, and preserving inflammatory environments and angiogenesis (26, 27). Therefore, tumor prognosis is closely related to inflammation and immunity (28, 29). In addition, nutritional status also affects tolerance to the drugs used in neoadjuvant therapy and postoperative recovery (30, 31). Studies have validated the prognostic value of inflammatory, immunological, and nutritional indicators in cancer (32–34). Research has further extended to composite scores that integrate these markers, demonstrating utility beyond that of individual utility (35, 36). The NPS is composed of albumin, cholesterol, LMR, and NLR (37). It has been confirmed to have prognostic value for multiple tumors (38–42). Previous research has shown that NPS demonstrates superior prognostic performance for resected locally advanced NSCLC patients undergoing neoadjuvant chemotherapy and surgery (20, 43). Based on the cohort of this study, the cut-off values of the four indicators were used to construct the mNPS. This study aimed to evaluate the prognostic significance of mNPS in locally advanced NSCLC. We found that mNPS was an independent prognostic factor for PFS in patients with locally advanced non-small cell lung cancer undergoing surgery after neoadjuvant chemotherapy, although it was not an independent prognostic factor for OS. The observed significance of mNPS for PFS but not OS may be attributed to several factors, including limited sample size for long-term survival events, heterogeneity in postoperative systemic therapies, and the possibility that inflammatory–nutritional influences are more pronounced in earlier disease progression rather than long-term overall survival. Although the AUC value of mNPS(0.640 for OS, 0.623 for PFS) had relatively moderate statistical significance, it was superior to a single indicator such as albumin, cholesterol, NLR, and LMR. Therefore, mNPS still has a relatively high research prospect.

Limitations
This research has certain limitations. This study is a retrospective one, a small-sample study from a single center, and the results are greatly influenced by the sample size. Another limitation is the difference in treatment methods. All the patients with locally advanced NSCLC included in this study received neoadjuvant chemotherapy and surgical treatment, but there were differences in postoperative treatment methods. To some extent, it might affect the prediction of survival by mNPS.

Conclusion
Although mNPS was an independent prognostic factor for PFS in patients with locally advanced non-small cell lung cancer undergoing surgery after neoadjuvant chemotherapy, it had no statistical significance for OS. In comparison, NPS demonstrated superior prognostic performance for resected locally advanced NSCLC patients undergoing neoadjuvant chemotherapy and surgery.