**Background:** Systemic inflammation is recognized as a hallmark of cancer contributing to tumor development and progression. The Naples prognostic score (NPS), based on serum albumin, total cholesterol, neutrophil-to-lymphocyte ratio (NLR), and lymphocyte-to-monocyte ratio (LMR), was previously established as a prognostic indicator for colorectal cancer (CRC). However, confounding factors such as distant metastasis and baseline characteristic differences needed to be minimized to validate its predictive value. This study aimed to examine the predictive value of NPS for survival in CRC patients undergoing curative resection using propensity score matching (PSM) analysis, and to compare NPS against other inflammation-based and/or nutritional markers (NLR, LMR, PLR, PNI, CONUT).
**Methods:** This retrospective study enrolled 533 consecutive CRC patients who underwent curative surgical resection at Osaka Metropolitan University Hospital from January 2008 to December 2016. Patients with pathological Stage 0 or IV, non-curative resection, neoadjuvant therapy, synchronous other cancer, or atypical histology were excluded. NPS was scored based on four parameters: serum albumin (<4.0 g/dL = 1 point), total cholesterol (≤180 mg/dL = 1 point), NLR (>2.96 = 1 point), and LMR (≤4.44 = 1 point). Patients were classified as low NPS (groups 0 and 1; score 0–2) or high NPS (group 2; score 3–4). PSM was performed 1:1 using nearest neighbor matching with a caliper width of 0.2 SD, matching on age, sex, BMI, CCI, tumor location, histological type, pT and pN stage. Overall survival (OS) and disease-free survival (DFS) were compared using Kaplan–Meier curves and log-rank tests. Multivariable Cox proportional hazards regression identified independent prognostic factors. Time-dependent ROC analysis compared AUCs for 5-year OS across markers.
**Key Results:** Before PSM, 400 patients (75.0%) were low NPS and 133 (25.0%) were high NPS. After PSM, 123 patients were included in each group (note: text states 127 in one location and 123 in another; the matched analysis tables report n=123 per group). High NPS was significantly associated with older age (p<0.001), male sex (p=0.033), low BMI (p=0.047), high CCI (p<0.001), high ASA-PS (p=0.001), rectal cancer (p=0.025), undifferentiated histology (p=0.008), advanced pT stage (p<0.001), advanced pStage (p=0.002), and open surgery (p=0.023) before matching. After PSM, baseline characteristics were well balanced. The median follow-up was 63 months (IQR 51–75). The 5-year OS and DFS for the overall population were 82.6% and 85.9%, respectively; after PSM, they were 74.4% and 81.9%. Before PSM, OS and DFS were significantly lower in the high-NPS group (p<0.001 and p=0.017, respectively). After PSM, OS remained significantly lower in the high-NPS group (p=0.003), while DFS showed a trend but did not reach statistical significance (p=0.058). In subgroup analysis by pStage, high NPS was significantly associated with worse OS in pStage I (p<0.001), II (p=0.007), and III (p=0.034). After PSM, multivariate analysis identified high NPS as independently associated with OS (HR 2.23, 95% CI 1.27–3.89, p=0.005) and DFS (HR 1.97, 95% CI 1.02–3.78, p=0.043). Other independent factors for OS after PSM were ASA-PS ≥3 (HR 2.25, 95% CI 1.27–3.98, p=0.005) and undifferentiated tumor type (HR 5.32, 95% CI 2.28–12.4, p<0.001). For DFS after PSM, advanced T stage ≥3 (HR 3.07, 95% CI 1.02–9.27, p=0.047) and N stage ≥1 (HR 3.44, 95% CI 1.59–7.34, p=0.002) were also independent predictors. In the time-dependent ROC analysis, NPS had the greatest AUC for 5-year OS (0.643), compared to PNI (0.637), LMR (0.622), CONUT (0.601), NLR (0.579), and PLR (0.511).
**Clinical Implications:** This study validates the NPS as a useful prognostic indicator for CRC patients undergoing curative resection, demonstrating that it independently predicts both OS and DFS even after rigorous propensity score matching to minimize selection bias. The NPS outperformed other commonly used inflammation-based and nutritional markers (NLR, LMR, PLR, PNI, CONUT) in predicting 5-year OS. The finding that high NPS was significantly associated with worse OS across all pStages (I, II, and III) suggests that NPS could help refine risk stratification beyond TNM staging alone. This may aid in personalizing postoperative follow-up plans and adjuvant chemotherapy decisions. However, limitations include the single-center retrospective design, lack of genetic status data (KRAS, BRAF, MSI), and enrollment of only Japanese patients, necessitating further large-scale external validation in diverse populations.