OBJECTIVES: To assess the efficacy of the AuNP-enhanced ELISA in the detection of salivary TNF-α in OSCC and evaluate its predictive value for survival. DESIGN: A longitudinal study was conducted on 40 OSCC patients and 10 healthy controls. Saliva was collected at a regular interval and TNF alpha levels were measured using ELISA and AuNP-enhanced ELISA. Descriptive statistics were carried out for demographic, clinical, and biomarker data. The efficacy of ELISA and AuNP-enhanced ELISA was estimated and compared using the ROC curve. Kaplan-Meier survival analysis and Cox proportional hazards models were used to assess the survival outcome. A neural network model was performed to estimate TNF-α levels over time in OSCC patients. RESULTS: OSCC patients (ELISA is 47.52 ± 20.23 and Gold nano-enhanced ELISA is 57.63 ± 24.99) exhibited significantly elevated TNFα levels compared to controls (ELISA is 10.13 ± 3.07and Gold nano-enhanced ELISA is 12.07 ± 3.66). Gold nano-enhanced ELISA (AUC of 0.995) demonstrated superior sensitivity than ELISA (AUC of 0.986), while Gold nano-enhanced ELISA achieves an even higher in detecting elevated TNFα levels. Kaplan-Meier analysis shows that Gold nano-enhanced ELISA outperforms ELISA in capturing survival trends, with better survival for TNF-α levels above the cutoff at 9 months (70 % vs. 60 %) and 24 months (40 % vs. 0 %). The neural network model poorly predicted TNF- levels over the period (AUC = 0.447). CONCLUSIONS: The gold nano-enhanced TNFα detection method is effective in detecting TNFα levels between OSCC patients and controls, demonstrating superior sensitivity in identifying survival trends over time compared to traditional ELISA.