Background: There is a lack of effective treatments for recurrent or metastatic nasopharyngeal carcinoma (RM-NPC). Furthermore, the response rate of NPC patients to programmed death 1 (PD-1) inhibitors is approximately 20% to 30%. Thus, we aimed to explore reliable and minimally invasive prognostic indicators to predict the efficacy of PD-1 inhibitors combination therapy in RM-NPC. Methods: The serum markers of 160 RM-NPC patients were measured before and three weeks after the first anti-PD-1 treatment. The least absolute shrinkage and selection operator (LASSO) logistic regression was carried out to select dynamic serum indicators and construct a prediction model. Furthermore, we carried out univariate, multivariate, nomogram and survival analyses to identify independent prognostic factors that were associated with 1-year progression-free survival (PFS). Results: Based on two markers that were screened by Lasso logistic regression, we constructed a risk score prediction model for the prediction of anti-PD-1 efficacy at 8–12 weeks with an AUC of 0.737 in the training cohort and 0.723 in the validation cohort. Risk score and metastases were included in the nomogram, and the Kaplan–Meier survival curves demonstrated that the high-risk group has shorter PFS compared to the low-risk group. The concordance index (C-index) of the nomogram for PFS is higher than that of the TNM stage in the training and validation cohort. Conclusion: We proposed a strategy to monitor dynamic changes in the biochemistry markers and emphasized their importance as potential prognostic biomarkers for the treatment of advanced NPC treated with PD-1 inhibitors. Our risk score prediction model was based on the dynamic change of LDH and AST/ALT, which has predictive and prognostic value for NPC patients who were treated with PD-1 inhibitors.

Immune checkpoint blockade therapy targeting the programmed death-1(PD-1) pathway has shown remarkable efficacy and durable response in patients with various cancer types. Early prediction of therapeutic efficacy is important for optimizing treatment plans and avoiding potential side effects. In this work, we developed an efficient machine learning prediction method using routine hematologic and biochemical parameters to predict the efficacy of PD-1 combination treatment in Pan-Cancer patients. A total of 431 patients with nasopharyngeal carcinoma, esophageal cancer and lung cancer who underwent PD-1 checkpoint inhibitor combination therapy were included in this study. Patients were divided into two groups: progressive disease (PD) and disease control (DC) groups. Hematologic and biochemical parameters were collected before and at the third week of PD-1 therapy. Six machine learning models were developed and trained to predict the efficacy of PD-1 combination therapy at 8–12 weeks. Analysis of 57 blood biomarkers before and after three weeks of PD-1 combination therapy through statistical analysis, heatmaps, and principal component analysis did not accurately predict treatment outcome. However, with machine learning models, both the AdaBoost classifier and GBDT demonstrated high levels of prediction efficiency, with clinically acceptable AUC values exceeding 0.7. The AdaBoost classifier exhibited the highest performance among the 6 machine learning models, with a sensitivity of 0.85 and a specificity of 0.79. Our study demonstrated the potential of machine learning to predict the efficacy of PD-1 combination therapy based on changes in hematologic and biochemical parameters.