Background: Recently, the hemoglobin to albumin ratio (HAR) has been shown to be closely associated with the survival of certain malignancies. However, its prognostic value in nasopharyngeal carcinoma (NPC) remained to be elucidated. Herein, we aimed to explore the correlation between HAR and overall survival (OS) in NPC patients treated with concurrent chemoradiotherapy (CCRT).

Methods: This retrospective study included a total of 858 patients with NPC receiving CCRT between January 2010 and December 2014 in Sun Yat-sen University Cancer Center. We randomly divided them into the training cohort (N = 602) and the validation cohort (N = 206). We performed univariate and multivariate Cox regression analyses to identify variables associated with OS, based on which, a predictive nomogram was constructed and assessed.

Results: In both the training and validation cohorts, patients were classified into low- and high-HAR groups according to the cutoff value determined by the maximally selected rank statistics. This HAR cutoff value effectively divided patients into two distinct prognostic groups with significant differences. Multivariable Cox analysis revealed that higher T-stage, N-stage, and HAR values were significantly related to poorer prognosis in NPC patients and served as independent prognostic factors for NPC. Based on these, a predictive model was constructed and graphically presented as a nomogram, whose predictive performance is satisfactory with a C-index of 0.744 [95%CI: 0.679-0.809] and superior to traditional TNM staging system [C-index = 0.609, 95%CI: 0.448-0.770].

Conclusion: The HAR value was an independent predictor for NPC patients treated with CCRT, the predictive model based on HAR with superior predictive performance than traditional TNM staging system might improve individualized survival predictions.