A57

Purpose: To evaluate the effect of combining plasma/serum Epstein-Barr viral (EBV) DNA load data with clinical staging information in prognostication of nasopharyngeal carcinoma (NPC).Patients and Methods: 376 patients of all stages of NPC, but excluding those with distant metastases on presentation, were studied. The histology of the tumors was World Health Organisation Type II or III in all except two patients. Clinical stage designation (Stages I-IV) was according to the 5th edition of the UICC-AJCC on staging. Pre-treatment plasma/serum EBV DNA concentrations were quantified by a real-time PCR assay that amplified a DNA segment in the BamHI-W fragment region of the EBV genome. A cutoff level of 4000 copies/mL was used to define low and high EBV-DNA load. All patients received radiation therapy and 41 patients also received chemotherapy. Kaplan-Meier plots of overall survival were established for patient groups of clinical stages I to IV, and for patients sub-groups with high, and low EBV DNA loads within early (Stages I-II), and advanced (Stages III-IV) UICC stages. Differences in survival probabilities were analysed by the log-rank test. P values of <0.05 were taken as signficant. Determinants of overall survival were also assessed by multivariate analysis.Results: The median follow-up duration was 5.8 years. Prognostic segregation among the different clinical stages was more obvious for advanced clinical stage. Survival probabilities were significantly different between Stages III and IV, but only marginally between Stages II and III, and not between Stages I and II. In contrast, among patients with Stage I-II disease, those with high EBV DNA load had significantly inferior survival than those with low EBV DNA load (p=0.0001). The high EBV DNA group had a survival probability similar to that of Stage III patients while the low EBV DNA group had a survival probability similar to Stage I patients. Within Stage III-IV disease, high EBV DNA load also predicts inferior survival (p=0.034), but the discrimination was not more powerful than that by clinical stage alone (p=0.0002). Multivariate analysis showed that EBV DNA was an independent prognostic factor for overall survival (p=0.017) after UICC stage (p<0.0001), while use of concurrent chemotherapy, use of a boost dose of radiotherapy, age and gender were not.Conclusion: Peripheral-blood cell-free EBV DNA load is an independent prognostic factor to clinical staging in NPC. Combined interpretation of EBV DNA data with clinical staging data leads to an alteration of risk definition of patient subsets, with improved risk discrimination in early-stage disease, with implications on the choice of therapy. Validation studies are awaited.

[First AACR International Conference on Molecular Diagnostics in Cancer Therapeutic Development, Sep 12-15, 2006]