Proposal of a Clinical Typing System and Generation of a Prognostic Model in Patients with Nasopharyngeal Carcinoma from Southern China.
Peng Sun,Cui Chen,Xin-Lin Chen,Yi-Kan Cheng,Lei Zeng,Zhi-Jian Zeng,Li-Zhi Liu,Yong Su,Mo-Fa Gu
2014-01-01
Abstract:PURPOSE:To propose a novel clinical typing classification focusing on the distinct progression patterns of nasopharyngeal carcinoma (NPC), to supplement our knowledge of the clinical-biological behavior, to provide useful knowledge for treatment planning, and to contribute to basic research in NPC.METHODS:632 consecutive patients were retrospectively reviewed and analyzed according to the novel typing system. We considered that NPC can be divided into 5 types as follows: limited (L), ascending (A), descending (D) ascending- descending (mixed) (AD), and distant metastasis types (M). The distribution of these clinical types, their association with Epstein-Barr virus (EBV) serology and prognostic value were explored.RESULTS:55 (8.70%), 59 (9.34%), 177 (28.01%), 321 (50.79%) and 20 (3.16%) patients were classified as Type L, A, D, AD and M, respectively. EBV (VCA)-IgA titers, EBV early antigen (EA)-IgA serum titers, and capsid antigen lg(EBV DNA) were positively associated with the clinical typing (p<0.05). The 3-year overall survival (OS) rates for Types L, A, D, AD and M were 100, 100, 95.10, 88.20 and 59.30%, respectively (p<0.001). A prognostic model was constructed based on pretreatment lg (EBV DNA) and clinical type, which were independent predictors of OS (multivariate Cox proportional model). The prognostic model stratified patients into 4 risk subgroups. The 3-year OS rates of the low, intermediate, high and extremely high risk groups were 99.5, 90.0, 85.5 and 53.2%, respectively (p<0.001). Compared with the low-risk group, the risk of death was 4.96, 8.75 and 35.9 in the intermediate, high and extremely high risk groups, respectively (p<0.001). The model also predicted OS independently of TNM classification.CONCLUSION:This novel clinical typing system and prognostic model can supplement TNM classification, and may help design novel treatment strategies, evaluate risk stratification and investigate the varied biological characteristics of NPC.