Familial Nasopharyngeal Carcinoma 6

Prognostic Factors in Nasopharyngeal Cancer 97considered revolutionary, and has significantlychanged the paradigm of the management of NPC.Recently, several studies have proposed some modificationsfor T-classification, N-classification, andoverall stage grouping (Au et al. 2003; Lu et al. 2006;Ma et al. 2007; Ng et al. 2007; Liu et al. 2008; Maoet al. 2008, 2009; Tang et al. 2008). Each modificationwas based on their survival data and showed betterprognostic discrimination between patients with differentsubgroup. Future revision of the staging systemshould incorporate these suggestions.It is reasonable to question the magnitude of thepredictive value of the 1997 version of staging system(or its 2002 updates) in the new era of combinedchemoradiotherapy and IMRT for NPC. Nevertheless,the follow-up periods of literatures on IMRT for NPCwere limited. The long-term effect of these new treatmenttechnology or strategy on the patients’ outcomeawaits further investigation. We expect that the effectof presenting staging on prognosis will change moreor less under different treatment modality in thefuture. Since the locoregional control has been greaterthan 90% using modern radiation technique, distantfailure rate will continue to be the critical pointaffecting patient’s prognosis. Future staging modificationshould focus on risk grouping according tothe distant failure rate. For patients with high risk ofdistant failure, more intensive systemic treatmentstrategy deserves to be investigated.9.3Predictive Value of the Volumesof the Primary Disease and CervicalLymph Adenopathy on PrognosisTumor volume or size has been well recognized asone of the major prognostic factors in the managementof most malignancies. The main purpose ofstaging systems in cancer is to segregate patients intosubgroups with different prognosis, and to guideappropriate treatment. Thus, tumor bulk has beenadopted in the staging systems of several malignancies,which often employed a crude measurement oftumor diameter as well as assessment of tumor extension.For resectable tumors, tumor size may be of lessimportance to predict local control. However, for diseasestreated with radiation therapy or chemotherapy,tumor bulk will have a great impact in local controlbecause bulky tumor usually have increasing numbersof tumor clonogens to be sterilized. In addition,larger tumor volume is also likely to contain hypoxicregions thus more resistant to radiation therapy. Anumber of publications have suggested a positivecorrelation between tumor volume and treatmentoutcome (Hamilton et al. 2004; Plataniotis et al.2004; Cheng SW et al. 2006).NPC is a highly infiltrative tumor, with propensityto spread along adjacent soft tissues as well as the skullbase and foramina (Sham and Choy 1991). Tumor volumeof NPC cannot be easily assessed clinically withoutCT scan and MRI. However, accurate assessment oftumor volume in NPC requires delineation of thetumor extent from series imaging slides and calculationof tumor volume from a 3D perspective. Withthe prevailing utilization of computerized treatmentplanningsystem in radiation therapy, the calculationof tumor volume becomes easier. As the T-classificationof NPC is based on the anatomical location and cranialnerve involvement, a large variation in tumor extensionor tumor volume exists within the same T-classification,especially in advanced primary diseases. Althoughadvanced T-disease is usually associated with poorerlocal control and survival, patients with different tumorvolume within the same T-category may possess differentprognosis. Several studies have addressed the prognosticimpact of tumor volume in NPC (Chua et al.1997b, 2004b; Chang et al. 2002; Chen et al. 2004, 2009;Sze et al. 2004; Kim and Lee 2005; Lee et al. 2008; Shenet al. 2008). Overall, the results of these studies showeda large variation in primary tumor volume in the sameT-category, irrespective of the staging systems or imagingmodalities used. Although the median and rangeof primary tumor volume differs largely in differentstudies, most studies reported a positive correlationbetween primary tumor volume and various survivals(Table 9.2).In a study reported by Chua et al. (1997b) from theQueen Mary Hospital of Hong Kong, the primary andnodal tumor of 290 NPC patients were contoured followedby summation of areas in sequential CT scanslides. All patients received radiation therapy and 67patients also received neoadjuvant chemotherapybefore radiotherapy. On univariate analyses, largeprimary disease (>60 ml) was associated with a significantlypoorer local control (p < 0.001) and disease-specificsurvival (p < 0.001). In patients with asmall primary disease (£20 ml), there were no significantdifferences in local control among differentT-classification. Large nodal volume was significantlyassociated with poor distant failure rate (p = 0.002),nodal control rate (p = 0.001), and disease-specificsurvival (p < 0.001). In multivariate analyses, primary