Better prediction of prognosis for patients with nasopharyngeal carcinoma using primary tumor volume.

Abstract

Heterogeneity of primary tumor volume within tumors of the same classification indicates a need to elucidate the effects of primary tumor volume on treatment outcomes in patients with nasopharyngeal carcinoma (NPC). From 1994 through 1996, 129 patients with newly diagnosed NPC who were treated with high-dose radiotherapy were enrolled in the study. Computed tomography-derived primary tumor volume was measured using the summation-of-area technique. Correlations between American Joint Committee on Cancer (AJCC) disease stage, primary tumor volume, and disease-specific survival were assessed using a Cox regression model. Cross-validation based on receiver operating characteristic (ROC) curve also was examined. Compared with the AJCC staging system and the TNM classification system, primary tumor volume was better at determining cumulative survival for patients with NPC. Hazard ratios increased with tumor volume, ranging from 6.68 (95% confidence interval [95% CI], 1.89-23.67) for tumor volumes between 20-40 mL, 18.03 (95% CI, 4.80-67.75) for tumor volumes between 40-60 mL, and 26.06 (95% CI, 7.70-88.20) for tumor volumes > 60 mL. With both tumor volume and T classification in the same Cox regression model, only tumor volume remained statistically significant in the prognosis of NPC. The validation results with ROC curves also revealed that, in predicting patient outcome, primary tumor volume (area under the ROC = 83.33%) was superior to disease stage (area under the ROC = 66.53%) and TNM classification (area under the ROC = 58.61%). The incorporation of primary tumor volume may lead to a further refinement of the current AJCC staging system, particularly for patients with large primary tumor volumes (> 60 mL), who require more aggressive treatment. Cancer 2004.