Reply to "Prognostic Value of Fluorodeoxyglucose-Positron Emission Tomography".

e102 Journal of Thoracic Oncology ® • Volume 10, Number 10, October 2015 importance of the T-stage and have reanalyzed our primary data. By analysis of variance and the Wilcoxon rank sum test, we show that the distribution of maximum standardized uptake value (SUV max ) within the three “T” groups (T1a/T1b/T2a) is significantly different (p < 0.0001). Cox proportional hazards model was then used to assess the relationship between T-stage and survival with and without adjustment for SUV max (uncategorized). After adjustment for T-stage, the effect of SUV max remains statistically significant (p = 0.003). In our study, the effect of T-stage was not a statistically significant predictor of survival with or without adjustment for SUV max . The lack of statistical significance for T-stage does not necessarily imply that the variable does not have prognostic importance. Rather, the power of this statistical analysis may be limited by there being only 134 deaths, which in general drives the power of a study to detect a clinically meaningful effect. As we suggested, prospective treatment trials in early-stage disease could incorporate SUV max to stratify patients for therapy, although we acknowledge that other features including T-stage need to be considered.