Re: Prognostic value of FDG uptake in early stage non-small cell lung cancer

Abstract

I read with interest the article by Hanin and colleagues regarding the prognostic value of 2-[18F] fluoro-2-deoxy-Dglucose (FDG) uptake in early stage non-small cell lung cancer (NSCLC), published recently in the May issue of the European Journal of Cardio-thoracic Surgery [1]. Standardised uptake value (SUV) of FDG has previously been shown to prognosticate disease recurrence and survival in patients with early stage NSCLC [2,3]. The current study also reported similar findings. However, there are several faults with the data presentation and analysis in this article. First, unlike previous studies which all have reported disease-free and overall survival for any malignancies in terms of 5-year survival percentages, Hanin and colleagues have chosen to report their findings in terms of median survival in months from time of surgery for both low and high SUV groups. The median survival was 127 months and 69 months in patients with SUV 7.8, respectively, both of which had median survival of more than 5 years. Beyond 5 years after a curative resection for early stage NSCLC, the overall influence of the initial cancer on survival and mortality becomes unclear as other causes of mortality can come into play. Of the 40 patients who died during the follow-up period, more than 50% (21/40) of the deaths were due to cancer unrelated deaths [1]. This group should not be included in the survival analysis as this will bias the outcome. Unfortunately, this cohort was included in the overall analysis by the authors, which is obvious from Fig. 1 where the number of events (death) is more than 30. Hence the authors’ conclusion is misleading. Second, a multivariate analysis would have been ideal to distinguish the effect of other variables on survival such as age, stage 1A or 1B, etc. Although a univariate analysis was performed, the authors did not report this in the result section nor did they perform a multivariate analysis as the other variables tested (histologic subtype, stage, differentiation) failed to achieve any statistical significance. Considering that the median age of the study population was 65.3 years with the oldest patient at 82.2 years and a reported median survival of more than 5 years, age may have been a significant factor as was found previously [4]. Lastly, previous reports have shown that patients with high SUV are more likely to have poorly differentiated tumours, advanced stages and are less likely to have their disease completely resected [2,4]. Hanin and colleagues have also confirmed these findings that advance stage tumours such as T2 and T3 tumours, undifferentiated tumours, were more likely to have high SUV [1]. Hence high SUV can be considered as a marker as well as an independent prognostic factor for clinical TNM stage but it has been shown not to add to the prognostic significance of pathologic TNM stage [5].