Prognostic Value of Fluorodeoxyglucose Uptake in Non-small Cell Lung Cancer: Time for Standardization and Validation

Abstract

Most clinical studies on the use of positron emission tomography with fluorodeoxyglucose (FDG-PET) in non-small cell lung cancer (NSCLC) have focused on its role in noninvasive diagnosis and staging.1 Another striking feature of the PET lung cancer literature is that (semi) quantitative measures of FDG uptake may also serve as a surrogate marker of the biologic aggressiveness of the tumor. Initial in vitro studies demonstrated that FDG uptake is related to the number of viable tumor cells and their doubling time.2 This led to the hypothesis that FDG uptake might be related to outcome, and numerous studies reported on the potential use of FDG-PET in the prediction of recurrence and survival rates of NSCLC, either newly diagnosed or after induction therapy in multimodality approaches.3 More recent studies unraveled how FDG uptake might be related to the different biologic pathways involved in lung cancer progression.4,5 A high-baseline FDG uptake was also reported to be a predictor of poor outcome for breast, thyroid, esophageal, and colorectal cancer, as well as sarcomas and gliomas. In this issue of Journal of Thoracic Oncology, Paesmans et al.6 report on an updated systematic review of 21 evaluable studies relating the degree of FDG uptake of the primary tumor at diagnosis to the prognosis of 2637 patients. Most of these studies were retrospective single-institution studies in predominantly surgical cohorts of patients. They examined the correlation of FDG uptake of the primary tumor, measured as standardized uptake value (SUV), and survival. Patients were dichotomized in low and high SUV categories by using the median SUV or the “best cutoff ” SUV. Overall, patients with high SUV had a significantly worse prognosis: hazard ratio 2.1 (95% confidence interval: 1.72–2.59), a result in line with their previous report,7 but with sharper confidence intervals. Sensitivity analyses for period of publication or method of dichotomization gave small variations in the results. Several clinical and methodological elements need to be considered when trying to interpret this observation. First, the tumor, node, metastasis stage at diagnosis remains the most important prognostic factor, even if it does not give an explanation for the differences in survival within a similar stage, probably due to biologic factors of tumor aggressiveness, which could be estimated by metabolic imaging of FDG uptake. However, as the authors acknowledge, this systematic review does not allow to assess the independent prognostic value of SUV compared with tumor, node, metastasis stage, because many studies in it did not report multivariate analysis of the different prognostic factors. Their results also suggest that the prognostic value of FDG uptake may be more important in early-stage NSCLC, possibly because the prognostic impact of FDG uptake is too much diluted in patients with advanced stage and major anatomic extent of the tumor. Second, the measuring tool SUV is not “standardized” at all.8 Indeed, a variety of factors can affect the (semi)quantitative determination of FDG uptake. Patient factors are body weight and plasma glucose level. Imaging acquisition factors are the dose of FDG