Discriminant analysis of ¹⁸F-fluorothymidine kinetic parameters to predict survival in patients with recurrent high-grade glioma.

PURPOSE The primary objective of this study was to investigate whether changes in 3'-deoxy-3'-[¹⁸F]fluorothymidine (¹⁸F-FLT) kinetic parameters, taken early after the start of therapy, could predict overall survival (OS) and progression-free survival (PFS) in patients with recurrent malignant glioma undergoing treatment with bevacizumab and irinotecan. EXPERIMENTAL DESIGN High-grade recurrent brain tumors were investigated in 18 patients (8 male and 10 female), ages 26 to 76 years. Each had 3 dynamic positron emission tomography (PET) studies as follows: at baseline and after 2 and 6 weeks from the start of treatment, ¹⁸F-FLT (2.0 MBq/kg) was injected intravenously, and dynamic PET images were acquired for 1 hour. Factor analysis generated factor images from which blood and tumor uptake curves were derived. A three-compartment, two-tissue model was applied to estimate tumor ¹⁸F-FLT kinetic rate constants using a metabolite- and partial volume-corrected input function. Different combinations of predictor variables were exhaustively searched in a discriminant function to accurately classify patients into their known OS and PFS groups. A leave-one-out cross-validation technique was used to assess the generalizability of the model predictions. RESULTS In this study population, changes in single parameters such as standardized uptake value or influx rate constant did not accurately classify patients into their respective OS groups (<1 and ≥ 1 year; hit ratios ≤ 78%). However, changes in a set of ¹⁸F-FLT kinetic parameters could perfectly separate these two groups of patients (hit ratio = 100%) and were also able to correctly classify patients into their respective PFS groups (<100 and ≥ 100 days; hit ratio = 88%). CONCLUSIONS Discriminant analysis using changes in ¹⁸F-FLT kinetic parameters early during treatment seems to be a powerful method for evaluating the efficacy of therapeutic regimens.