Integration of dynamic parameters in the analysis of ¹⁸F-FDopa PET imaging improves the prediction of molecular features of gliomas.

Abstract

¹⁸F-FDopa PET imaging of gliomas is routinely interpreted with standardized uptake value (SUV)-derived indices. This study aimed to determine the added value of dynamic ¹⁸F-FDopa PET parameters for predicting the molecular features of newly diagnosed gliomas. We retrospectively included 58 patients having undergone an ¹⁸F-FDopa PET for establishing the initial diagnosis of gliomas, whose molecular features were additionally characterized according to the WHO 2016 classification. Dynamic parameters, involving time-to-peak (TTP) values and curve slopes, were tested for the prediction of glioma types in addition to current static parameters, i.e., tumor-to-normal brain or tumor-to-striatum SUV ratios and metabolic tumor volume (MTV). There were 21 IDH mutant without 1p/19q co-deletion (IDH+/1p19q-) gliomas, 16 IDH mutants with 1p/19q co-deletion (IDH+/1p19q+) gliomas, and 21 IDH wildtype (IDH-) gliomas. Dynamic parameters enabled differentiating the gliomas according to these molecular features, whereas static parameters did not. In particular, a longer TTP was the single best independent predictor for identifying (1) IDH mutation status (area under the curve (AUC) of 0.789, global accuracy of 74% for the criterion of a TTP ≥ 5.4 min) and (2) 1p/19q co-deletion status (AUC of 0.679, global accuracy of 69% for the criterion of a TTP ≥ 6.9 min). Moreover, the TTP from IDH- gliomas was significantly shorter than those from both IDH+/1p19q- and IDH+/1p19q+ (p ≤ 0.007). Prediction of the molecular features of newly diagnosed gliomas with ¹⁸F-FDopa PET and especially of the presence or not of an IDH mutation, may be obtained with dynamic but not with current static uptake parameters.