An effective MRI perfusion threshold based workflow to triage additional $^{18}$F-FET PET in posttreatment high grade glioma

MRI is the preferred method for follow-up imaging of post-treatment WHO grade 3 or 4 gliomas. While positron emission tomography with O-(2-[F]fluoroethyl)-L-tyrosine) (F-FET PET) offers higher diagnostic accuracy, its use is limited due to low availability. We propose a sequential, threshold-based workflow to triage patients for additional F-FET PET scans based on MRI dynamic susceptibility contrast (DSC) perfusion-derived rCBV values, to optimize F-FET PET resource allocation. Patients with high-grade gliomas who had undergone standard-of-care treatment and developed new or enlarging contrast-enhancing post-treatment lesions on MRI were included, with a F-FET PET study performed within 4 months of the MRI. Patients were excluded if there were significant changes in lesion size or treatment between the MRI and F-FET PET scan. An rCBV threshold was determined and the performance of a threshold-based imaging workflow was evaluated compared to the gold standard defined here as surgical verification or long-term imaging follow-up without further intervention. Forty-one patients with a total of 49 lesions were included (tumor progression n = 40, treatment-related changes n = 9). Above the rCBV threshold of 2.4, MRI was 100% accurate (21/21 patients) in diagnosing tumor progression. Below the threshold, MRI identified 9 true negatives but produced 19 false negatives. F-FET PET reclassified 18/19 (95%) false negatives resulting in an overall accuracy of 48/49 (98%) for the workflow. Our MRI DSC perfusion rCBV-based threshold workflow for triaging patients for additional F-FET PET imaging in post-treatment high grade glioma has the potential to optimize F-FET PET resource allocation.