18F-Fluciclovine for Detection of Recurrent Brain Metastases After Radiation Therapy: Image Interpretation Criteria and Diagnostic Performance From the PURSUE Study
Publication Title
International journal of radiation oncology, biology, physics
Document Type
Article
Publication Date
2-26-2026
Keywords
california; pni; sjci; santa monica
Abstract
Purpose: Differentiating recurrent brain metastases (BM) from treatment-related changes after radiation therapy remains challenging. PURSUE (NCT04410367) evaluated various lesion uptake metrics to establish image interpretation criteria to detect recurrent BM with 18F-fluciclovine positron emission tomography (PET).
Methods and materials: Patients with BM were enrolled if they had a previously irradiated MRI-equivocal "reference" lesion and were planned for craniotomy. PET with 18F-fluciclovine (185 MBq) was performed < 42 days post-MRI and 1-21 days precraniotomy. Performance of different qualitative thresholds of lesion 18F-fluciclovine uptake on static PET images (10-20 minutes postinjection) versus histopathological standard of truth was assessed by 3 blinded readers, as were semiquantitative (standardized uptake value [SUV]) and dynamic uptake measures, to establish 18F-fluciclovine interpretation criteria. A separate committee reviewed all data to establish interpretation criteria. Following initial review, additional blinded reads were conducted at three timepoints (10-20, 15-25, and 20-30 minutes) to explore visual-only reads (uptake ≥ parotid/pituitary as reference) and SUV ratio (SUVR)-based criteria (SUVpeak(lesion-to-pituitary)).
Results: Twenty-three reference lesions from 23 subjects underwent histopathological analysis; 10 (43%) were confirmed as recurrence. At 10-20 minutes postinjection, the highest performing qualitative measure was "uptake higher than parotid" (specificity: 92%-100%; sensitivity: 40%-80% across readers). SUVmax was a high-performing metric on receiver operating characteristic analysis (area under the curve: 0.87, sensitivity: 80%, specificity: 85% for SUVmax threshold = 4.8). Dynamic measures provided no further diagnostic value. Additional reads showed visual-only interpretation was optimal at 15 to 25 minutes (sensitivity: 60%-90%; specificity: 77%-100%). SUVR-based interpretation using a threshold of 1.1 provided 70% to 80% sensitivity and 85% specificity.
Conclusions: PURSUE established effective, histopathologically verified 18F-fluciclovine interpretation criteria for diagnosing recurrent BM after radiation therapy that will help evaluate 18F-fluciclovine PET in future studies.
Area of Special Interest
Cancer
Area of Special Interest
Neurosciences (Brain & Spine)
Specialty/Research Institute
Oncology
Specialty/Research Institute
Neurosciences
Specialty/Research Institute
Diagnostic Imaging
DOI
10.1016/j.ijrobp.2026.02.239
