NEIM-05 PURSUE: RESULTS FROM A PROSPECTIVE, PHASE 2B TRIAL TO DEFINE IMAGE INTERPRETATION CRITERIA FOR18F-FLUCICLOVINE-PET FOR DETECTION OF RECURRENT BRAIN METASTASES AFTER RADIATION THERAPY

Abstract BACKGROUND After radiation therapy (RT) for brain metastases (BM), differentiating necrosis from tumor recurrence remains challenging. PURSUE (NCT04410367) evaluated various measures of lesion 18F-fluciclovine uptake to establish image interpretation criteria (IIC) 18F-fluciclovine-PET accurately diagnose recurrent BM. METHODS Patients with solid BM and a previously irradiated ‘reference’ equivocal on MRI who were scheduled craniotomy underwent (185 MBq) <42 days post-MRI 1-21 pre-craniotomy. Lesion was qualitatively rated as ‘mild’ (up blood pool), ‘moderate’ (above pool parotid gland), or ‘marked’ gland) by three independent blinded readers. The primary endpoint diagnostic performance different thresholds qualitative reads vs central histopathology. Secondary endpoints included performance, based thresholds, quantitative (e.g., standardized value [SUV]) dynamic measures. IIC determined committee subsequently used assess the 18F-fluciclovine. RESULTS All 23 reference lesions enrolled patients histopathological analysis; 10 (43%) pathologically-confirmed recurrence. highest performing measure ‘marked’, rendering 92-100% sensitivity 40-80% specificity across SUVmax reader-independent, high-performing metric (AUC 0.87, threshold 4.8, 80% sensitivity, 85% specificity). Dynamic did not add value. Thus, established as: “Lesions equal greater than visually gland, should be considered suspicious Otherwise, unlikely.” Application resulted in 77-85% specificity, 73-80% PPV, 83-85% NPV CONCLUSIONS is first prospective, multicenter trial suspected after RT, verified Finalized demonstrated excellent will help evaluate future studies.