Fluid-Suppressed Amide Proton Transfer-Weighted Imaging Outperforms Leakage-Corrected Dynamic Susceptibility Contrast Perfusion in Distinguishing Progression from Radionecrosis in Brain Metastases

Background: Differentiating brain radionecrosis (RN) from tumor progression (TP) is a persistent clinical difficulty. Here, we compared the diagnostic accuracy of leakage-corrected relative cerebral blood volume (rCBV) and fluid-suppressed amide proton transfer-weighted (APTw) imaging in distinguishing between RN and TP in metastases. Methods: Subjects with enlarging lesions after stereotactic radiosurgery were prospectively examined at 3T. APTw data were acquired with a 3D snapshot-gradient echo sequence. B0 and B1 inhomogeneities were corrected using the WASAB1 protocol. rCBV was calculated according to established guidelines. Image analysis was performed using Olea Sphere 3.0 software. ΔAPTw and ΔrCBV were calculated as the average signal within the lesion normalized against the average signal in the contralateral white matter. A diagnosis of TP or RN was assessed by histology or imaging at follow-up. Independent samples t-tests of ΔAPTw and ΔrCBV and the areas under the curve (AUCs) were computed. Results: Twenty-one metastases (10 RN, 11 TP) were evaluated. APTw differentiated between RN and TP (U = 120, p < 0.001), in contrast to rCBV (U = 71, p = 0.174). The AUC was 0.991 (95% CI = 0.962–1.020) for ΔAPTw, and 0.636 (95% CI = 0.352–0.921) for ΔrCBV. The optimal cutoff points were 0.4 and 2.1 for ΔAPTw and ΔrCBV, respectively. The sensitivity and specificity for RN-TP were 100% and 90% for ΔAPTw and 63.6% and 36.4% for ΔrCBV. Conclusions: Fluid-suppressed APTw metrics enabled more accurate diagnostic performances than leakage-corrected rCBV metrics in distinguishing between RN and TP. These promising results suggest that APTw imaging could valuably complement current multiparametric MRI protocols in brain metastases follow-ups.