RADT-19. PILOT STUDY OF VELOCITY-SELECTIVE ARTERIAL SPIN LABEL MRI (VSASL-MRI) DERIVED RELATIVE CEREBRAL BLOOD FLOW TO DISTINGUISH IMAGING CHANGES AFTER RADIOSURGERY (SRS) FROM PROGRESSIVE BRAIN METASTASIS (BM)

Abstract

Abstract Management of imaging progression on standard contrast enhanced MRI (cMRI) after SRS for brain metastasis is challenging as treatment effect/necrosis appears similar to progressive tumor. VSASL-MRI uses water labeled based on velocity, rather than spatial location, as an endogenous contrast agent and is therefore insensitive to transit delay impacting other ASL-MRI techniques. Other advantages include no IV contrast, limited susceptibility artifact, and absence of confounding leakage of gadolinium contrast into tissues. VSASL-MRI was performed (median 15 (4-123) months after SRS) to evaluate SRS treated solid tumor BM with clinician determined imaging progression on cMRI to evaluate the hypothesis that VSASL-MRI may distinguish biologic tumor (hyperperfused compared with normal brain) from treatment effect/necrosis (hypoperfused). 17 patients (19 progressing lesions) were enrolled. 13/19 index lesions had hypoperfusion. Of these, 3/13 had resection with no tumor present, 8/13 were conservatively managed and appeared to be treatment effect/necrosis as they stabilized or improved on follow-up cMRI without further BM directed intervention, and 2/13 had resection for further progression with active tumor present. For 6 index lesions that were hyperperfused: 3/6 had pathology were confirmed to be tumor, whereas the other 3/6 continued to progress (one got additional RT, two expired with progression). In addition, two pts had concurrent identification of new non-index previously untreated lesions on cMRI that were hyperperfused on VSASL-MRI and were treated. One of these patients had subsequent VSASL-MRI demonstrating post-SRS resolution of perfusion 2 months after treatment. This pilot data supports the hypothesis that VSASL-MRI may be useful in distinguishing tumor progression from radiation related imaging changes and/or be an early biomarker of radiosurgery response. We are initiating further study (1R01CA282928-01, PI; Qin) to refine the technique and prospectively assess the value of this approach in a larger number of patients.