⁶⁸Ga-DOTA Positron Emission Tomography for Diagnosis of Spinal Cerebrospinal Fluid Leaks

Abstract

Spontaneous intracranial hypotension (SIH) due to spinal cerebrospinal fluid (CSF) leakage causes substantial disease burden. In many patients, the course is protracted and refractory to conservative treatment, requiring targeted therapy. We propose positron emission tomography (PET) of the CSF space with ⁶⁸Ga-DOTA as a state-of-the-art approach of radionuclide cisternography (RC) and validate its diagnostic value. Methods: Retrospective analysis of patients with suspected intracranial hypotension due to spinal CSF leaks who underwent whole-body PET/CT at 1, 3 and 5 hours after intrathecal lumbar injection of ⁶⁸Ga-DOTA. Two independent raters blinded to clinical data analyzed all scans by for direct and indirect RC signs of CSF leakage. Volume of interest analysis was performed to assess the biological half-life of the tracer in CSF space (T1/2,biol) and the ratio of decay-corrected activity in CSF space at 5 and 3 hours (R5/3; simplified marker of tracer clearance). Comprehensive stepwise neuroradiological work-up served as reference, which was additionally validated by surgical findings and follow-up. Results: Of 40 consecutive patients, 39 patients with a working diagnosis of intracranial hypotension due to a spinal CSF leak (n = 31 spontaneous and n = 8 post-interventional) could be analyzed. A spinal CSF leak was verified by the neuroradiological reference method in 18 of 39 patients. As the only direct and indirect diagnostic signs, extrathecal tracer accumulation at the cervicothoracic junction and lack of activity over the cerebral convexities (5h) showed a high diagnostic value for spinal CSF leaks (sensitivity/specificity: 67%/90% and 94%/67%, respectively). Their combination provided little improvement (71%/95%). Additional quantitative analyses yielded no benefit (T1/2,biol: 94%/53%; R5/3: 94%/58%). Of note, the location of direct signs (extrathecal tracer accumulation) did not correlate with verified sites of spinal CSF leakage. Conclusion: We propose CSF-PET with ⁶⁸Ga-DOTA as a novel, fast and convenient approach of RC for verification, though not localization, of spinal CSF leaks with high sensitivity and specificity. CSF-PET may fulfill an important gatekeeper function to stratify patients towards escalation (rule-in) or de-escalation (rule-out) of diagnostic and therapeutic measures. Further prospective studies are needed to validate the present results and the potential of the methods to reduce the burden to the patient.