Successful Streamlined Photon-Counting CT Myelography Protocol for Detection of CSF-Venous Fistulas in Busy Practice...

CSF leaks are difficult to detect with energy-integrating CT myelography due to its limitation in spatial resolution. The purpose of this study was to develop a streamlined protocol for entire-spine evaluation for CSF leaks in patients with spontaneous intracranial hypotension using photon-counting CT (PCCT).

In this internal review board-approved study, 35 patients (25 women, mean age: 51.6 ± 15.6 years) with prior inconclusive spine imaging for CSF leak detection received PCCT myelographic imaging. Contrast was injected under fluoroscopic guidance, and patients were transported to the PCCT scanner in decubitus positioning. PCCT was performed with a protocol minimizing patients' time in the scanner to 15 minutes, while acquiring full spine images at 0.2×0.1 mm resolution in 3 breathing phases: resistive inspiration, end inspiration, and Valsalva (6 series total). All but 2 patients had imaging performed on 2 consecutive days, once for each decubitus position. The number of CT/MRI investigations and blood patches before the PCCT were documented for each patient. One neuroradiologist with subspecialty training in spinal interventions assessed each PCCT to define whether a CSF leak was identified. Three separate readers assessed best breathing phase for leak detection. Treatment was considered a success if patients received an intervention that provided symptom relief.

Patients had received 4.6 ± 5.0 prior CTs (range: 0-26) and 4.0 ± 3.0 prior MRIs (range: 1-13); this had led to a mean of 2.1 ± 1.3 prior blood/fibrin glue patches (range: 1-6) without lasting symptom relief. A CSF leak was definitively identified in 12/35 (34%) patients; an additional 8/35 (23%) patients had potential CSF leak sites. Following PCCT imaging, 4 patients received surgical treatment, 8 received percutaneous embolization, and 17 received blood/fibrin glue patches. Resistive inspiration was most frequently considered the best phase for leak detection. Of patients who underwent treatment, it was considered a success in 18/29 (62%) patients, 5/29 (17%) had no relief, 6/29 (21%) had incomplete relief, and 6 patients had not been treated at the time of last follow-up.

Using a streamlined PCCT protocol, CSF leak detection is feasible and provides a diagnosis, allowing targeted treatment options for patients whose prior imaging was inconclusive.