Management of cerebrospinal fluid leaks from lateral sphenoid sinus wall: Tailored approach based on proposed zone-wise classification.
Objective: Management of cerebrospinal fluid (CSF) leaks of the lateral sphenoid sinus wall (LSW) can be challenging to accomplish via an endoscopic transsphenoidal approach. While transcranial repair is a less chosen approach in the era of the endoscopic transpterygoid approach, its role is still relevant. We review our experience with LSW CSF leak repair via both techniques to decide the success of these techniques based on preoperative radiological images.
Methods: Electronic medical records of patients with LSW CSF leaks operated at a tertiary referral centre were retrospectively reviewed. Preoperative computed tomography (CT), CT cisternography and MRI CISS (Magnetic resonance imaging with Construction interference in steady state) were reviewed. Based on preop Coronal CT (27) / CT-cisternography (7), each defect was categorised in Zone 1, 2 or 3 based on its relation to the VR line (an imaginary line connecting vidian canal and foramen rotundum). Outcomes were measured in terms of successful repair of CSF leak and complications.
Results: Twenty-seven patients with 34 LSW CSF leaks were identified. CSF leak was spontaneous in 23 cases, and four patients had CSF leak following trauma. Based on the zone-wise classification, there were ten, seven and seventeen defects in zones 1,2 and 3, respectively. The transcranial subtemporal (TC) approach was used for 13 defects, and the endoscopic transsphenoidal (ETS) approach was used for 21 defects. Successful repair of CSF leak was achieved in 85% (23/27 cases). Success rates for the TC vs ETS approach for zone 1,2, and 3 defects were NA*1 vs 100 %, 0% vs 100% and 100% vs 50 %, respectively. One patient had persistent CSF rhinorrhoea, and three patients had recurrent CSF rhinorrhoea. Three patients required re-exploration or another approach for successful repair, while one was managed conservatively. Complications were noted in 3 patients (11%), namely temporal lobe contusion with seizures, surgical site infection and cheek sensory loss. All complications were seen in the TC approach.
Conclusions: This study demonstrates that the endoscopic transsphenoidal approach is more successful for zone 1 and zone 2 defects, while the transcranial subtemporal approach is more successful in zone 3 defects. Complications, though rare, were seen only with the transcranial approach. Hence, zone-wise classification helps in a clear understanding of the surgical approach for LSW CSF leaks.