Discussion
We propose two clinical issues to be addressed in this case. First, rigid reconstruction was possible even for a narrow fistulous skull base defect. Second, thin-slice HRCT images were effective in identifying the fistula.
The main feature of this procedure is the use of bone for rigid reconstruction. In this case, a multilayer reconstruction was performed, which included a collagen matrix laid against the dura mater as the first layer, bone pile fitted into the bone as the second layer, and a pedicle nasoseptal flap covering the bone as the third layer. Previously, for small defects, only soft reconstruction was performed with fascia or fat and rigid reconstruction was not recommended7-9. However, since spontaneous CSF leakage is related to chronic intracranial hypertension2-6, a strong reconstruction is preferable. In this case, the fistula was in the form of a narrow canal, therefore we processed the vomer bone into a pile and inserted it in a pile-driving manner. This technique provides a rigid bone-based reconstruction for canalicular fistulas.
The fistula in this case was thin and canal-like; thus, a 0.5-mm slice HRCT was useful in identifying it. According to the previous reports, spontaneous CSF leakage in the sphenoid sinus has often been associated with meningoencephaloceles3,10-12. CT and MRI are commonly used imaging modalities, with a sensitivity of approximately 90%9. However, in this case, CT and MRI of the brain did not raise suspicion of spontaneous CSF leakage. Even in the absence of the lateral recess of the sphenoid sinus and meningoencephalocele, it is important to recognize the possibility of CSF leakage via a canal-like fistula and an ascending bacterial infection.
We performed reconstructive surgery of the fistula on the lateral wall of the sphenoid sinus using a transseptal approach. The transpterygoid approach is useful for observation and surgical manipulation of the defect in patients with a well-developed lateral recess of the sphenoid sinus13. Since the lateral recess of the sphenoid sinus was not observed in the current case, the surgical field was expanded laterally via the transseptal approach. We were able to observe and manipulate the defect area sufficiently using a 30-degree telescope.
No complications were observed 6 months after surgery. However, it is necessary to investigate the possibility of resorption of the injected bone and recurrence of CSF leakage.