The MR Imaging Appearance of Endoscopic Endonasal Skull Base Defect Reconstruction Using Free Mucosal Grafts

Background The endoscopic endonasal approach in skull base surgery has remarkably progressed and is now being increasingly utilized for resection of multiple skull base pathologies. With such expansion of indications, successful reconstruction of the resulting skull base defect has become equally critical and a subject of great interest. At our institution, the standardized repair after resection of pituitary adenomas via transsellar approach includes multilayer reconstruction using a free mucosal graft harvested from the nasal floor. This study is designed to describe the expected appearance of the reconstruction on postoperative MRI surveillance scans, and to assess graft integrity and impact on the sinonasal cavity.

Methods Fifty consecutive patients in a 2-year period, from November 2014 to October 2016, who underwent endoscopic endonasal resection of pituitary adenomas with multilayer skull base reconstruction using a free mucosal graft from the nasal floor were retrospectively reviewed. Both the inpatient and outpatient EMRs were reviewed for each patient's postoperative course, SNOT-22 scores, and nasal endoscopy reports. A total of 98 postoperative MRI surveillance scans were reviewed, obtained at 3 months (n = 47), 12 months (n = 39), or 24 months (n = 12) after surgery according to our departmental protocol. These scans were collected to evaluate (1) the appearance and thickness of the graft, (2) the enhancement of the graft, and (3) the T2 signal in the sphenoid sinus, as a surrogate marker of inflammation.

Results By 1 month, the graft was well healed, as visualized by nasal endoscopy. The donor site on the nasal floor was completely remucosalized as early as 1 week. There was no significant change in the thickness of the graft over time, with the average being 1.9 (0.6–6.4) mm at 3 months, 1.4 (0.6–6.4) mm at 12 months, and 1.6 (0.7–5.1) mm at 24 months. Except for the five scans that were obtained without IV contrast, all scans showed enhancement of the graft. The percentage of patients with postoperative T2 signal in the clival recess /− inferolateral recess of the sphenoid sinus decreased over time, with 91.5% at 3 months, 76.9% at 12 months, and 66.6% at 24 months. As such, more than two-thirds of the patients displayed persistent T2 signal in the mucosal lining of the sphenoid sinus at 12 and 24 months. However, this pattern was not reflected clinically, as SNOT-22 scores demonstrated a statistically significant downward trend over time.

Conclusion Our study demonstrated that multilayer reconstruction of a skull base defect using a free mucosal graft from the nasal floor is a viable option, and may be added to the armamentarium of all skull base surgeons. Postoperative MRI surveillance scans demonstrated the stable appearance of the graft with enhancement through time, reflecting its robust vascularization and integration to the skull base. Although persistent T2 signal was detected in the sphenoid sinus, clinical evidence based on nasal endoscopy and SNOT-22 scores indicated minimal sinonasal morbidity from this technique.