Role of Intraoperative MRI in Endoscopic Endonasal Transsphenoidal Pituitary Surgery

The transsphenoidal corridor for pituitary adenoma surgery was established as early as 1906 by Schloffer and was subsequently refined by Cushing throughout the early 20thcentury [1]. The use of intraoperative magnetic resonance imaging (iMRI) in endoscopic endonasal transsphenoidal resections, however, is a relatively contemporary addition to the surgical treatment of pituitary tumors. The morbidity of these cases has decreased over the years in light of advances in intraoperative navigation as well as improvements in endoscope dynamics and surgical instruments. Despite such improvements, a substantial number of patients require repeat surgeries or subsequent radiotherapy for residual and/or recurrent disease. This can be largely attributed to cavernous sinus invasion or suprasellar extension, which pose technical challenges to achieving gross total resections (GTRs). The rate of GTR for pituitary tumors cited in the literature varies from 59–88%.[2–3] The advantage of iMRI is that it provides the surgeon with immediate feedback regarding their progress and ability to safely achieve GTR which, in pituitary surgery, is critical for long term cure. Additionally, although there is concern for increased risk of postoperative endocrine dysfunction, Zhibin et al prove that this is not necessarily the case. In their series, 133 patients who underwent iMRI had higher rates of GTR and did not have a significant difference in postoperative hypopituitarism. [4]

This study includes a combined retrospective and prospective comparative analysis of 238 patients who underwent transsphenoidal resection of a pituitary tumor from January 2013 until May 2019. All patients were operated on by one of four experienced neurosurgeons and one of three experienced otolaryngologists. There were 203 patients who did not undergo iMRI and 25 patients who did. A 3 tesla MRI magnet was used in all cases. All intraoperative images were read and interpreted by a senior neuroradiologist at our institution. Amongst the two groups, there was no statistically significant difference in patient age (p = 0.488), tumor size (microadenoma versus macroadenoma, p = 0.878), and primary versus recurrent tumor (p = 0.837). The use of iMRI did not yield a decrease in the length of stay (4.84 days in the no iMRI group and 5 in the iMRI group, p = 0.777). There were zero cases of a return to the OR for residual tumor in the intraoperative MRI group versus the non-MRI group. However, this did not reach statistical significance. This study did not yield a statistically significant difference in GTR (p = 0.75), near total resection (NTR, p = 0.167), or subtotal resection (p = 0.083). This is likely secondary to a low sample size and therefore power in the iMRI group. Finally, there was no significant difference in the number of patients requiring postoperative DDAVP (p = 0.099) or hydrocortisone (p = 0.873) after discharge.

Preliminary results reveal a potential benefit of iMRI use to assess for residual disease which can be addressed immediately during the initial operation, thus decreasing the need for re-operations. Furthermore, the ability to correlate intraoperative findings with an intraoperative structure may lead to more precise identification and preservation of normal gland, which can possibly decrease the incidence of postoperative endocrine dysfunction.

No conflict of interest has been declared by the author(s).