projects II

Fabio Guglielmetti (1), Luigi Mariani (2), Sven Berkmann (1)

10-years’ experience using low-field intraoperative MRI in transsphenoidal surgery

(1) Department of Neurosurgery, Kantonsspital Aarau, Aarau/Switzerland

(2) Department of Neurosurgery, University Hospital of Basel, Basel/Switzerland

  • retrospective
  • monocentric
  • start scheduled 01.08.16
  • conception


INTRODUCTION. Transsphenoidal surgery is the first-line treatment in the majority of symptomatic sellar tumors. The goals are cure of endocrinological syndromes (e.g. acromegaly), neurological deficits (e.g. optic chiasm compression syndrome), restoration of pituitary function and long-term tumor control. With the evolution of the transsphenoidal approach, several tools like the X-ray intensifier, cisternography, sonography, computed tomography, the endoscope, and the intraoperative MRI (iMRI) have been used to increase visualization of the tumor and thereby to make surgery safer and increase the extent of resection. The two significant advances in pituitary surgery during the past 15 years have been the tailoring of the endonasal endoscopic approach and the implementation of intraoperative magnetic resonance imaging. Each provides improved visualization of intra- and parasellar anatomy with the goal of attaining a complete resection. The aim of this retrospective single center outcome study is to assess the 10 years’ experience using low-field intraoperative MRI in transsphenoidal surgery in >200 patients.

AIM. The primary objective of this study is to evaluate the outcome of intraoperative low-field magnetic resonance imaging in a large number of patients (n>200) and during a time frame of 10 years, which may result in a mean follow-up >3 years.

The following outcomes will be addressed:

  • oncological results (rates of gross total resection, progressions and recurrences; need for further surgery, medical tumor therapy or radiotherapy)
  • endocrinological results (number of biochemical remission, hormone substitution dependency and hormonal recovery; number of salt and water balance disorders)
  • ophthalmological outcome (course of visual acuity, visual fields, and ocular palsies)
  • adverse events due to surgery
Secondary objectives will be (1) time efficiency (i.e. prolongation of the procedure by intraoperative scans), (2) analysis of sensitivity and specificity of the iMRI compared to postoperative MRI.


METHODS. All patients who suffer from sellar lesions and who were operated at the neurosurgical department of the Kantonsspital Aarau, Switzerland, between 2006 and 2016 (n=274) will be retrospectively screened for inclusion in this study. This period corresponds with the routine use of the 0.15T iMRI unit (PoleStar, Medtronic). Criteria for inclusion are (1) transsphenoidal surgery performed by a neurosurgeon at the Kantonsspital Aarau between 2006 and 2016 and (2) use of iMRI. Patients who were not operated in the iMRI will be excluded from outcome assessment; however, the reasons for not using the iMRI will be recorded. The course of the extent of surgical resection during follow-up imaging, surgery-related complications, endocrinological outcome (e.g. hormone substitution dependency, biochemical remission rates), ophthalmological outcome (i.e. improvement of vision, visual fields und ocular palsies) will be assessed based on clinical records and radiological imaging. The details of the intraoperative workflow (e.g. additional resection possible due to results of the iMRI) will be described. Intraoperative MRI results will be compared to postoperative follow-up MRI. Prognostic factors for oncological, endocrinological and ophthalmological outcome will be evaluated. The data will be collected using the institutions pituitary tumor registry, which is based on an electronic data capture system, named secuTrial®. This system runs on a server maintained by the IT-department of the University Hospital Basel. Patients will be informed about the inclusion of their data into the database. Approval of the local ethical committee for conducting this retrospective study and for use of the institutions pituitary tumor database was obtained. For statistical analysis, we will use univariate and multivariate cox regression models for all time to event data (i.e. time to relapse of disease) and logistic and linear regression for binary and continuous outcomes as appropriate. Effect estimates will be presented with 95% confidence intervals (CI). We will investigate evidence for effect modification by including interaction terms into the statistical models with p value of 0.05 indicating significant effects. All reported P values will be two-sided, with a significance level of 0.05.

last update 08.12.2017