Characterization of tumor remnants in intraoperative MRI-assisted microscopic and endoscopic transsphenoidal resection of less invasive pituitary adenomas.

INTRODUCTION Intraoperative magnetic resonance imaging (iMRI) improves the intraoperative detection of adenoma remnants in transsphenoidal surgery. iMRI might be redundant in endoscopic pituitary surgery in non-invasive tumors (Knosp 0-2) due to a superior visualization of anatomical structures in the periphery of the sella turcica compared to the microscopic technique. We identified the anatomical location of tumor remnants in iMRI and evaluated risk factors for secondary resection after iMRI and hereby selected patients with pituitary adenomas who may benefit from iMRI-assisted resection. METHODS We conducted a retrospective monocenter study of patients who underwent iMRI-assisted transsphenoidal surgical resection of pituitary adenomas at our department between 2012 and 2020. A total number of 190 consecutive iMRI-assisted transsphenoidal surgeries of pituitary adenomas graded as Knosp 0-2 were selected for analysis. Exclusion criteria were missing iMRI availability or pathologies other than adenomas. Of these 190 cases, 46.3% (N = 88) were treated with microscopic, 48.4% (N = 92) with endoscopic, and 5.3% (N = 10) with endoscopic-assisted technique. Volumetric measurement of preoperative, intraoperative, and postoperative tumor extension was performed. Demographic data, tumor characteristics, and MRI features were evaluated. Additionally, analysis of adenoma remnants identified by iMRI was performed. RESULTS An additional resection after iMRI was performed in 16.3% (N = 31). iMRI helped to reach gross total resection (GTR) in 83.9% (26/31) of these cases. False-positive resection was found in 1 patient (0.5%). Multivariable logistic analysis identified tumor volume (OR = 1.2, p = 0.007) recurrence (OR = 11.3, p = 0.002) and microscopic technique (OR = 2.8, p = 0.029) as independent risk factors for additional resection. Simultaneously, the endoscopic technique was significantly associated with GTR as evaluated by iMRI (OR = 2.8, p = 0.011) and postoperative MRI (OR = 5.8, p = 0.027). The detailed analysis of adenoma remnants on iMRI revealed the suprasellar location in a diaphragm fold, penetrating tumor above the diaphragm, or undetected invasion of cavernous sinus as well as in case of microscopic resection tumor location outside the line of sight as the main reasons for incomplete resections. CONCLUSION Tumor volume, recurrence, and microscopic technique were identified as independent predictors for additional resection in patients with Knosp 0-2 adenomas. iMRI might increase the extent of resection (EOR) safely even after the endoscopic visualization of the sella with very low risk for false-positive findings. Remnants of tumors hidden within the diaphragmic folds, intrathecally, or behind the infiltrated wall of cavernous sinus not recognized on preoperative MRI were the most common findings in iMRI.