The reliability of identifying the Omega sign using axial T2-weighted magnetic resonance imaging

Abstract

Preoperative identification of the eloquent brain is important for neurosurgical planning. One common method of finding the motor cortex is by localizing “the Omega sign.” No studies have tested the reliability of imaging to identify the Omega sign. We identified 40 recent and consecutive patients who had undergone preoperative functional magnetic resonance imaging for identification of the hand motor area prior to tumor resection. We recruited 11 neurosurgical residents of various levels of training and one board-certified neurosurgeon to identify the hand motor cortex Omega. Testees were given axial images of T2-weighted MRI and placed marks where they expected to find the Omega. Two board-certified radiologists graded and quantified the localization attempts. Inter-rater reliability was assessed using the kappa statistic, and Rao–Scott chi-square tests were used to examine the relationship between clinical factors and testees’ experience with correct identification of the Omega sign. The overall correct identification rate was 69.9% (95% CI = 63.4–75.7), ranging from 36.6% to 92.7% among all raters for the tumor side and from 46.2% to 97.4% for the non-tumor side. Anatomic distortion greatly affected correct identification ( p < 0.005). Senior residents had a significantly higher rate of identification of the Omega than junior residents ( p < 0.001). Overall, inter-rater reliability for the Omega sign is poor, with a Fleiss kappa of 0.23. We concluded that correct identification of the Omega sign is affected by tumor distortion and experience but overall is not reliable. This underscores the limitations of anatomic landmarks and the importance of utilizing multiple scanning planes and preoperative fMRI for appropriate localization.