We randomly assigned patients (1:1) with computer-generated blocks of four and a sealed-envelope design to undergo intraoperative MRI-guided surgery or conventional microsurgery (control group).