Facebook
Twitter
LinkedIn
Email
 |
Checking for direct PDF access through Ovid | |
Excerpt
A collaborative project of the Brigham and Women's Hospital and General Electric Corporation (Milwaukee, WI) has led to the development of a magnetic resonance imaging (MRI) scanner than can be used for intraoperative imaging guidance. The authors recently reported results on a series of 60 craniotomies for tumor resection, which were performed with this prototype unit. Twenty-nine women and 31 men underwent craniotomies for treatment of intracranial mass lesions during a one-year period. The ages of the patients ranged from 21 to 72 years.
The intraoperative MRI unit was designed by making several major modifications to a conventional MRI scanner. The 0.5 tesla superconducting “open-configuration” magnet was produced with coils in two separate but communicating cryostats resulting in a 56 cm-wide vertical gap at the center of the magnet. This configuration allows the surgeon to stand on either side of a patient placed within the scanner. Shielded gradient coils were engineered with a corresponding central gap, and flexible radiofrequency coils were contoured around the head, framing the planned craniotomy incision. The design yields a spherical imaging volume of 30 cm with no significant reduction in the gradient strength, linearity in the field of view, or eddy current performance. A light-emitting diode (LED)-based Flashpoint navigational system (Flashpoint; Integrated Boulder Technology, Boulder, CO) is an integrated component of the MRI unit and provides tracking capabilities within the three-dimensional space at the center of the magnet.
In this group of patients, images were obtained before opening of the dura, after opening the dura in all cases, after the administraion of gadolinium, and during tumor resection. A final set of images was obtained for all patients after closure of the craniotomy to ensure that no hemorrhage was present in the resection bed. Some patients with lesions in or near eloquent cortex underwent surgery under local anaesthesia and remained awake throughout the procedure to be tested for motor/speech function during the procedure. The remainder of the patients received general anesthesia.
The surgeons reported no technical difficulties during the operations performed using intraoperative MRI. Marked shifting (>1.0 cm) of the brain parenchyma was regularly observed in serial images as the operations progressed. In more than one-third of the cases, tumor resection was considered complete on the basis of the surgical view alone, but MRI revealed residual tumor requiring further resection.
One patient experienced delayed hemorrhage into the resection bed 48 hours after surgery and required surgical evaluation of the clot. A second patient exhibited a cerebrospinal fluid leak from the wound and required placement of additional sutures. Six patients with lesions adjacent to the motor strip and two with lesions involving the language cortex exhibited transient deficits after surgery; however, these changes were secondary to localized brain tissue edema and fully resolved by 1 month follow-up. In two cases, routinely obtained surveillance images during closure revealed bleeding in the tumor resection bed, and this potential complication was treated before the final craniotomy closure. The authors estimate that the use of the intraoperative MRI added approximately 1 hour to the operative procedure time.
Overall, the authors' enthusiasm for this new technology is subjective, particularly in view of the comment that one-third of the tumor resections were considered complete before MRI imaging, and then further surgery was undertaken based upon this imaging.
