By Dr. Laurent Hermoye
Intraoperative MRI can provide the neurosurgeons with updated MRI images during an operation.
The aim of most intracranial neurosurgical operations is the comprehensive treatment of a brain lesion, by directing the procedure very precisely on the target without damaging, or minimally, the adjacent normal parenchyma. This is made possible through the development of neuroimaging, which, coupled with stereotactic localization techniques with or without a frame (neuronavigation), can be used to guide the surgical procedure. However, as part of the surgical excision of brain tumors, this traditional approach has limitations.
It has been shown that during a craniotomy, the displacement of brain structures ("brain shift") can vary between 3 and 24 mm from the preoperative status (1). The conventional neuronavigation is thus limited because it is based on preoperative imaging and does not allow the position of the deformed tissue contours to be updated during surgery. The completeness of tumor resection is largely based on visual inspection which is limited by the experience of the neurosurgeon, the restricted field of view and the difficulty to distinguish pathologic tissue from normal tissue.
Intraoperative imaging can solve the limitations of neuronavigation in providing updated images during the surgical procedure. One of the first implementations of such a system dates back to 1995, when General Electric installed at Boston's Brigham and Women's Hospital its 0.5-Tesla open magnet, known as the "double-donut" in reference to its shape (2). Its usefulness was quickly demonstrated by Dr. Ferenc Jolesz's team initially to guide biopsies, then to monitor tumor resections and verify the success of other operations (3).
A few years later, other low field systems emerged, such as the 0.12 Tesla PoleStar system developed by the Israeli company Odin (which was upgraded to 0.20 Tesla and is now marketed by Medtronic) (4). Their permanent magnet, in the form of two discs reassembled around the patient's head when an image must be acquired, allows for easy integration in the operating room. However the weak signal-to-noise ratio (SNR), due to their low field strength, and limited field of view reduce the image quality.
Only MRI systems with higher fields strength could overcome these limitations. Several 1.5 Tesla intraoperative MRI systems then saw their day in operating rooms (5, 6). Patients may be transported towards the magnet when imaging is necessary. In the configuration proposed by the Canadian company IMRIS, it is the magnet that moves towards the patient (7).
MRI systems dedicated to the operating room suffer from a major economic limitation: they are used only for a few tests each week. Some MRIs have therefore been shared between intraoperative use and standard diagnostic use (5). At the Cliniques Universitaires Saint-Luc in Brussels, a 3 Tesla MRI was installed in a room adjacent to the operating room (Figure 3) (8).
Intraoperative imaging requires MRI-compatible equipment, including the headrest tip, the respirator, monitoring equipment and some surgical instruments. Since the patient's position usually does not allow his head to enter into the head coil, flexible coils may be placed around his head. It is also necessary to clean the room and to switch the airflow to operating conditions.
On high field MRIs, all sequences can be used with the same quality as in diagnostic mode. On the T1-injected sequence, the existence of a non-tumoral enhancement on the edges of the surgical cavity can be due to the rupture of the blood-brain barrier. T2 imaging allows for the distinguishing of a residual tumor. In addition to anatomical sequences, perfusion, diffusion and diffusion tensor sequences can be used intra-operatively.
(1) Nimsky et al. Neurosurgery. 2000; 47:1070-1079.
(2) Schenck et al. Radiology 1995;195:805-814.
(3) Alexander et al. Stereotact Funct Neurosurg. 1997; 68:10-17.
(4) Kanner et al. J Neurosurg 2002; 97:1115-1124.
(5) Martin et al. Radiology 2000;215:221-228.
(6) Nimsky et al. Radiology 2004;233:67-68.
(7) Sutherland et al. Acta Neurochir (Wien) 2003;85:21-28.
(8) Jankovski et al. Neurosurgery 2008;63:412-424.
No cost. No obligation. 30-day free trial.