3-T ultrahigh-field intraoperative MRI for low-grade glioma resection

Neurosurgery is an ever-changing and -developing medical field. The history of this field, which is merely longer than a century, is loaded with dramatic advancements comparable to the rapid development of a child in infancy. In the second half of this century almost every decade has brought a new development to make surgery safer and more effective. Microsurgery in the 1960s, computerized tomography in the 1970s, MRI in the 1980s, skull-base surgery in the 1990s and, finally, the introduction of high technology to the operating room in the 2000s. Development of intraoperative MRI (ioMRI) is the most striking example for this latest phase, and with current developments, ioMRI is the single most prominent strategy that carries the potential to dramatically improve patient outcome in glioma surgery.

The drive for the development of intraoperative imaging came from the field of neuro-oncology, owing to a need to improve the effciency of tumor resection. The role of surgery in the treatment of glial tumors is one of the controversial issues in neurosurgical practice. Glial tumors grow in an infiltrative fashion, which precludes a true oncological resection of these tumors [1]. However, studies have shown that a more radical resection is associated with longer survival, better quality of life and lower chance of malignant degeneration over time [1–3]. In addition to this, prospective randomized, multicenter trials have shown that neither radiotherapy nor chemotherapy have a significant effect on patient survival in low-grade glioma (LGG). Currently, the most effective form of treatment for LGG is complete surgical excision [2]. Studies have also indicated that patients do benefit from the extent of this resection, with better outcome after more complete resections [1–3].

Factors that limit a complete resection in glioma surgery are several fold. The first and foremost factor is the infiltrative nature of the tumor and, therefore, surgery is aimed at removing the mass lesion identifiable by neuroimaging. Obtaining tumor-negative margins is not an objective. Infiltration of functional brain tissue of eloquent or indispensable function is another absolute contraindication for radical resection. However, even resection of tumor in relatively silent areas of the brain poses significant problems. The tumor–normal brain interface cannot be reliably determined by visual inspection alone. Studies with the application of early postoperative MRI have shown that the surgeon’s impression during surgery is far from an objective evaluation on the extent of resection [4]. This led to a search for supportive technologies to guide the surgeon during resection of LGG. Technologies such as ultrasonography, navigation and tumor fluorescence have all created much enthusiasm; however, each of these methods was limited by technological difficulties. Intraoperative application of MRI, on the other hand, has resulted in a significant reduction in recurrence and death rate in LGG [5]. This is not unexpected, because MRI is currently the standard in preoperative LGG imaging and postoperative evaluation of the surgical result (extent of resection), and intraoperative availability of this information could certainly help to improve surgical results.

The first ioMRI unit to become operational was installed at Brigham and Women’s hospital in Boston (MA, USA) in 1994 [6]. This was a 0.5-T scanner with a double-donut design (SIGNA SP, General Electric Medical Systems, WI, USA). In this design, both the patient and the surgeon were inside the magnet, where the surgery was performed using all nonferromagnetic surgical equipment. Several other low-field (0.2–0.5 T) designs worldwide followed this first application, including the Erlangen [7] and Toronto Siemens low-field systems (Magnetom OPEN; Siemens Medical Systems, Erlangen, Germany) and the Hitachi system (Fonar, NY, USA; Hitachi Medical, OH, USA) [8]. Currently, we are starting to see short- to mid-term follow-up results of surgeries performed using low-field systems [5,9]. In a recent review of their results, Claus et al. have shown that the odds ratio for recurrence is 1.4, and the odds ratio for death is 4.9 when subtotal versus total resection using ioMRI were compared [5].

The next major development in the field came with the adoption of high-field equipment in the operating room. The image quality and signal–noise ratio of the MRI improves with increasing magnet strength and magnetic gradients. Low-field systems suffer from low anatomical resolution, slow speed and the lack of the most sophisticated MRI sequences, which are indispensable in today’s MRI technology. Use of 1.5-T equipment dramatically improved anatomical resolution and enabled acquisition of T2-weighted images, which are the standard MRI sequence for LGG. Several high-field systems were installed worldwide, including the Philips system at the University of Minnesota (MN, USA) [10], Siemens systems at the University of Erlangen (Germany) [11] and the University of California (CA, USA) and the IMRIS system at Calgary (Canada) [12]. Reports have shown that these systems were capable of increasing the extent of resection [13,14].

Increased magnet strength is associated with improved image quality, albeit at the expense of bulky equipment. An alternative strategy was the development of lower-cost, low-bulk systems, exemplified by the 0.12-T system, which had permanent magnets of 40 cm diameter placed 25 cm apart from each other [15]. Use of a 0.15-T version of ultra-low field designs for LGG surgery was reported by Johann Wolfgang von Goethe University in Germany (Polestar N20 system, Medtronic Navigation, CO, USA) [9].

High-field scanners were followed by the introduction of ultra-high-field (3 T) systems. Very few ultrahigh-field systems have been reported in the literature so far, including the 3-T-Philips systems at University of Minnesota [10,16], the 3-T-Siemens at Acibadem University, Turkey [17,18] and finally the second Philips system at Cliniques Universitaires St-Luc, Université Catholique de Louvain, Belgium [19]. The first clinical LGG series was reported by our institution and showed an increased gross total resection rate comparable to the results obtained by low- and high-field systems [17]. One major development with ultrahigh-field systems was the routine application of very high-resolution T2-weighted images as the standard imaging sequence. Other exciting developments also followed: Application of proton magnetic resonance spectroscopy and diffusion-weighted imaging (DWI) to differentiate peritumoral T2-weighted MRI changes in normal parenchyma from residual tumor, which can be problematic even in postoperative imaging, was performed without technical difficulties [17]. DWI was used to monitor surgically induced ischemia [17]. Previous studies using low-field imaging had already shown that complications, such as intraoperative bleeding could be demonstrated using ioMRI [20]. With the use of high- and ultrahigh-field ioMRI, intraoperative diffusion tensor imaging was also performed to visualize the corticospinal tract [21,22]. All of these technologies are still being validated, investigated further and undergoing refinement, and their impact on long-term patient outcome and quality of life still needs to be investigated.

The design of the operating room has also varied. The initial SIGNA SP system was designed to be dedicated operative equipment [6]. Furthermore, as the surgery was performed inside the magnetic field, all surgical equipment, including the operative microscope, had to be nonferromagnetic. All of these factors accounted for high setup and running costs. These costs increase further with increasing magnet strength. Financing such a big system is a major burden for most academic or private institutions. An alternative was offered by Siemens [7] and Hitachi [8] low-field systems by designing shared-resource facilities where the MRI gantry and the operating room were separated and the MRI was used both for operative and out-patient purposes. Similar designs were developed for high-field systems [23]. Siemens ioMRI design reverted to a single dedicated operating room in the Siemens 1.5-T brain suite design [11,24]. The latest system by Siemens, installed at our institution, again used the twin-room philosophy to reduce costs by allocating the equipment to diagnostic out-patient imaging at times when MRI is not needed intraoperatively [17,18]. A similar design was also adopted by the Philips system installed in Belgium [19].

The conclusions after 10 years of experience with ioMRI for neurosurgery are as follows [2,5,8,9,11,13,14,17,21–26]. Current ioMRI technology has very high anatomical resolution but also provides functional information (e.g., functional MRI, DWI, magnetic resonance spectroscopy, diffusion tensor imaging, magnetic resonance angiography and magnetic resonance venography). ioMRI complements preoperative and postoperative imaging. The technology is near real-time, improves precision of navigation, aids in preservation of normal brain tissue and function and provides a reliable assessment of the completeness of tumor resection. Finally, ioMRI can exclude immediate surgical complications.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.