Tumors arising around the skull base, such as pituitary adenomas, meningiomas and craniopharyngiomas, represent approximately 35% of all intracranial tumors (1-3). Because of their location and proximity to numerous delicate structures, these tumors pose some of the most complex challenges in neurosurgery(4). By avoiding skin incision, facial bone flap or craniotomy, endoscopic endonasal surgery (EES) presents a less externally invasive option for removal of these tumors, resulting in reduced trauma, decreased morbidity and mortality, and shorter hospital stays. The success of EES is heavily dependent on preoperative imaging and detailed evaluation before surgery (5). Magnetic resonance imaging (MRI) at ultrahigh field strengths, such as 7 Tesla (7T), is particularly effective at elucidating the delicate vasculature and cranial nerves adjacent to and involved with tumors as well as tumor composition and extent. The purpose of this pilot study is to assess the impact of high-resolution 7T anatomical and vascular imaging for improved surgical planning and navigation.

MR imaging: In this IRB-approved prospective study, 10 patients (8 females, age 48 ± 11) with skull base tumors (7 pituitary adenoma, 1 meningioma, 1 macroadenoma, 1 craniopharyngioma) underwent a high resolution scan on a 7T whole body scanner (Magnetom, Siemens) equipped with a 32Rx / 1Tx head coil. The following 3 acquisitions were performed: (i) 3D T1-weighted Magnetization-Prepared-Rapid-Gradient-Echo (MP2-RAGE) at 0.79 mm isotropic resolution and TI1/TI2/TR of 1050/3000/6000 s; (ii) Axial, sagittal and coronal T2-weighted turbo spin echo (TSE) at 0.39 x 0.39 x 2 mm3 resolution, turbo factor 7 and TE/TR 59 / 6000 ms; (iii) time of flight (TOF) non-contrast angiography at 0.35 x 0.26 x 0.40 mm3 resolution and TE/TR 5.6 / 18 ms. All acquisitions used parallel imaging acceleration R=2, and total acquisition time was 40-50 min, including initial localization and calibration scans.

Image analysis: Image quality was assessed by an expert neuroradiologist by comparing the 7T series (uniform T1-weighted (6), T2-weighted and angiogram) to clinical exams performed either at 3T (n=2 subjects) or a 1.5T (n=8). A 3-point scale (0 = not visualized, 1 = maybe visualized, 2 = confidently visualized) was used to assess visualization of the following key structures/features at each field strength: the pituitary stalk, delineation of normal pituitary gland from tumor, cranial nerves (CN II-VI) in the plan of the cavernous sinus, and arteries (OA = ophthalmic, ACA = anterior cerebral, MCA = middle cerebral, PcomA = posterior communicating, CA = choroidal).

Surgical procedure: All subjects underwent EES after the MR scans, and 7T images were loaded to the surgical navigation software (Brainlab) as shown in Fig. 1.

Table 1 summarizes the performance evaluation of 7T vs. clinical scans for the detection of CNs and arteries with medium to high confidence (score ≥ 1). All structures were better visualized at 7T compared to 1.5 and 3T. In addition, gland-tumor delineation was observed in 7/10 patients at 7T while only observed in 2/8 patients at 1.5T and not visible at 3T (0/2 patients). Figure 2 displays a sagittal view of the high resolution TSE of a meningioma patient at 7T, showing the large number of fine structures visible around the tumor. Figure 3 shows example images comparing 7T to 3T scans in two patients with pituitary adenomas. Both nerves and arteries appear more conspicuous on the higher resolution 7T images.As summarized in Table 1 and shown in Figures 2 and 3, many of the nerves, arteries and other structures proximal to the tumor, critical to surgical planning and resection, are often clearly resolved at 7T and insufficiently resolved at 1.5T and 3T. 7T imaging can be seamlessly integrated into neurosurgical planning and projected on the surgical navigation systems for guidance in the operating room. Using 7T, surgeons qualitatively reported a significant increase (approx. 50%) in confidence of critical decisions determining resection method, as well as decreased need for invasive procedures such as extensive drilling, which resulted in the total surgical time shortening on average from 179 mins (mean of 30 previous similar cases) to 125 mins (31%). The results of this work are immediately applicable to neurosurgical planning and guidance for brain tumors and could lead to the integration of 7T imaging into standard surgical procedure for other neurological conditions. The next phase of this study will be to perform a randomized clinical trial with a larger cohort, using a quantitative scale to assess improvement of neurosurgical efficacy as well as patient outcome after the integration of this technology.