Lowfield Intraoperative

As demonstrated by other groups (10,11,16-20) the use of MR scanners in the operating environment seems to be safe and reliable, as well as applicable to neurosurgical procedures, even if these procedures have to be adapted to the MR environment to a certain extent. Even intraoperative patient transport, which was necessary in our first conceptual design combining intraoperative imaging with microscope-based neuronavigation, did not cause any problems. The development of new navigation microscopes, which can be used in the fringe field of the MR scanner (21), allowed us to abandon the cumbersome and time-consuming intraoperative patient transport. It is now possible to integrate full microscope-based neuronavigational support with the concept of surgery in the fringe field, which was first applied in transsphenoidal surgery (12,22) and afterwards proposed for open cranial surgery (23).

There are two main designs for intraoperative MRI. In the first, the patient is placed directly inside the scanner, using a system specially designed for intraoperative use, like the Signa SP "double-doughnut" scanner (10,24), which offers almost real-time imaging with the drawback of restricted patient access. The 0.12-T PoleStar N-10 (Odin Technologies, Yokneam, Israel) has a similar design: the disk-shaped magnets are placed below the operating table and moved upward for scanning (25,26). In the second type, the standard diagnostic scanners adapted to the operating environment require some kind of patient transport. The patient is operated on either in the fringe field of the scanner, lying on the movable MR tray (21), or on a rotating operating table (27), or else surgery is performed in a separate operating room, as in the twin operating room setup, necessitating a longer intraoperative patient transport (11,12). Both of these designs are implemented with the 0.3-T AIRIS II scanner (Hitachi Medical Corporation, Tokyo, Japan) (28).

Presently we consider that intraoperative MRI is indicated for the surgical treatment of gliomas (especially low grade), ventricular tumors, epilepsy (29), and complicated larger suprasellar pituitary tumors (22). In addition, intraoperative MRI should be used to compensate for the effects of brain shift, if ongoing neuronavigational guidance is needed in complicated cases with major distortion of the surgical field (30-33). Further indications, not investigated by our group, may be biopsy procedures with additional therapy control provided by the scanner, e.g., temperature monitoring in cryoablation or laser ablation (27).

Operating Room Setup

All procedures were performed under general anesthesia. The local ethical committee approved the intraoperative MRI with intraoperative patient transport, and all patients signed an informed consent. Timing of intraoperative MRI was decided by the neurosurgeon. MRI was performed either when the neurosurgeon had the impression of complete tumor resection or, in case of incomplete tumor removal, when the neurosurgeon felt that no further tumor removal was possible, e.g., because of an infiltration of eloquent brain areas.

Intraoperative MRI was performed using a 0.2-T Magnetom Open MR scanner (Siemens Medical Solutions, Erlangen, Germany), which was located in the radiofrequency-shielded part of a twin operating theater (12,34). Two operating sites were mainly used: either an adjacent conventional operating theater, necessitating intraoperative patient transport, or the extended table of the MR scanner at the 5-G line.

The concept of surgery in the fringe field of the scanner was first applied in transsphenoidal surgery, because in these cases no navigation microscope was needed, so that a conventional operating microscope could be used. The head of the patient was placed directly on the movable table of the MR scanner at the 5-G line. A standard flexible coil was attached around the head. For safety reasons, when working closer than 1.5 m from the magnet isocenter, MR-compatible instruments, especially developed for this purpose, were used (35). Prior to imaging, an MR-compatible speculum was inserted. For intraoperative scanning the table was placed in the center of the magnet, and then data acquisition was started. In the routine setup coronal and sagittal T1-weighted spin-echo sequences were acquired (slice thickness 3 mm, TR 340 ms, TE 26 ms, bandwidth 39 Hz, FOV 200 mm, matrix 192 x 256) . In selected cases, e.g., in cystic adenomas, T2-weighted turbo spin-echo sequences were used in addition (slice thickness 3 mm, TR 5700 ms, TE 117 ms, bandwidth 33 Hz, FOV 230 mm, matrix 224 x 256). For craniopharyngiomas, T1-weighted imaging was repeated after contrast media was applied (gadolinium-DTPA 0.2 mL/kg body weight IV).

Fig. 1. Twin operating room setup: the initial concept combining microscope-based neuronavigation using the MKM microscope with intraoperative low-field imaging (0.2-T Magnetom Open). (A) View from the RF-cabin to the MKM microscope, which is placed in an adjacent operating room. (B) Opposite view. (C) Intraoperative patient transport, the draped patient, laying on an air-cushioned OR table has just been moved to the scanner. (D) Schematic drawing of the twin operating room setup.

Fig. 1. Twin operating room setup: the initial concept combining microscope-based neuronavigation using the MKM microscope with intraoperative low-field imaging (0.2-T Magnetom Open). (A) View from the RF-cabin to the MKM microscope, which is placed in an adjacent operating room. (B) Opposite view. (C) Intraoperative patient transport, the draped patient, laying on an air-cushioned OR table has just been moved to the scanner. (D) Schematic drawing of the twin operating room setup.

For transcranial surgery of gliomas, other brain tumors, and drug-resistant epilepsy, the head was fixed in a ceramic, MR-compatible head holder. Using the MKM navigation microscope (Zeiss, Oberkochen, Germany), which was the only navigation microscope available in 1996, surgery was performed in an adjacent operating room (twin operating room concept). The patient was placed on an air-cushioned OR table, and the table was transported into the scanner during surgery over a distance of 5 m (Fig. 1) (12,34). With the introduction of the NC4 navigation microscope (Zeiss), which consists of only a few magnetic parts, intraoperative patient transport could be abandoned, and all procedures could be performed in the same position as for transsphenoidal surgery (Fig. 2) (21). For imaging, an experimental separable coil was used. The lower, nonsterile part was applied prior to surgery; the sterile upper part of the coil was placed onto sterile adapters just before the head was moved into the center of the scanner.

For glioma surgery, axial inversion recovery (slice thickness 6 mm, TR 6000 ms, TE 48 ms, TI 300 ms, bandwidth 65 Hz, FOV 250 mm, matrix 182 x 256) and dark fluid sequences (slice thickness 6 mm, TR 6000 ms, TE 93 ms, TI 1600 ms, bandwidth 65 Hz, FOV 230 mm, matrix 210 x 256) were measured. Additionally, volume data were obtained routinely using a T1-weighted 3D fast low-

Fig. 2. Surgery at the fringe field, the NC4 microscope is placed at the 5-G line, so no long, time-consuming intraoperative patient transport is necessary. (A) Same view as in Fig. 1A. (B) Microscope position near the scanner, a flat screen depicts the navigational views with integrated functional data. (C) For scanning, the MR tray is moved into the center of the scanner. (D) Schematic drawing.

Fig. 2. Surgery at the fringe field, the NC4 microscope is placed at the 5-G line, so no long, time-consuming intraoperative patient transport is necessary. (A) Same view as in Fig. 1A. (B) Microscope position near the scanner, a flat screen depicts the navigational views with integrated functional data. (C) For scanning, the MR tray is moved into the center of the scanner. (D) Schematic drawing.

angle shot (FLASH) gradient-echo sequence (slice thickness 1.5 mm, TR 16.1 ms, TE 7 ms, bandwidth 98 Hz, FOV 250 mm, matrix 256 x 256). This sequence was used for multiplanar reformatting to obtain standard projections, which compensates for the fact that the head was fixed intraoperatively in variable angles. This sequence was also the prerequisite for an intraoperative update of neuronavigation. MR contrast agent (gadolinium-DTPA 0.2 mL/kg body weight, IV.) was administered just prior to scanning only if the tumor showed enhancement in the preoperative images. For resection control in cavernoma, other brain tumors, and nonlesional cases of epilepsy surgery, a reduced scanning protocol was used.

Each OR position allowed unlimited patient access and full use of standard neurosurgical microinstrumentation. Either pointer-based neuronavigation (Stealth, Medtronic, Broomfield, CO) or microscope-based neuronavigation (MKM or NC4 navigation microscopes, Zeiss) were used. In addition to anatomical neuronavigation, functional data from MEG or fMRI to identify eloquent brain areas were integrated. Functional data were measured with a MAGNES II biomagnetometer (BTI, San Diego, CA) and a 1.5-T Magnetom Symphony (Siemens Medical Solutions) (13-15,36-38).

In case of a suspected tumor remnant, several MR-visible bone fiducials (Howmedica-Leibinger, Freiburg, Germany) were placed around the cran-iotomy opening prior to intraoperative scanning, allowing an intraoperative registration of the 3D image data to update neuronavigation. The 3D image data were then transferred to the navigation system via intranet, and the tumor remnant was segmented manually with the software of the respective neuronavigation systems (e.g., STP 4.0, Zeiss, Oberkochen, Germany). Then the neuronavigation system was reregistered with the intraoperative image data, allowing compensation for the effects of brain shift (31,32).

In case of further tumor removal, repeated imaging was performed either directly after removal of the tumor remnant or at the end of surgery, i.e., after craniotomy closure. Contrast media was not applied repeatedly.

Clinical Experience

Between March 1996 and July 2001, we performed intraoperative MRI in 330 patients. Histopathologic examination revealed pituitary adenoma in 61, craniopharyngioma in 26, glioma in 106, cavernoma in 18, and other diagnoses in 54 patients (including chordoma, gangliocytoma, germinoma, lymphoma, melanoma, meningioma, metastasis, neurocytoma, pinealoma, primitive neu-roectodermal tumor, and subependymoma). For remaining 65 cases, either resective or disconnective epilepsy surgery was carried out for nonlesional conditions (n = 40), or procedures were performed in which intraoperative MRI was used as an online control during catheter or electrode placements (n = 25). Among the 330 procedures, there were 240 craniotomies, 59 transsphenoidal approaches, and 31 burr hole procedures.

We did not observe any adverse effects of intraoperative imaging on the postoperative course of the patients. Intraoperative patient transport (applied in 166 patients) did not cause any harm. On average, a delay of some 10 min occurred until imaging could start. The same time was necessary until surgery could be continued after imaging, so that intraoperative patient transport was felt to be time consuming and cumbersome. A further time delay occurred when only an anesthesiology team without experience in intraoperative patient transport was available, since all anesthesia lines had to be switched twice for intraoperative imaging. In the patients who were operated on directly on the MR table at the 5-G line (n = 164), scanning started on average after 20 s to 1 min. The overall complication rate with respect to wound infection (n = 3) and rebleeding (n = 1) was a bit lower than the average in our department.

Transsphenoidal Procedures

Among 59 patients who underwent transsphenoidal pituitary surgery (52 pituitary adenomas, 7 craniopharyngiomas), image evaluation was limited in only 16, either because drilling artifacts from metal debris disturbed image interpretation in the sella region, or because blood in the resection cavity mimicked tumor remnants. With increasing experience in image interpretation, as well as repeated intraoperative scanning and the application of T2-weighted sequences, further evaluation was possible in most of the patients. In 30

patients a repeat inspection of the surgical field was performed. This led to further tumor removal in 17 out of the total 59 patients. Among a subgroup (n = 44) of resectable pituitary macroadenomas with a distinct suprasellar extension, the further resection of tumor increased the rate of complete removal from 45% (21/44) to 75% (33/44). In the remaining patients, an extensive supra- and parasellar tumor extension did not allow complete transsphenoidal removal, necessitating surveillance, secondary transcranial surgery, or radiation therapy. In only one patient was there a transient postoperative visual impairment, caused by a secondary empty sella after complete tumor removal. In none of the 59 patients was endocrine function disturbed owing to the surgical intervention (22).

In transsphenoidal pituitary surgery, intraoperative MRI provides a possibility for further tumor removal in the case of residual tumor, thus increasing the rate of complete tumor removals significantly (39,40). Removal of the suprasel-lar tumor portion could be evaluated reliably. However, low-field MRI did not seem to be suitable for evaluation of the removal of intrasellar microadenomas. Also, interpretation of the extent of removal of parasellar, i.e., intracavernous, tumor parts was limited. Image artifacts caused by metal debris from drilling or by blood accumulation in the resection cavity were further challenges. In case of incomplete tumor removal intraoperative MRI allowed additional therapy planning at a much earlier state; usually further diagnostics have to be delayed for 2-3 mo after surgery, to obtain artifact-free postoperative images (22).

Glioma Surgery

For glioma surgery, intraoperative MRI showed remaining tumor in a high percentage (63%). In 42% of these patients we continued tumor resection and tried to remove as much tumor as possible; in more than 50%, neuronavigation was updated by the intraoperative image data to localize the tumor remnants. Further tumor removal increased the gross total removal rate, especially for low-grade gliomas; however, in some patients, despite further tumor removal, complete resection was not performed since some tumor remnants invaded eloquent brain areas, which were identified by functional neuronavigation.

Intraoperative MRI in glioma surgery (n = 106) revealed incomplete tumor removal in 67 (63%); in one additional case intraoperative imaging did not allow a clear distinction between artifacts and tumor remnants. In 34 patients, i.e., 51% (34 of 67) of the patients with incomplete resection, the resection cavity was inspected again. In 28 patients (42% of the patients with tumor remnants, i.e., 26% of all glioma patients), the extent of resection could be expanded based on the results of intraoperative MR scanning. This increased the rate of gross total removal in World Health Organization (WHO) grade I tumors (n = 15) from 87 to 100%. In grade II tumors (n = 37) further tumor removal was performed in 19 patients. In 10 of them, complete tumor removal was finally achieved. Thus, the gross total removal rate in the grade II tumors increased from 27 to 54%. In the remaining nine patients, despite further tumor removal, complete resection was not possible because small tumor remnants had infiltrated eloquent brain areas. In the high-grade astrocytomas, the extent of resec tion was enlarged in seven patients only. This resulted in an increased gross total removal of 55% vs 41% in the grade III and 22% vs 19% in the grade IV patients.

In eight of the low grade gliomas (WHO grade II), the resection could not be extended primarily, because eloquent brain areas were infiltrated. In the majority (82%: 32 out of 39) of the high-grade gliomas (grades III and IV), for which intraoperative imaging had depicted incomplete removal, it was the policy above all to avoid new neurological deficits.

Forty-five glioma patients had preoperative neurological deficits, which resolved at least partially in 32 and worsened in only 2 patients. In the group of patients without preoperative neurological deficits (n = 61), three developed a postoperative neurological deterioration (two pareses, and one aphasia). In five glioma patients (4.7%), a neurological postoperative deficit occurred; in only two of them had functional neuronavigation been applied.

We used functional neuronavigation if the tumor was located near eloquent areas to preserve neurological function (14,15,36,37), as a limitation for total resection. In addition to anatomical data, functional data from MEG (n = 64) or fMRI (n = 15), identifying either motor, sensory, or language-related cortex in 66 patients, were integrated into the neuronavigation, which allowed identification and avoided damage to eloquent brain areas during surgery. The identification of these cortical areas at the beginning of tumor removal was not affected by brain shift, since the functional markers from MEG or fMRI were used to identify the eloquent cortex just after dural opening. The application of functional neuronavigation was associated with low postoperative morbidity. In only two of the 66 patients (3%) in whom functional neuronavigation was applied did a persistent neurological deterioration occur, which in both patients fortunately improved in the further postoperative course. Anatomical and functional neuronavigation were used as guidance to identify relevant structures. Intraoperative MRI allowed a delineation of the extent of resection, so the combination of both allowed a maximum possible resection with the least neurological deficits, while taking incomplete tumor removal into account, when eloquent brain areas were infiltrated.

Besides the possibility of evaluating the completeness of a tumor resection, intraoperative MRI not only allowed us to delineate and visualize the extent of brain shift, which results in great inaccuracies of neuronavigation systems in an ongoing operation, but also offered the possibility of compensating for the effects of brain shift. This could be achieved reliably by an intraoperative update of the neuronavigation system with the intraoperative 3D image data (31-33). Of the 28 patients in whom resection was continued after intraoperative imaging, an intraoperative update of neuronavigation was performed in 18 (64%; 17% of the total 106 gliomas). The update procedure was performed only when it was difficult to localize the remaining tumor in the surgical field. In all these cases the intraoperative update of neuronavigation was technically successful. Image transfer, new segmentation, and rereferencing added another 15-20 min to the whole procedure. The registration error, calculated as the root mean square error, was low (mean 0.85 mm, standard deviation 0.44 mm, max imum error 2.3 mm), resulting in a reliable navigation with high accuracy that quickly led to the suspicious areas (31,32). In three patients the histopathologi-cal examination revealed no tumor in the tissue that was removed additionally because of intraoperative imaging. Complete tumor removal was confirmed by early scanning in all three of them.

One of the general limitations of glioma surgery is that by histological definition a 100% tumor removal is not possible. Furthermore, the extension of the tumor into eloquent brain areas will also result in only partial removal. According to our experience with intraoperative MR evaluation of glioma removal, we still doubt the long-term benefit of intraoperative MRI in surgery of high-grade gliomas, although the first reports published on this topic claim a benefit even in high-grade tumors (16,18,41). In the high-grade tumors, differentiation among tumor remnants, surgically induced imaging changes, edema, and normal brain may be very difficult. Controversial reports on life expectancy in high-grade gliomas emphasize that life expectancy depends more on low postoperative deficits than on "macroscopic" total tumor removal (42). In the low-grade gliomas (43), at a microscopic level complete resection also will not be possible, but since survival of patients seems to be correlated with the extent of tumor resection (44), we believe that surgery of these tumors will benefit more from intraoperative imaging (45). It is still too early to determine the effects on life expectancy in these patients, but it can be stated that more radical resections are possible without increasing morbidity, especially when intraoperative imaging is supported by the use of functional neuronavigation (14,15). Long-term survival studies are still needed.

Other Procedures

Epilepsy surgery is another indication for intraoperative MRI; it has led to reliable evaluation of the extent of resection or disconnection procedure. Not only did it allow us to evaluate completeness of resection in the lesional cases (e.g., glioma, cavernoma), but also in the nonlesional cases the exact extent of a callosotomy (n = 6) or the size and shape of a tailored temporal lobe resection (n = 34) in its neocortical and hippocampal extent could be documented (29). Increased knowledge of structure-function relationships as partially defined by intraoperative imaging may reduce adverse neuropsychological sequels of epilepsy surgery in the future (29,46,47).

Besides the application of intraoperative MRI in epilepsy surgery, the extent of tumor resection was also evaluated in transcranial pituitary tumor surgery and in various other nongliomatous brain tumors. Furthermore, intraoperative MR resection control of ventricular tumors (n = 8) in particular proved helpful. In surgery of ventricular tumors, the cerebrospinal fluid CSF loss quickly resulted in inaccurate neuronavigation. Intraoperative imaging, with the possibility of an update of the neuronavigation system, allowed us to obtain a good orientation in the surgical field and to evaluate the extent of the resection. Catheter and electrode placements, as well as cyst punctures (craniopharyn-giomas and other cysts), were reliably controlled by intraoperative MRI in all cases (n = 31). We did not encounter problems in visualizing the catheters and electrodes. In only one case was an electrode not easily visualized. Intraoperative X-ray showed that in this patient the depth electrode deviated considerably from the preplanned path into the subdural space.

Was this article helpful?

0 0
Reducing Blood Pressure Naturally

Reducing Blood Pressure Naturally

Do You Suffer From High Blood Pressure? Do You Feel Like This Silent Killer Might Be Stalking You? Have you been diagnosed or pre-hypertension and hypertension? Then JOIN THE CROWD Nearly 1 in 3 adults in the United States suffer from High Blood Pressure and only 1 in 3 adults are actually aware that they have it.

Get My Free Ebook


Post a comment