HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Martin, A. J. Articles by Truwit, C. L. Search for Related Content PubMed PubMed Citation Articles by Martin, A. J. Articles by Truwit, C. L.

Brain Tumor Resection: Intraoperative Monitoring with High-Field-Strength MR Imaging-Initial Results¹

¹ From the Departments of Radiology (A.J.M., H.L. C.H.P., E.M., S.O.C., C.L.T.) and Neurosurgery (W.A.H., R.E.M.), University of Minnesota, 420 Delaware St, SE, Box 292, Suite J2-447, Minneapolis, MN 55455-0392. From the 1998 RSNA scientific assembly. Received February 3, 1999; revision requested March 1; final revision received June 30; accepted August 11. Address reprint requests to A.J.M. (e-mail: marti154@tc.umn.edu).

	Abstract

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 
PURPOSE: To investigate the challenges and benefits of magnetic resonance (MR) imaging during brain tumor resection.

MATERIALS AND METHODS: A short-bore 1.5-T MR system equipped with echo-planar–capable gradients was used in resection of brain tumors in 30 patients. MR sequences and need for contrast material enhancement were determined on the basis of the targeted lesion. MR images were acquired before, during, and after surgery. Tissue obtained at biopsy or excised as a result of intraoperative MR findings was examined histopathologically.

RESULTS: MR images of enhancing lesions proved to be the most challenging to interpret intraoperatively, and relative enhancement at the resection cavity boundary was not specific for residual tumor. The ability to detect residual tumor intraoperatively resulted in a radiologically complete resection in 24 (80%) of 30 patients. The frequency of complications was low, and no untoward effects related to the MR environment were observed.

CONCLUSION: Intraoperative MR imaging provided valuable information on the completeness of resection, and resection progress was well demonstrated during surgery.

Index terms: Brain, surgery, 10.45 • Brain neoplasms, 13.3611, 13.362, 13.363, 13.366, 13.38, 14.36, 15.3637 • Brain neoplasms, MR, 10.121411, 10.121412, 10.121413, 10.121416, 10.121417, 10.12143, 10.12144 • Magnetic resonance (MR), guidance

	Introduction

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 
The introduction of image-guidance systems into neurosurgical practice has helped improve noninvasive visualization of the surgical field. The majority of such image-guidance systems, however, rely on computer manipulation of data that were acquired hours or days prior to surgery. For tumor resection, these image-guidance systems can be helpful for determination of the optimal placement of the craniotomy, provided that no substantial changes in brain morphology have occurred between imaging and surgery. However, the degree to which preoperative data accurately reflect surgical reality after a bone flap has been turned, the dura punctured, and the tumor partially resected is in question (1–4).

For these reasons, some form of intraoperative imaging is desirable to allow intermittent monitoring of the progress of surgery. Ultrasonography (US) (5–7), computed tomography (CT) (8–10), and magnetic resonance (MR) imaging (11–13) have all been proposed as potential modalities to help obtain this information. US has the obvious advantages of mobility and cost but may not provide satisfactory image quality or help delineate the tumor boundary with high specificity (7). Moreover, virtually all preoperative images are acquired with MR imaging or CT; thus, correlation with intraoperative US can be difficult.

Early demonstrations of intraoperative CT took place more than 15 years ago (8,9), but intraoperative CT has not achieved an important role in neurosurgical practice. The development of mobile CT (10), which precludes the need for a dedicated imaging and surgical suite, makes intraoperative use of CT more practical. CT uses ionizing radiation, however, and has restrictions on scanning plane orientation. CT image contrast also is more limited than and tends to be inferior to MR image contrast. In addition, CT lacks the diverse functional capabilities that MR imaging can provide. Many of these capabilities, such as those that are possible with functional MR brain activation studies, can be used to aid neurosurgical planning and monitoring.

It is, therefore, desirable to use MR imaging to monitor the progress of brain tumor resection. The integration of an MR imaging system into the neurosurgical suite, however, requires substantial compromises with regard to the surgical environment and/or the capabilities of the MR imager. Indeed, there are several fundamentally different types of imagers and configurations of surgical suites that have been proposed for intraoperative MR imaging. If imaging is required during surgery, then a configuration such as that developed by Schenk et al (14) and demonstrated by Black et al (11) might be desirable. Because image updates for guidance and monitoring of tumor resection typically are required only at discrete stages, however, it may be preferable to remove the patient from the MR unit when imaging is not being performed. If patient transit is acceptable, then the restrictions placed on the MR imager and surgical environment are markedly reduced. The choices of magnet configuration and field strength then largely become a trade-off between cost, siting, and system capabilities.

The justification for the use of MR imaging, rather than CT or US, stems from the fact that it can provide clearer and more comprehensive diagnostic information. A high-field-strength MR system with appropriate gradient capabilities produces high-quality anatomic images, provides excellent MR angiographic capabilities, permits identification of the eloquent cortices by means of blood oxygenation level–dependent imaging (15), and allows MR spectroscopic interrogation. These properties make high-field-strength interventional MR imaging ideal for comprehensive monitoring of tumor resection at neurosurgery. The purpose of the present study was to investigate the challenges and benefits of MR imaging during brain tumor resection. This article presents our initial experiences with high-field-strength intraoperative MR imaging during brain tumor resection in 30 patients.

	MATERIALS AND METHODS

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 
MR Imaging System
The MR imaging system was a short-bore 1.5-T MR unit (Gyroscan ACS-NT; Philips Medical Systems, Best, the Netherlands) equipped with strong imaging gradients (23 mT/m, 105 mT/m/msec) and commercially available, state-of-the-art pulse sequences such as echo-planar imaging (EPI), half-Fourier rapid acquisition with relaxation enhancement (16), and gradient–spin-echo (17) sequences. The main field magnet was actively shielded and produced a 5-G (0.5-mT) "footprint" that covered a peanut-shaped area with principal axial dimensions of 7.8 x 5.0 m. The total length of the MR system was 180 cm, and the inner-bore diameter of 60 cm extended to 100 cm at the flared openings. A head coil consisting of two circular loops that were combined as a phased array was designed for interventional applications. This coil offered excellent access to the patient while helping maintain high image quality.

The MR system we used has the same imaging capabilities as other state-of-the-art diagnostic imagers. The EPI-capable gradient set permitted performance of functional studies such as brain activation (functional MR imaging) and perfusion- and diffusion-weighted imaging. The high field strength permitted acquisition of spectroscopic data by means of single-volume or chemical shift imaging. Real-time interactive MR imaging had been implemented and allowed on-the-fly manipulation of geometric and contrast parameters. Images were reconstructed and displayed immediately after acquisition both at the console and at a set of four MR-compatible liquid-crystal display screens that were positioned in the MR suite next to the MR system. The four screens were attached to a flexible arm and could be freely moved along a ceiling-mounted rail system. This made in-room viewing during interventional procedures possible at any location, both at the front and at the back of the magnet.

Surgical Suite
A pedestal table suitable for neurosurgery was positioned parallel to the long axis of the imager and was adapted to attach to the MR system (Fig 1). The tabletop moved freely along the resultant track and could be locked at the isocenter of the magnet, on the surgical pedestal, or at the distal end of the magnet, where the table extended 40 cm beyond the opening of the bore. A carbon-fiber frame (Malcom-Rand; Elekta Instruments, Atlanta, Ga) and titanium stereotactic frame (model CRW; Radionics, Burlington, Mass) were customized to fit the tabletop and to permit exact reproduction of imaging planes even after extensive surgical periods.

View larger version (106K): [in this window] [in a new window] [Download PPT slide]   Figure 1. The interventional MR suite is shown from the front, with the 1.5-T MR imager in the background. The surgical and angiography pedestal is visible in the foreground and is just outside the 5-G (0.5-mT) footprint of the imager. The floating tabletop moves on a continuous track between the MR unit and the surgical pedestal. The ceiling-suspended liquid-crystal display panel (arrow) is visible on its articulating arm. A secondary surgical area is located at the rear opening of the magnet, where the table extends 40 cm beyond the end of the imager to improve access to the patient.

The surgical suite also contained MR-compatible anesthesia equipment (Ohmeda, Liberty Corner, NJ), patient monitoring equipment (In Vivo, Orlando, Fla), surgical microscope (Leica, Heerbrugg, Switzerland), craniotome (Midas Rex, Fort Worth, Tex), fiberoptic headlight (Cogent Light Technologies, Santa Clarita, Calif), bipolar cautery equipment (Codman and Shurtleff, Randolph, Mass), and numerous surgical instruments. A blunt-tipped side-cutting prototype biopsy needle (Elekta Instruments) was used when biopsy samples were obtained beyond the surgical cavity.

The interventional MR suite met operating room specifications and was also used for routine diagnostic imaging between surgical procedures. Before a surgical procedure, the room was terminally cleaned and subsequently treated as a sterile surgical environment. Owing to the potential safety risk caused by the fringe magnetic field, a special color-coded, pocketless uniform was used to both identify interventional MR personnel and minimize the opportunity for inadvertent introduction of ferrous objects into the suite.

Patients
MR imaging–monitored brain tumor resection was performed in 30 patients between May 1997 and September 1998. Informed consent was received from all patients, in compliance with the requirements of our institutional review board. There were 18 male and 12 female patients aged 14 months to 70 years (mean, 35 years), including nine pediatric patients. Seventeen patients were undergoing initial resection of their lesion, and the remainder were undergoing repeat surgery.

At the time of arrival at the interventional MR suite, all patients received general anesthetics, and invasive arterial blood pressure monitoring was established. After anesthetic induction, patients were placed in a head frame in the surgical position, and a portion of the scalp was shaved. An MR-visible marker grid was then placed on the patient's head to help the neurosurgeon optimize the size and location of the craniotomy bone flap. Immediately after the scalp had been shaved and prior to the initial incision, patients were transferred to the magnet for baseline MR imaging.

Perioperative MR Imaging
Preoperative MR imaging was performed to obtain baseline images with which intraoperative images would be compared and to delimit the necessary extent of the craniotomy. Functional MR imaging was performed immediately before anesthetic induction in three patients with lesions in close proximity to eloquent cortex. Functional MR imaging data were used for spatial mapping of the motor cortex, language, and memory areas. The functional MR imaging protocol was a single-shot EPI sequence (3,000/50 [repetition time msec/echo time msec], 220-mm field of view, 64 x 64 matrix, 14 6-mm-thick sections acquired with no intersection gap). This acquisition was repeated 72 times in sequential fashion while the patient performed a series of activity and rest phases. Areas of blood oxygenation level–dependent activation were calculated at the imager console immediately after acquisition, were superimposed on high-quality anatomic images, and were presented to the neurosurgeon prior to surgery.

After initial imaging, the patient was removed from the magnet and transferred to the surgical pedestal (Fig 2), where the craniotomy and preliminary resection were performed. Additional MR images were subsequently obtained when the neurosurgeon requested an image update or judged the resection to be complete on the basis of visual inspection. At this time, all non–MR-compatible instruments were removed, and the patient was moved into the magnet bore.

View larger version (151K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Surgical configuration during tumor resection is illustrated. The MR imager (arrow) is in the background, and a neurosurgeon is working with an MR-compatible microscope just outside the 5-G (0.5-mT) footprint of the imager. This arrangement does not encumber the neurosurgeon and permits use of non-MR-compatible surgical instruments (shown in the foreground).

Tissue samples from lesions were routinely sent for histopathologic assessment; however, tissue removed as a direct result of its intraoperative MR imaging appearance was specially labeled, and a separate histopathologic examination was performed. Additional image updates were repeated as necessary until a radiologically complete resection was achieved or the surgery was terminated for other clinical reasons. Resection completeness was assessed by the attending neuroradiologist (C.L.T., C.H.P., E.M., or S.O.C.), in collaboration with the neurosurgeon (W.A.H. or R.E.M.). At the completion of the procedure, additional images were obtained to screen for complications.

Patient Follow-up
All patients were routinely imaged at 3-month intervals for the 1st year after tumor resection. The frequency of follow-up imaging beyond the 1st year varied from patient to patient but typically was one to two times per year. The results of these follow-up examinations were evaluated by the attending neuroradiologist (C.L.T., C.H.P., E.M., or S.O.C.) in collaboration with the neurosurgeon (W.A.H. or R.E.M.) to determine whether disease recurrence or progression had occurred.

	RESULTS

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 
All 30 patients successfully underwent MR imaging–monitored brain tumor resection. There were no untoward events related to MR-compatible instrumentation or intraoperative patient monitoring. Tumor histologic results included 11 cases of glioblastoma multiforme (GBM), eight of astrocytoma (seven low-grade, one anaplastic astrocytoma), three of meningioma, two of ganglioglioma, two of metastases to brain, one of oligodendroglioma, one of craniopharyngioma, one of medulloblastoma, and one of teratoma.

Intraoperative MR Imaging
MR imaging immediately prior to surgery provided baseline images with which to compare subsequent intraoperative MR images. This imaging, combined with the MR-visible surface markers, proved to be an effective way to localize and minimize the size of the craniotomy bone flap.

Preoperative functional MR data were acquired in three patients, and areas of activation were identified directly at the MR system console within 5 minutes. Superimposition of these data on high-quality diagnostic MR images (Fig 3) enabled the neurosurgeon to identify lesion margins where particular caution was warranted. One patient experienced postoperative apraxia after complete resection of a lesion that directly abutted the MR image–defined motor strip. The apraxia resolved completely within a week and likely arose from involvement of the supplementary motor area.

View larger version (57K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Low-grade glioma in the medial portion of the left frontal lobe in a 40-year-old man who underwent MR imaging-monitored tumor resection. A, Transverse single-shot EPI functional MR imaging (3,000/50; EPI factor of 63; field of view, 210 mm; section thickness, 7.0 mm with no intersection gap; matrix, 64 x 64; 72 repetitions) was performed to help identify the proximity of the motor strip to the lesion. Activation patterns during a paradigm of right finger tapping (light blue pixels) were determined with inversion-recovery MR imaging (4,869/14/170; echo train length of nine; field of view, 225 mm; section thickness, 4.0 mm with 0.8-mm intersection gap; matrix, 256 x 192) and are superimposed on the anatomic image. The "on-off" period (dark blue line) in this activation paradigm and the signal intensity within the region of activation (yellow line) are also superimposed. B-D, Resection progress was monitored with coronal turbo FLAIR MR imaging (6,000/100; field of view, 210 mm; section thickness, 3.0 mm with 0.01-mm overlap; echo train length of 27; matrix, 256 x 192). B, The lesion (arrow) is demonstrated on this preoperative image. C, After initial resection, residual tumor (arrow) is seen. D, After complete resection, no residual tumor is seen, although biopsy results from tissue sampled at the margin of the MR image-demarcated area of resection showed infiltration of tumor cells.

Cases of a nonenhancing lesion were routinely followed with turbo FLAIR and half-Fourier rapid acquisition with relaxation enhancement acquisitions. Images obtained with these sequences demonstrated a consistent appearance of the lesion throughout surgical interventions. The half-Fourier rapid acquisition with relaxation enhancement sequence allowed rapid T2-weighted imaging, and the turbo FLAIR sequence allowed high-quality T2-weighted imaging (Fig 3, B–D) with suppression of signal intensity from cerebrospinal fluid. Biopsy results from a sample from the periphery of a radiologically complete resection of the low-grade glioma illustrated in Figure 3, however, demonstrated cellular infiltration beyond the MR image–demarcated (T2 high-signal-intensity) boundary of the lesion.

Radical resection of the full edematous extent of a high-grade enhancing GBM lesion was performed in one patient (Fig 4, A–D). Owing to the potential for important neurologic deficit, only the enhancing portion of a high-grade lesion typically was resected. At the completion of all procedures, additional images were routinely obtained to help assess for complications. Postoperative MR imaging often included turbo FLAIR (Fig 4, E) and T2*-weighted fast low-angle shot (Fig 4, F) imaging to screen for hemorrhage and diffusion-weighted (Fig 4, G, H) imaging to screen for ischemia.

View larger version (86K): [in this window] [in a new window] [Download PPT slide]   Figure 4. High-grade GBM in a 57-year-old man who underwent initial resection. A-D, Transverse turbo FLAIR MR images (6,000/100; field of view, 210 mm; section thickness, 3.0 mm with 0.01-mm intersection gap; echo train length of 27; matrix, 246 x 192). A small enhancing focus of this lesion (arrow in A) is surrounded by a large region of increased signal intensity (arrowheads). Because the patient was in the left decubitus position, a fluid level (arrow in C) was visible on many sections. Surgical resection of the full extent of this lesion was monitored to completion. E-H, Postoperative transverse MR images to assess for hemorrhage, stroke, or both, included turbo FLAIR images (E), which were acquired with same parameters as for A-D; T2*-weighted fast low-angle shot images (F) (450/30; flip angle, 25°; field of view, 230 mm; section thickness, 4.0 mm with 0.8-mm intersection gap; matrix, 256 x 192); and diffusion-weighted multishot EPI images (G, H) (repetition time, duration of two heartbeats; echo time, 105 msec; field of view, 230 mm; section thickness, 6.0 mm with 1.0-mm intersection gap; EPI factor, 15; matrix, 128 x 96). In G, b = 0 sec/mm2; in H, b = 1,000 sec/mm2. After resection, the surgical cavity was filled with saline solution, which produced the high-signal-intensity area in F and G.

View larger version (131K): [in this window] [in a new window] [Download PPT slide]   Figure 5a. Right frontal lobe brain metastasis in a 63-year-old man. Coronal contrast-enhanced T1-weighted spin-echo MR images (400/14; field of view, 230 mm; section thickness, 5.0 mm with 0.5-mm intersection gap; matrix, 256 x 192). (a) Preoperative image shows the superficial lesion (arrow). (b) After complete resection, diffuse enhancement around the surgical bed remains.

View larger version (135K): [in this window] [in a new window] [Download PPT slide]   Figure 5b. Right frontal lobe brain metastasis in a 63-year-old man. Coronal contrast-enhanced T1-weighted spin-echo MR images (400/14; field of view, 230 mm; section thickness, 5.0 mm with 0.5-mm intersection gap; matrix, 256 x 192). (a) Preoperative image shows the superficial lesion (arrow). (b) After complete resection, diffuse enhancement around the surgical bed remains.

View larger version (33K): [in this window] [in a new window] [Download PPT slide]   Figure 6. Brain metastasis in the left occipital lobe of a 53-year-old man. Transverse T1-weighted spin-echo MR images (400/14; field of view, 230 mm; section thickness, 5.0 mm with 0.5-mm intersection gap; matrix 256 x 192). A, Image obtained before craniotomy to demarcate the necessary extent of the bone flap shows the metastatic lesion (arrow). B, C, After initial resection, images obtained immediately before (B) and after (C) administration of contrast material shown relative enhancement (arrow in C) along the inferior and anterior wall of the resection cavity. Histologic results in tissue from the wall of the cavity did not reveal confluent tumor. D, E, After additional resection, images obtained before (D) and after (E) administration of contrast material demonstrate the increased difficulty in interpretation after multiple doses of contrast material.

View larger version (110K): [in this window] [in a new window] [Download PPT slide]   Figure 7a. GBM in the left frontal lobe of a 57-year-old man. Intraoperative transverse T1-weighted spin-echo MR images (400/14; field of view, 230 mm; section thickness, 5.0 mm with 0.5-mm intersection gap; matrix, 256 x 192) obtained immediately (a) before and (b) after administration of contrast material show relative enhancement (arrow in b) along the posterior wall of the resection cavity. Tissue from this region was histologically determined to contain confluent tumor. A fluid level (arrow in a) at the posterior aspect of the resection cavity is evident in a and b.

View larger version (114K): [in this window] [in a new window] [Download PPT slide]   Figure 7b. GBM in the left frontal lobe of a 57-year-old man. Intraoperative transverse T1-weighted spin-echo MR images (400/14; field of view, 230 mm; section thickness, 5.0 mm with 0.5-mm intersection gap; matrix, 256 x 192) obtained immediately (a) before and (b) after administration of contrast material show relative enhancement (arrow in b) along the posterior wall of the resection cavity. Tissue from this region was histologically determined to contain confluent tumor. A fluid level (arrow in a) at the posterior aspect of the resection cavity is evident in a and b.

The radiologically incomplete cases included one GBM lesion, two astrocytomas, one ganglioglioma, the oligodendroglioma, and the teratoma. The neurosurgical reasons for incomplete resection included excessive brain swelling in the patient with GBM; infiltration of tumor into eloquent cortex in two patients with astrocytoma, one with ganglioglioma, and one with oligodendroglioma; and substantial intraoperative bleeding in the patient with a teratoma. In two of the four patients who underwent limited resection because of the proximity of the lesion to eloquent cortex (one with astrocytoma, one with ganglioglioma), surgery was performed to improve or alleviate epileptic seizure activity. The remaining two patients (one with astrocytoma, one with oligodendroglioma) who underwent incomplete resection due to the intracerebral location of the lesion were treated with radiation therapy.

The neurosurgeons were asked to comment on whether they believed that intraoperative MR imaging improved the completeness of the resection they performed. An improvement in resection completeness was claimed to have occurred in 17 (57%) of 30 patients. In four of the remaining 13 patients, MR was considered to have been helpful for avoiding vessels or eloquent cortex during resection.

Recurrence or Progression
All tumor resections reported in this article were performed in 1997 and 1998; thus, only preliminary trends can be established at this time. Of the 30 patients, 15 (50%) underwent initial resection of a primary brain tumor, 13 (43%) underwent resection of a primary brain tumor that had previously been resected at least once, and two (7%) underwent initial resection of brain metastasis.

At the time of this writing, none of the 15 patients who underwent initial resection of a primary brain lesion showed disease progression or recurrence. The average time between surgery and the most recent imaging follow-up was 6.6 months in these patients. This group included seven patients with low-grade astrocytoma (average follow-up, 8.2 months), three with meningioma (average follow-up, 5.8 months), two with ganglioglioma (average follow-up, 5.5 months), two with GBM (average follow-up, 4.75 months) and one with teratoma (follow-up, 3 months). Complete resection was achieved in 11 (73%) of these 15 patients. The four patients who underwent incomplete resection either responded well to postoperative chemotherapy (one patient with teratoma) or radiation therapy (one with astrocytoma) or underwent surgery to alleviate epileptic seizures (one with ganglioglioma, one with astrocytoma).

Of the 13 patients who underwent repeat resection, 11 experienced recurrence or progression of disease, and five of the 11 died. The patients with recurrent or progressive disease included eight with GBM (average time to recurrence, 6.5 months), one with anaplastic astrocytoma (time to recurrence, 11 months), one with craniopharyngioma (time to recurrence, 6.5 months) and one with medulloblastoma; the latter patient died after shunt failure. In only one patient who experienced recurrence or progression of GBM was resection radiologically incomplete. The two patients in whom disease had not recurred or progressed included the patient with oligodendroglioma (follow-up, 5.5 months) and a patient with GBM (follow-up, 2.5 months). The two patients who underwent resection of a brain metastasis showed no progression of brain lesions (average follow-up, 2.75 months).

Complications
The frequency of complications was low and is comparable to that encountered after procedures performed in a conventional operating room. Among the 30 patients, one experienced a postoperative infection with Propionibacterium acnes, and one had cerebral peduncular ischemia during a hippocampectomy. The brain abscess due to P acnes infection occurred 6 weeks after surgery and was presumed to be due to intraoperative failure of sterile technique. The patient with the small peduncular infarct, although initially hemiparetic, responded well to physical therapy and had only a minor deficit. The overall infection rate for the MR-surgery suite, spanning more than 100 neurosurgical procedures (including brain biopsies, functional neurosurgical procedures, shunt placements, and laminectomies), was less than 2% (two of 101) at the time of this writing.

	DISCUSSION

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 
The role of resection completeness for the long-term prognosis in patients with brain tumor has yet to be well established and certainly will be a function of the location, type, and grade of tumor. There is, however, increasing evidence that optimization of the initial resection of a low-grade glioma is correlated with improved clinical outcome, as measured in terms of neurologic function or survival (18,19). Results of studies (20–22) of malignant gliomas have suggested a possible benefit for radical initial resection. Albert et al (20) further established that the surgical impression of resection completeness can markedly differ from the early postoperative radiologic evaluation. Thus, the development of intraoperative imaging systems that provide the feedback necessary to ensure a radiologically complete resection is of crucial importance in this debate. A thorough analysis of rates of recurrence, time to recurrence, and survival in individuals who are determined at the time of the procedure to have undergone radiologically complete resection will be necessary to establish the effectiveness of intraoperative MR imaging monitoring.

The interpretation of intraoperative MR images can be challenging. Intraoperative MR images can demonstrate air, saline, edema, hemorrhage, and diffuse residual contrast enhancement that were not present on preoperative images. Indeed, Forsyth et al (23) determined that postoperative assessment of malignant gliomas should be performed 3–5 days after the resection. This interval was found to minimize the confounding effects of postoperative nontumorous marginal enhancement and of methemoglobin in the surgical bed.

Ambiguity at the boundary of the resection cavity can be a problem with any lesion. The region of confluent tumor for a low-grade lesion often is well demarcated by the region of high signal intensity on T2-weighted or turbo FLAIR MR images. However, as has been reported by other investigators (24), we detected cellular infiltration by tumor cells in biopsy samples from tissue around the surgical cavity after radiologically complete resection of low-grade glioma. The presence of infiltration beyond the MR image–demarcated tumor boundary may be an important indicator of the need for subsequent therapy, even after complete resection.

Lesions that are best demarcated after injection of contrast material create additional challenges for interpretation of the completeness of resection. These lesions typically have a marked edematous region of high T2 signal intensity that extends well beyond the focal region of enhancement. Cellular infiltration of the tumor beyond the region of enhancement and into the region of elevated T2 signal intensity is known to occur. In most instances, however, it is practical to target only the region of enhancement for surgical resection.

The difficulties associated with perioperative use of contrast material stem from the fact that the material can accumulate in the resection cavity and infiltrate along its borders. This phenomenon is likely due to the intravascular nature of the MR contrast material and to surgical disruption of the blood-brain barrier. Whenever possible, preoperative administration of contrast material was avoided to prevent these anomalous enhancement patterns. If contrast material had previously been administered, then MR imaging was performed immediately before and after injection of additional contrast material. This approach was helpful, because regions of diffuse enhancement typically did not show acute relative enhancement after fresh injection of contrast material. However, relative enhancement at the boundary of the resection cavity was noted in some cases in which no tumor was evident in excised tissue specimens. By monitoring enhancement dynamics, some additional insight may be gained for classifying these lesions (25); however, similar inconsistencies at the resection boundary may be noted. In addition, surgical resection can alter or eliminate feeding vessels to the tumor and thereby substantially change or influence its enhancement properties. These limitations make it difficult to obtain a conclusive intraoperative characterization of residual tumor solely on the basis of enhancement properties.

The ability to achieve a complete resection must always be weighed against the potential for causing an important neurologic deficit. Thus, simple demarcation of lesional volume does not provide the neurosurgeon with sufficient information to perform a successful procedure. Identification of eloquent cortex, blood vessels, or both, provides equally valuable data to the clinician. Indeed, the identification of these structures may become increasingly important as more aggressive surgical resections are performed. The need for high-quality anatomic, angiographic, and functional data emphasizes the value of a state-of-the-art high-field-strength MR system. Whereas high-field-strength systems offer limited patient access during imaging, it has proven to be practical to move the patient into the magnet at the discrete time points when image updates are needed.

In conclusion, MR imaging monitoring of brain tumor resection was practical with a short-bore, high-field-strength interventional MR system. Resection progress was well demonstrated with intraoperative MR imaging, and a radiologically complete resection was achieved in a large percentage of patients. Intraoperative MR images of enhancing lesions proved to be the most challenging to interpret, and relative enhancement at the boundary of the resection cavity was found not to be specific for residual tumor. The results of a preliminary outcome analysis support the hypothesis that optimal initial resection of brain lesions may result in improved patient outcome. No incidents related to the MR system or the use of MR-compatible equipment were noted, and sterility was not compromised by operating in this environment.

	Footnotes

 
Abbreviations: EPI = echo-planar imaging FLAIR = fluid-attenuated inversion recovery GBM = glioblastoma multiforme

Author contributions: Guarantors of integrity of entire study, A.J.M., W.A.H., C.L.T., H.L.; study concepts and design, A.J.M., W.A.H., C.L.T., H.L.; definition of intellectual content, A.J.M., W.A.H., C.L.T.; literature research, A.J.M., W.A.H.; clinical studies, W.A.H., C.L.T., R.E.M., C.H.P., E.M., S.O.C.; data acquisition, A.J.M., H.L.; data analysis, A.J.M.; manuscript preparation, A.J.M.; manuscript editing, A.J.M., W.A.H., C.L.T.; manuscript review, all authors.

	References

TOP Abstract Introduction MATERIALS AND METHODS RESULTS DISCUSSION References

 

1. Dorward NL, Alberti O, Velani B, et al. Postimaging brain distortion: magnitude, correlates, and impact on neuronavigation. J Neurosurg 1998; 88:656-662.[Medline]
2. Hill DLG, Maurer CR, Maciunas RJ, Barwise JA, Fitzpatrick JM, Wang MY. Measurement of intraoperative brain surface deformation under a craniotomy. Neurosurgery 1998; 43:514-526.[Medline]
3. Roberts DW, Hartov A, Kennedy FE, Miga MI, Paulsen KD. Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery 1998; 43:749-758.[Medline]
4. Maurer CR, Jr, Hill DL, Martin AJ, et al. Investigation of intraoperative brain deformation using a 1.5-T interventional MR system: preliminary results. IEEE Trans Med Imaging 1998; 17:817-825.[Medline]
5. Le Roux PD, Berger MS, Wang K, Mack LA, Ojemann GA. Low grade gliomas: comparison of intraoperative ultrasound characteristics with preoperative imaging studies. J Neurooncol 1992; 13:189-198.[Medline]
6. LeRoux PD, Winter TC, Berger MS, Mack LA, Wang K, Elliot JP. A comparison between preoperative magnetic resonance and intraoperative ultrasound tumor volumes and margins. J Clin Ultrasound 1994; 22:29-36.[Medline]
7. Woydt M, Krone A, Becker G, Schmidt K, Roggendorf W, Roosen K. Correlation of intra-operative ultrasound with histopathologic findings after tumor resection in supratentorial gliomas. Acta Neurochir 1996; 138:1391-1398.[Medline]
8. Shalit MN, Israeli Y, Matz S, Cohen ML. Intra-operative computerized axial tomography. Surg Neurol 1982; 17:376-382.[Medline]
9. Lundsford LD, Parrish R, Albright L. Intraoperative imaging with a therapeutic computed tomographic scanner. Neurosurgery 1984; 15:559-561.[Medline]
10. Butler WE, Piaggio CM, Constantinou C, et al. A mobile computer tomographic scanner with intraoperative and intensive care unit applications. Neurosurgery 1998; 42:1304-1310.[Medline]
11. Black PM, Moriarty T, Alexander E, et al. Development and implementation of intraoperative magnetic resonance imaging and its neurosurgical applications. Neurosurgery 1997; 41:831-842.[Medline]
12. Tronnier VM, Wirtz CR, Knauth M, et al. Intraoperative diagnostic and interventional magnetic resonance imaging in neurosurgery. Neurosurgery 1997; 40:891-900.[Medline]
13. Steinmeier R, Fahlbusch R, Ganslandt O, et al. Intraoperative magnetic resonance imaging with the Magnetom open scanner: concepts, neurosurgical indications, and procedures—a preliminary report. Neurosurgery 1998; 43:739-747.[Medline]
14. Schenk JF, Jolesz FA, Roemer PB, et al. Superconducting open-configuration MR imaging system for image-guided therapy. Radiology 1995; 195:805-814.[Abstract/Free Full Text]
15. Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci USA 1990; 87:9868-9872.[Abstract/Free Full Text]
16. Kiefer B, Grassner J, Hausman R. Image acquisition in a second with a half-Fourier-acquisition single-shot turbo spin echo (abstr) In: Proceedings of the First Meeting of the Society of Magnetic Resonance. Berkeley, Calif: Society of Magnetic Resonance, 1994; 86-87.
17. Feinburg DA, Oshio K. GRASE (gradient- and spin-echo) MR imaging: a new fast clinical imaging technique. Radiology 1991; 181:597-602.[Abstract/Free Full Text]
18. Berger MS, Deliganis AV, Dobbins J, Keles GE. The effect of extent of resection on recurrence in patients with low grade cerebral hemisphere gliomas. Cancer 1994; 74:1784-1791.[Medline]
19. Gajjar A, Sanford RA, Heideman R, et al. Low grade astrocytoma: a decade of experience at St Jude Children's Research Hospital. J Clin Oncol 1997; 15:2792-2799.[Abstract]
20. Albert FK, Forsting M, Sartor K, Adams HP, Kunze S. Early post-operative magnetic resonance imaging after resection of malignant glioma: objective evaluation of residual tumor growth and its influence on regrowth and prognosis. Neurosurgery 1994; 34:45-61.[Medline]
21. Shapiro WR, Shapiro JR. Biology and treatment of malignant glioma. Oncology 1998; 12:233-240.[Medline]
22. Black P. Management of malignant glioma: role of surgery in relation to multimodality therapy. J Neurovirol 1998; 4:227-236.[Medline]
23. Forsyth PAJ, Mahallati H, Cairncross JG, et al. Prospective study of postoperative magnetic resonance imaging in patients with malignant gliomas. J Clin Oncol 1997; 15:2076-2081.[Abstract/Free Full Text]
24. Watanabe M, Tanaka R, Takeda N. Magnetic resonance imaging and histopathology of cerebral gliomas. Neuroradiology 1992; 34:463-469.[Medline]
25. Schwartz RB, Hsu L, Kacher DF, et al. Intraoperative dynamic MRI: localization of sites of brain tumor recurrence after high-dose radiotherapy. J Magn Reson Imaging 1998; 8:1085-1089.[Medline]

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Martin, A. J. Articles by Truwit, C. L. Search for Related Content PubMed PubMed Citation Articles by Martin, A. J. Articles by Truwit, C. L.