Answers are grounded in this paper and cite their sources.
Intraoperative magnetic resonance imaging (iMRI) was first introduced to the neurosurgical operating suite in 1995.1,2 Multiple iMRI systems have since been developed that assist in the neurosurgical treatment of various nervous system diseases, but perhaps none more so than the field of neuro-oncology. iMRI is emerging as a valuable, high-resolution tool for neurosurgeons to improve the extent of resection of intracranial tumors and other lesions. The impact of extent of resection on outcomes for patients with intracranial gliomas has long been a topic of much debate. A Cochrane Database Systematic Review concluded that “there is no high quality evidence …” and “further large multicenter RCTs [randomized controlled trials] are required….”3 Given the substantial logistical and ethical limitations that will likely preclude any such future trial that randomizes solely on the basis of extent of resection, we are forced to rely on well-controlled observational studies. Such studies are accumulating and highlight the importance of increased extent of resection in the treatment of both high- and low-grade gliomas.4-9 Multiple observational studies also attest to the ability of both low- and high-field-strength iMRI to increase the extent of resection for glial tumors.10-13 Level 1 evidence, as defined by the Oxford Center of Evidence-based Medicine (http://www.cebm.net/index.aspx?o=1025), now exists that shows that iMRI also improves the extent of resection for gliomas.14 However, the utility of iMRI is not restricted solely to cases in which gross total resection is considered an achievable goal a priori. iMRI has been shown to safely maximize a subtotal resection during an awake craniotomy when the proximity of eloquent brain tissue makes complete removal prohibitive.15 The use of iMRI confers a nearly real-time assessment of the extent of a lesion resection, a frequent operative objective. Yet, several questions remain regarding the optimal application of this complex technology. To accurately assess the advantages and disadvantages of iMRI, outcomes analyses based on prospective data are required. Here, we report our initial experience using high-field-strength iMRI for the resection of intracranial gliomas, with particular emphasis on operating room efficiency, extent of resection, and patient outcome. PATIENTS AND METHODS Patients A ceiling-mounted, moveable 1.5-T iMRI was installed between 2 neurosurgical operative suites at the Barnes-Jewish Hospital in St. Louis, Missouri, in April 2008. With prior institutional review board approval, patients were identified from a prospective iMRI database established at the time of iMRI installation. Data collected included room use, operating time, patient demographics, symptomatology, surgery performed, clinical and pathological diagnoses, tumor dimensions, extent of resection, survival, and use of adjuvant treatments. In parallel, retrospective data have been collected specifically on patients treated for brain tumors with conventional microsurgery to compare treatment with and without the use of iMRI. iMRI Protocol All craniotomies in this study were performed with the specific preoperative goal of maximal safe resection. Biopsy cases were excluded. Chicoine et al16 previously outlined our iMRI protocol in detail. The time required to complete an iMRI case was strictly monitored and further categorized as time from room entry to skin incision, iMRI preparation time, and iMRI acquisition time. iMRI preparation time was defined as the time interval between the surgeon’s decision at the operative field that all surgical objectives were complete and the moment of the entrance of the iMRI scanner into the room. iMRI acquisition time was defined as the interval between the moment of the entrance of the iMRI scanner into the room and the moment it was completely resecured in a storage bay. The location of residual tumor tissue was examined and defined by 3 parameters: pattern, depth, and presence of overlying parenchyma. Residual tumor pattern was categorized as either focal or circumferential. The resection cavity was then divided into thirds to approximate the depth of residual tissue from the cortical surface. Finally, the presence or absence of overlying normal parenchyma on axial, coronal, and sagittal images was noted. Volumetric Extent of Resection Preoperative, intraoperative, and postoperative glioma tumor volumes were determined by manual segmentation of corresponding MRI scans with the Medtronic StealthStation S7 Surgical Navigation System (Medtronic Inc, Minneapolis, Minnesota). Only gadolinium-enhancing regions were selected as regions of interest during manual segmentation of glioblastomas. Tumor volumes were determined by a neurosurgeon without regard to survival data. A volumetric percentage of extent of resection was calculated from the following equation: [(preoperative tumor volume − postoperative tumor volume)/preoperative tumor volume] × 100. Statistics Time efficiency analyses were performed with a standard linear regression model and the Student t test. The χ2 or Fisher exact test was used to determine the influence of categorical variables. Patient survival differences were assessed with the Kaplan-Meier method with calculated P values from the log-rank test. Statistical significance was defined as P < .05. All analyses were performed with the SAS 9.1 statistical software package (SAS Inc, Cary, North Carolina). RESULTS Since the iMRI installation, > 1200 neurosurgical cases have been safely performed. Of these, 667 cases included the use of at least 1 iMRI scan, and > 500 cases were performed as if in a “standard” neurosurgical suite without iMRI. The most common indication for iMRI-guided surgery was the evaluation of extent of tumor resection (587 cases, 88%), with glioma representing the most frequent pathology treated (240 cases, 41%). An analysis of our initial 3½ years of iMRI operation included 180 intracranial glioma patients who underwent 188 iMRI craniotomies. The Table details pertinent demographic and tumor characteristics.TABLE: Patient Demographics and Tumor Characteristics (n = 188)Operative Time For an initial craniotomy on an adult glioma patient in the supine position, the mean operative time was 306 minutes when performed in a non-iMRI, standard room. Average operative time was 319 minutes for cases completed in an iMRI room when no scan was performed. This difference was not statistically significant compared with surgery in a standard room (P = .64; Figure 1). Glioma resections performed with iMRI were longer than standard room cases. Mean operative time was 418 minutes when no additional tumor resection followed iMRI (P < .001), whereas mean operative time was 468 minutes when additional tumor resection ensued (P < .001; Figure 1).FIGURE 1: Mean operative time according to room type and intraoperative magnetic resonance imaging (iMRI) use for an initial craniotomy performed in the supine position on an adult glioma patient. *Statistically increased operative time compared with a craniotomy performed in a standard room.Multiple time parameters were further monitored for efficiency and quality improvement analysis. Room entry to skin incision time (P = .01), iMRI acquisition time (P = .01), and total operative time (P < .001) all significantly declined with increasing iMRI experience (Figure 2A, 2C, and 2D). Average iMRI preparation time also declined but not significantly (P = .35; Figure 2B).FIGURE 2: Specific time component analysis for an initial craniotomy performed in the supine position on an adult glioma patient. Cumulative experience yielded significantly decreased the time from room entry to skin incision (A), intraoperative magnetic resonance imaging (iMRI) acquisition time (C), and total operative time (D). iMRI preparation time also declined but not significantly (B). Each data point represents the average of 5 consecutive patients.iMRI Impact Overall, iMRI use led to further surgical resection in 113 glioma cases (60%), with additional resection slightly more common for high-grade tumors (54% of 90 low-grade tumors vs 65% of 98 high-grade tumors; Figure 3A). Separately labeled post-iMRI specimens were taken in 74 cases. Histopathological examination confirmed the presence of residual tumor tissue in 85% of these cases, again with a higher frequency for malignant lesions (Figure 3B).FIGURE 3: Impact of intraoperative magnetic resonance imaging (iMRI) on glioma surgery. iMRI led to further glioma resection in 113 of 188 cases (60%; A). Further resection was more common for high-grade tumors (64 of 98 cases, 65%) than low-grade tumors (49 of 90 cases, 54%; A). Residual tumor tissue was identified at histopathological examination in 85% of all cases with specifically labeled post-iMRI specimens (n = 74), again with a higher frequency for high-grade tumors (B).Glioblastoma Experience Examination of the radiographic features of residual glioblastoma tissue noted at iMRI showed that a circumferential pattern was observed in 28% of iMRI scans, whereas 72% contained a focal tissue deposit (Figure 4A and 4B). Twelve percent of residual tumor foci were identified in the superficial third of the resection cavity, with 19% in the middle third and 69% in the deep third (Figure 4C and 4D). Overlying normal parenchyma was present for 31% of residual tumor foci (Figure 4E and 4F).FIGURE 4: Radiographic features of residual glioblastoma tissue found at intraoperative magnetic resonance imaging. Circumferentially oriented tumor (A) was less common than focal tumor deposits (B). Residual tumor was more common at greater depths from the cortical surface; only 12% of deposits were located in the superficial third of the resection cavity (C) compared with 69% in the deep third (D). Overlying normal parenchyma was absent in 69% of residual tumor deposits (E), whereas normal structures obstructed the view of residual tumor in 31% of cases (F). Arrows delineate residual tumor tissue.Complete volumetric imaging data were available for 41 glioblastoma cases. Mean preoperative glioblastoma volume was 36.5 cm3 (range, 2.9-97.4 cm3). Mean tumor volume noted at iMRI was 1.6 cm3 (range, 0-7.1 cm3), whereas the mean extent of resection at iMRI was 94%. Postoperatively, mean tumor volume was 0.4 cm3 (range, 0-7.1 cm3) with a 98% mean extent of resection. In total, the number of complete resections (ie, 100% extent of resection) increased from 24% to 57% after iMRI interpretation and subsequent additional resection. Awake Craniotomy With iMRI A cohort of 8 glioblastoma patients who underwent an awake craniotomy with iMRI was compared with age- and Karnofsky Performance Scale score--matched non-iMRI awake craniotomy control subjects (age: iMRI, 53 years; non-iMRI, 54 years; P = .55; Karnofsky Performance Scale score: iMRI, 80; non-iMRI, 80; P = .44). All patients completed adjuvant radiation and temozolomide-based chemotherapy. Median overall survival was 18.3 and 20.3 months for the iMRI and non-iMRI groups, respectively (P = .47; Figure 5A). Median progression-free survival was 9.3 months for the iMRI group but only 5.0 months for the non-iMRI group (P = .02; Figure 5B).FIGURE 5: Kaplan-Meier curves comparing glioblastoma patient survival after intraoperative magnetic resonance imaging (iMRI; solid line) and non-iMRI (dashed line) awake craniotomy. No difference was noted in overall survival between the 2 groups (iMRI, 18.3 months; non-iMRI, 20.3 months; P = .47; A). Significantly improved progression-free survival was noted for iMRI awake craniotomy patients compared with non-iMRI awake craniotomy controls (iMRI, 9.3 months; non-iMRI, 5.0 months; P = .02; B).DISCUSSION iMRI has been in existence for nearly 2 decades. Although initial systems were at times cumbersome compared with traditional microsurgery, recent advances like high-field-strength and moveable scanners have improved the feasibility of this intricate technology. Since the installation of a 1.5-T moveable iMRI at Barnes-Jewish Hospital, > 667 iMRI surgeries have been safely performed by the neurosurgeons of Washington University in St. Louis, Missouri. Although multiple nonneoplastic diseases have been effectively treated with the assistance of iMRI at our institution, the most widely applied use of this technology is for intracranial tumors. Mounting evidence shows improved outcomes after greater extent of resection for glial tumors.4-6,9 As cytoreduction becomes an increasingly important operative goal, techniques that gauge the extent of resection at the time of surgery become ever more valuable. This is especially true given that neurosurgeons, despite our best efforts, are particularly poor judges of extent of resection. Orringer et al17 recently reported their glioblastoma experience in which postoperative imaging confirmed residual tumor in 70% of cases previously deemed gross total resections by surgeon estimation. Radiographically confirmed complete resections occurred in only 24% of cases.17 iMRI provides nearly real-time feedback on the adequacy of tumor resection and, more important, does so at a time when further intervention remains possible. This particular advantage of iMRI is especially useful in pediatric neurosurgery. Shah et al18 reported a decreasing trend in the need for repeat surgery within 2 weeks of an iMRI-guided craniotomy for tumor or focal cortical dysplasia compared with conventional surgery. A common concern surrounding iMRI is increased operative time. In adult glioma patients undergoing initial surgery in the supine position, the average operative time in an iMRI suite was no different from that in a standard room if no intraoperative imaging was acquired. Not surprisingly, average operative time significantly increased when an intraoperative scan was performed and increased even further if the scan led to additional surgery. However, our specific time component analysis shows that improvements in efficiency accompany cumulative experience. Like other novel pieces of intraoperative equipment, iMRI has its own learning curve. Increased operative time is an unavoidable reality in iMRI cases. Yet, over the course of our iMRI program, the average operative time for glioma surgeries declined by 100 minutes. Improved nursing and physician experience, anesthesia specialization, equipment tray reorganization, and appropriate MRI sequence selection all contributed to the time reductions seen. A preliminary analysis of the first 3½ years of iMRI experience included 188 craniotomies for glioma. Of these total cases, iMRI led to additional tissue resection in 113 surgeries (60%), with a higher frequency for malignant lesions (65% high-grade vs 54% low-grade). Additional tissue resection does not guarantee additional tumor resection. To ensure that the residual abnormalities noted on intraoperative scans were indeed neoplastic tissue, cases with specifically labeled post-iMRI pathological specimens were examined. Of these 74 samples, 85% of cases contained residual tumor on histopathological examination. Interestingly, the location of residual glioblastoma tissue noted on intraoperative scans followed 3 trends. First, focal tissue deposits were more common than circumferentially oriented residual tumor surrounding the resection cavity. One might expect circumferentially oriented residual tissue with a more conservative approach to the resection. Indeed, circumferential tumor was noted in a number of awake procedures in which adjacent eloquent cortex was of particular concern. Second, the likelihood of residual tumor increased with increasing depth from the cortical surface. Finally, nearly a third of residual tumor foci were located beneath overlying normal parenchyma. Both increased distance from the cortical surface and normal overlying structures can obstruct a surgeon’s line of sight, thereby increasing the potential for occult tumor. To more precisely measure the effect of iMRI on glioblastoma management, we performed a volumetric analysis of extent of resection. Because of further surgery after iMRI evaluation, the mean extent of resection increased from 94% to 98%. Sanai et al8 examined 500 newly diagnosed glioblastoma cases and concluded that as little as a 78% extent of resection conferred an overall survival benefit. Interestingly, incremental improvements in survival were seen with increasing extent of resection, even when exceeding 95%.8 Alternatively, iMRI increased the number of 100% resections (ie, complete resections) from 24% to 57% in our series. These results are similar to the data reported in a recent clinical trial. Senft et al14 prospectively randomized patients with a suspected high-grade glioma to either conventional or iMRI-guided surgery. iMRI-guided surgery was statistically more likely to achieve a complete resection compared with conventional surgery, and a complete resection was associated with prolonged progression-free survival compared with an incomplete resection.14 Despite its limited size of 58 patients, this study provides the highest level of evidence that iMRI significantly affects the extent of resection achieved for glioma patients. Additional follow-up is required to detail the survival of our entire patient cohort. However, a pilot case-control study comparing awake craniotomy for glioblastoma cases with and without iMRI provides further initial insight into the effects of iMRI. Patients who underwent an awake craniotomy with iMRI experienced prolonged progression-free survival compared with age- and Karnofsky Performance Scale score--matched controls. A sample size of only 16 patients is of course too small to draw firm conclusions regarding the effect of iMRI on patient survival. However, the ability of iMRI to maximize an otherwise subtotal resection is noteworthy. Gross total resection need not be a preoperative goal to still impart a survival advantage, as is often the case with awake craniotomies in which the proximity of eloquent cortex prevents a more aggressive resection. CONCLUSION iMRI is a valuable tool in the neurosurgical suite and provides a nearly real-time assessment of the adequacy of tumor resection. The location of residual tumor tissue found on iMRI follows specific patterns. Although iMRI-guided craniotomies take longer compared with conventional microsurgery, cumulative iMRI experience progressively shortens operative time. iMRI improves the extent of resection for glial tumors, even when a gross total resection cannot be performed. These improvements in surgical outcome likely impart improved clinical outcome for patients with intracranial gliomas. For related video content, please access the Supplemental Digital Content: http://www.youtube.com/watch?v=mRkZR5gkseE Disclosure An unrestricted educational grant from IMRIS, Inc has helped establish and maintain a prospective iMRI database controlled by the Department of Neurological Surgery at Washington University School of Medicine. The authors have no personal financial or institutional interest in the drugs, materials, or devices described in this article.
No takes yet. Share an insight, caveat, or question.
Haydon et al. (2013) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: