- Email: [email protected]
Low field intra-operative magnetic resonance imaging for brain tumour surgery: Preliminary experience
ARTICLE IN PRESS n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
NEUROCIRUGÍA www.elsevier.es/neurocirugia
Clinical Research
Low field intra-operative magnetic resonance imaging for brain tumour surgery: Preliminary experience夽 Pedro Roldán a,∗ , Sergio García a , Josep González a , Luis Alberto Reyes a , Jorge Torales a , ˜ a Ricard Valero b , Laura Oleaga c , Joaquim Ensenat a b c
Servicio de Neurocirugía, Hospital Clínic de Barcelona, Barcelona, Spain Servicio de Anestesiología y Reanimación, Hospital Clínic de Barcelona, Barcelona, Spain Servicio de Radiodiagnóstico, Hospital Clínic de Barcelona, Barcelona, Spain
a r t i c l e
i n f o
a b s t r a c t
Article history:
Objectives: Intra-operative magnetic resonance imaging (iMRI) is a recently introduced tool
Received 12 June 2016
in the most advanced neurosurgical operating rooms worldwide. We present our preliminary ®
Accepted 10 August 2016
experience in brain tumour surgery with low field PoleStar N30
Available online xxx
its introduction in 2013 in the Barcelona Clinic Hospital.
Keywords:
on using iMRI and intention of complete removal up to October 2015. A record was made of
intraoperative MRI since
Material and methods: A prospective non-randomised study was conducted on cases operated Glioma
the data as regards surgical times, resection rates, histological diagnosis, hospital stay, and
Craniotomy
survival rates during follow-up.
Intra-operative magnetic resonance
Results: The study included 50 patients, with a mean age of 55 years (±13.7), a preoperative
imaging
mean Karnofsky of 92 (being 81 post-operatively), and a mean follow-up of 10.5 months
Neuronavigation
(±6.5). There were 26% re-operations due to recurrence. High-grade gliomas were reported in
Intra-operative neuroimaging
56%, low-grade gliomas in 24%, and 20% “Other” tumours. Overall hospital stay was 10 days
Neurosurgery
(±4.5). Depending on the histologiacl diagnosis, the “Others” group had a longer hospital stay. Overall, there were 52% complete removal, 18% of maximum removals, and 30% of partial removals. The overall survival rates during follow-up was 84%. Conclusions: iMRI is a safe and effective tool for brain tumour surgery. Its use allows an increase in resection rates, and minimises post-operative complications. Its implementation involves an increase in surgical time, which improves with the characteristic learning curve. More studies are needed to establish its role in the long-term survival of patients. ˜ ˜ S.L.U. All rights © 2016 Sociedad Espanola de Neurocirug´ıa. Published by Elsevier Espana, reserved.
夽 Please cite this article as: Roldán P, García S, González J, Reyes LA, Torales J, Valero R, et al. Resonancia magnética intraoperatoria de bajo campo para la cirugía de neoplasias cerebrales: experiencia preliminar. Neurocirugía. 2017. http://dx.doi.org/10.1016/j.neucir.2016.08.001 ∗ Corresponding author. E-mail address: [email protected] (P. Roldán). ˜ ˜ S.L.U. All rights reserved. 2529-8496/© 2016 Sociedad Espanola de Neurocirug´ıa. Published by Elsevier Espana,
NEUCIE-254; No. of Pages 8
ARTICLE IN PRESS 2
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
Resonancia magnética intraoperatoria de bajo campo para la cirugía de neoplasias cerebrales: experiencia preliminar r e s u m e n Palabras clave:
Objetivos: La resonancia magnética intraoperatoria (RMi) es una herramienta recientemente
Glioma
introducida en los quirófanos de neurocirugía más vanguardistas mundialmente. Presenta-
Craneotomía Resonancia magnética
mos nuestra experiencia clínica preliminar con relación al empleo de la RMi de bajo campo, PoleStar N30® , desde su implementación en 2013 en el Hospital Clínic de Barcelona, para
intraoperatoria
el tratamiento de neoplasias cerebrales.
Neuronavegación
Material y método: Se realizó un estudio prospectivo no aleatorizado incluyendo los casos
Neuroimagen intraoperatoria
intervenidos mediante RMi con intención de resección completa hasta octubre de 2015. Se
Neurocirugía
registraron los tiempos quirúrgicos así como los grados de resección, diagnóstico histológico, estancia hospitalaria y la supervivencia durante el seguimiento. ˜ Resultados: Se incluyeron 50 pacientes con edad media de 55 anos (±13,7), un Karnofsky preoperatorio de 92 (siendo el postoperatorio de 81); y un seguimiento medio de 10,5 meses (±6,5). Un 26% fueron reintervenciones por recidiva. Un 56% eran gliomas de alto grado, un 24% gliomas de bajo grado y un 20% otras neoplasias. La estancia hospitalaria global fue de 10 días (±4,5). Según el diagnóstico histológico el grupo «otras» fue el que mayor estancia hospitalaria presentaba. Globalmente, se lograron un 52% de resección completa, un 18% de resecciones parciales máximas y un 30% de resecciones parciales. La supervivencia durante el seguimiento fue del 84%. Conclusiones: La RMi es una herramienta segura y eficaz en la cirugía de neoplasias cerebrales. Su uso permite aumentar el grado de resección disminuyendo las complicaciones posquirúrgicas. Su empleo conlleva una prolongación del tiempo quirúrgico que mejora con la curva de aprendizaje característica. Más estudios son necesarios para poder establecer su papel en la supervivencia a largo plazo de los pacientes. ˜ ˜ S.L.U. Todos © 2016 Sociedad Espanola de Neurocirug´ıa. Publicado por Elsevier Espana, los derechos reservados.
Introduction The development of imaging techniques has had an unquestionable impact on the development of neurosurgery. Computed tomography, introduced in the 1970s, and magnetic resonance imaging (MRI), introduced in the 1990s, have been implemented in neuronavigation systems. Thanks to surgical planning, simulation and navigation, more complex procedures can be performed with ever-fewer invasive techniques. At present, the most widespread navigation systems are socalled frameless systems. Navigation can be affected by events occurring during initial recording, or by changes in intracranial tissues and fluids during surgery. Brain shift may render recorded preoperative data inaccurate. It is estimated that shifts of up to 1 cm per 1 h occur after opening the dura mater in more than half of patients studied. This shift is greater in cases of tumour volume resection or hydrocephaly, thereby drastically decreasing the usefulness of navigation during the procedure.1–6 Intraoperative imaging was developed to overcome these disadvantages. Current intraoperative imaging modalities include intraoperative computed tomography,7 intraoperative ultrasounds8 and intraoperative MRI (iMRI). The problems encountered in bringing MRI into the operating theatre included electromagnetic interference as well as device size, mobility and magnetic isolation. In 1997, Black et al.9 first
used a novel iMRI system allowing high-quality images to be acquired and used as a navigation system during neurosurgical procedures.10–13 The potential benefits of this technology would allow intraoperative confirmation of degree of tumour resection (DoTR). This would contribute to an increase in complete resection (CR) and a decrease in postoperative complications, morbidity and length of hospital stay, which would in turn contribute to an increase in quality of life and overall survival.14–20 Since this first publication in 1997,11 the development of iMRI has gradually increased in significance and sites worldwide have equipped their surgeons with this very useful yet very expensive technology. There are 3 major types of iMRI depending on magnetic field intensity–high-field (1.5–3 teslas [T]), medium-field (0.5–1.5 T) and low-field (0.12–0.5 T)–with advantages and disadvantages inherent to the performance of each. In this study we present preliminary experience with the use of low-field iMRI, PoleStar N30 Medtronic® , to assist neuro-oncology surgery at Hospital Clínic Universitari in Barcelona, since it was implemented in 2013 (Fig. 1).
Material and methods All patients who underwent surgery for brain tumours with the intention of CR and iMRI assistance were analysed
ARTICLE IN PRESS 3
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
®
Fig. 1 – Panoramic view of the “smart” operating theatre with PoleStar N30 low-field iMRI system at Hospital Clínic in Barcelona.
prospectively. The sample was recruited from October 2013, when the system was implemented at our site, to October 2015. All patients signed the informed consent form authorising their enrolment in the framework intraoperative imaging study, “SMART OR”, conducted at the site. The same inclusion and exclusion criteria as those established for conventional neuro-oncology surgery were used and the standard safety regulations for patients examined with MRI were applied. iMRI was performed in all cases of intra-axial brain lesions with the intention of CR where MRI use was not contraindicated and was permitted by patient physiognomy. The neurosurgeon selected the imaging modality on a case-by-case basis taking into account the best technique in each case. The surgical technique for neuro-oncology surgery assisted by low-field iMRI has already been previously reported by other authors7,21 and is not discussed in this study. We evaluated parameters such as surgery time (ST), placement time (PT), DoTR, postoperative morbidity, length of hospital stay and survival. To measure DoTR, preoperative images (3 T) were compared to early postoperative images (up to 72 h after surgery). To evaluate DoTR, 2 separate neuroradiologists with no knowledge of the surgical procedure performed were asked to classify an excision as complete resection when there was no residual tumour, maximum partial resection when less than 10% of the tumour remained and partial resection when less than 90% of the tumour had been resected. All tumours were examined by an experienced neuropathologist at the site and were classified according to the categories on brain tumours established by the World Health Organisation.22 The clinical variables analysed included form of presentation (seizure, intracranial hypertension, focal neurological signs) and overall status through the Karnofsky scale both preoperatively and postoperatively after 6 months. The intraoperative and delayed-onset complications that occurred within the observation period and could be attributed to the process analysed in this study were gathered prospectively. The times were recorded on the operating sheet. PT starts with the securing of the MRI-compatible Mayfield head clamp and obtaining the first intraoperative images before beginning the procedure. A 24-s e-steady sequence to locate the injury, and 7–11 min sequences potentiated in T1 with contrast for gadolinium-enhancing lesions, or T2 or FLAIR for all others, were performed, as reported in previous studies.23 ST starts
and the i7 integrated navigation
when PT concludes and ends when the procedure is finished. To evaluate postoperative morbidity, preoperative and postoperative signs and symptoms were recorded and follow-up was performed for at least 3 months after surgery. To evaluate survival, regular clinical follow-up was performed, which in the case of this study was at least 6 months from the surgical procedure. We studied the correlation between immediate postoperative imaging with low-field iMRI and early high-field MRI, and analysed sensitivity (SEN), specificity (SPE), positive predictive values and negative predictive values to detect tumour remnants. For the statistical analysis, categorical variables were reported using frequencies and percentages, and continuous variables were reported using mean and standard deviation or median and interquartile range (25th percentile-75th percentile) depending on their distribution. The general strategy for analysis was established as follows: to compare categorical variables, Fisher’s exact test was used between groups and McNemar’s test was used within groups; to compare continuous variables, Student’s t test was used between 2 groups (for dependent or independent data as applicable) and ANOVA was used for more than 2 groups. If applicability assumptions were not met, non-parametric methods were used: Mann–Whitney U tests (2 groups) or Kruskal–Wallis tests (more than 2 groups) for independent data, and Wilcoxon’s test (2 groups). Pearson’s correlation coefficient, or Spearman’s correlation coefficient if parametric applicability requirements were not met, was used to study the correlation between variables. The level of statistical significance was set at 5% on either side and statistical analysis was performed with the statistical package SPSS version 20.0 (SPSS Inc., Chicago, IL, United States).
Results Table 1 shows the baseline characteristics of the patients who underwent surgery. Of them, 37 patients (74%) underwent initial tumour surgery. The remaining 13 patients underwent surgery due to tumour recurrence after CR, and had previously undergone surgery in operating theatres without iMRI. Average ST was 196 min (median 180 ± 71 min) and average PT was 51 min (±19 m). Fig. 2 features a scatter plot which
ARTICLE IN PRESS 4
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
Degrees of tumour resection
Table 1 – Characteristics of the study sample. Patients (No.)
50
Male Female
56% 44%
100% 90%
20.0 28.6
80%
Age (years) KPS
55 (±13.7) PreOP PostOP 6 m
First procedure (No. %) Follow-up (m)
37; 74% 10.5 (±4.5)
Histological type (No. %)
HGGs LGGs Other
28; 56% 12; 24% 10; 20%
Placement Surgery
51 (±19) 196 (±71)
10.1 (±4.5)
HGGs LGGs Other
92.2 (±10.3) 80.9 (±22)
30.0 0.0
41.7
70% 60%
18.0 28.6
8.3
50% 40%
Times (min)
Hospital stay (d)
20% 10% 0% High-gr. g. (%)
9.8 (±5.2) 9.6 (±4.1) 11.6 (±5)
illustrates how PT evolved from the first case in October 2013 to the last case recorded, in October 2015, from an average of 1.5 h to an average of 0.5 h. Fig. 3 summarises DoTR. We found no statistically significant evidence that tumour type influences DoTR (p = 0.17). Partial resections were due to the eloquence of the operated area with language mapping (4 cases); neurophysiological recording which showed proximity to a motor area (6 cases), thalamic remnants (3 cases) confirmed through intraoperative neuroimaging or intraoperative risk criteria (extreme bradycardia, 2 cases). In the remaining cases the resection planned preoperatively was achieved. In 26 cases (52%), tumour remnants were observed after initial resection, and in 15 of these cases (30% of the total), this initial resection could be extended. There were false negatives, wherein iMRI showed CR and
1.5
Placement time (hours)
52.0
50.0 42.9
d: days; HGGs: high-grade gliomas; KPS: Karnofsky performance status; LGGs: low-grade gliomas; m: months; No.: number; OP: operative.
1
Low-gr. g. (%)
Other (%)
Total (%)
Histological diagnosis PR (n = 15)
MPR (n = 9)
CR (n = 26)
Fig. 3 – Degrees of tumour resection overall and according to tumour type. CR: complete resection; MPR: maximum partial resection; PR: partial resection.
high-field postoperative MRI showed tumour remnants, in 2 cases of partial resection and one case of maximum partial resection. These false negatives accounted for 6% of the total. None of the patients who had to undergo additional resection when tumour remnants were found on iMRI showed functional decline in the postoperative period. None of the patients died during the immediate postoperative period. During follow-up, overall mortality amounted to 8 cases (16%). The mortality rates in the follow-up period were 18% for a histological diagnosis of high-grade glioma (HGG), 17% for a histological diagnosis of low-grade glioma (LGG) and 10% for a different histological diagnosis. Fig. 4 shows the tables for Kaplan–Meier survival estimates at 25 months overall and by tumour type. The Karnofsky scale after 6 months of follow-up was 80.9 (±22) overall, 78 (±19) for HGGs, 77 (±35) for LGGs and 91 (±10) for other. No statistically significant evidence was found that tumour type influences overall survival. Table 1 shows average age. Table 2 shows the results for SEN, SPE, positive predictive values and negative predictive values.
Discussion
.5
0 01Oct2013
80.0
30%
01Apr2014
01Oct2014
01Apr2015
01Oct2015
Date of procedure
Fig. 2 – Scatter plot which illustrates the evolution of placement times. The arrow indicates the trend whereby placement time decreased as experience in use increased, from the first cases, in which placement time was 1.5 h, to the last cases, in which placement time was around 0.5 h.
With the series presented, we attempted to assess the use of iMRI technology in treating potentially resectable brain tumours. iMRI has demonstrated its usefulness as a technique to monitor degree of tumour resection and as a guide during surgery. In general, iMRI enabled maintenance of a high DoTR while minimising the rate of complications and gradually decreasing the ST. These results supported the belief that iMRI is a useful technological resource for brain tumour surgery. In 2008, Sanai and Berger,23 conscious of the lack of a general consensus on the role of DoTR in improving patient outcomes, reviewed all significant clinical publications since 1990 addressing this subject. Their most notable results showed that DoTR is among the most significant prognostic
ARTICLE IN PRESS 5
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
A
Kaplan–Meier survival estimate
1.00
B
Kaplan–Meier survival estimates
1.00
0.75
0.75 0.50 0.50 0.25
0.25
0.00
0.00 0
5
10
15
20
25
0
5
10
Months
15
20
25
Months High-gr. g.
Low-gr. g.
OTHER
Fig. 4 – Curves for Kaplan–Meier survival estimates, overall (A) and for each tumour type (B).
factors, as it has a substantial impact on disease-free survival and overall survival in both patients with LGGs and patients with HGGs. Moreover, brain shift is known to potentially decrease the reliability of neuronavigation and preoperative planning as well as negatively influence DoTR and thus overall survival and postoperative morbidity.24 iMRI allows images to be updated throughout the procedure by compensating for different degrees of brain shift and distortion which could not be corrected with other neuroimaging modalities.24,25 Thus, iMRI assistance allows more precise, less invasive approaches to be performed while simultaneously permitting maximum resection to be confirmed and intraoperative complications such as haemorrhages and infarcts to be detected. Since P. Black implemented the first iMRI in Boston in ® 1996 (Signa SP, General Electric Medical Systems), technological advancement has overcome initial difficulties: poor image definition, limited surgical access, instruments, size of equip® ® ment, etc. PoleStar (Medtronic ) iMRI, with a 0.12-T field (version N10) or a 0.15-T field (version N20/N30) was developed as a portable device with an image quality comparable to that of high-field systems which avoided patient transfers and changes in material, and reduced surgery time. Thus, some of the main disadvantages for widespread use of this technology were overcome. The scientific evidence of the benefit of image-guided surgery is controversial. Recently, Cochrane commissioned a literature review that compiled the most significant publications referring to the advantages, in terms of DoTR, of image-guided surgery versus conventional surgery.26 It also attempted to analyse whether any tool or technology was more effective. It concluded that the currently existing evidence was of low or very low quality with respect to the
notion that iMRI-guided surgery, 5-aminolevulinic acid or DTIneuronavigation increases the proportion of CR in patients with HGGs. Neither beneficial effects nor non-beneficial effects of image-guided surgery on survival and quality of life could be confirmed. However, the review made it possible to rule out, through the studies included, the prospect that iMRI use will cause greater postoperative sequelae than other surgical treatment modalities. Our study on the prospectively analysed series contributed data that supported the existing results in the literature. A review by Nimsky et al.27 on the usefulness of iMRI (0.2T) in glioma surgery found no complications associated with iMRI. Intraoperative imaging found residual tumour tissue in 63% of cases and allowed for extensive resection in 26% of all cases, especially in low-grade tumours. The researchers concluded that iMRI achieves a higher DoTR in the same surgical session. In their study on 103 patients with gliomas who underwent low-field iMRI-assisted surgery, Senft et al.28 confirmed that intraoperative imaging allowed residual tumour tissue to be identified in 51 individuals (49.5%) and allowed DoTR to be improved in 31 of them (30.1%). Our results are coherent with these studies since iMRI contributed to continuing resection in 30% of cases. In all other cases, resection could not be extended due to neurophysiological limitations or morphological eloquence. In line with the conclusions of Nimsky et al.27 and Herrera et al.,29 we believe that iMRI helps increase the extent of resection and monitor potential intraoperative events. Like them, we did not find that the rates of surgical morbidity and mortality were negatively affected by iMRI use. For Senft et al.28 , Nimsky et al.27 and Herrera et al.29 , more extensive resection did not correspond to a higher
Table 2 – Values for sensitivity (SEN), specificity (SPE), positive predictive value (PPV) and negative predictive value (NPV) of iMRI in our study. Values in parentheses indicate 95% confidence intervals. Histological diagnosis Overall HGGs LGGs Other
SEN 90.9% (76.4–96.9%) 93.8% (71.7–98.9%) 83.3% (43.6–97%) 50% (9.5–90.5%)
SPE 100% (81.6–100%) 100% (75.7–100%) 100% (61–100%) 100% (67.6–100%)
PPV 100% (88.6–100%) 100% (79.6–100%) 100% (56.6–100%) 100% (20.7–100%)
NPV 85% (64–94.8%) 92.3% (66.7–98.6%) 85.7% (48.7–97.4%) 88.9% (56.5–98%)
ARTICLE IN PRESS 6
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
rate of neurological sequelae. In addition, when the survival of patients with HGGs was analysed, it was found that the subjects who underwent complete tumour resection had a significantly higher survival rate than the patients with residual tumour (a 50% survival rate after 57.8 weeks versus 33.8 weeks, p = 0.003). In their study, age had no effect on survival (p = 0.12). Given these results, the authors of the study concluded that low-field iMRI was a useful device for brain tumour resection with a maximum safety level. Its use leads to a better prognosis for patients with malignant tumours. Our results showed a mortality rate of 16% overall, 18% for patients with a histological diagnosis of HGG, 17% for patients with a histological diagnosis of LGG and 10% for patients with a different histological diagnosis. Mortality in the LGG group was due to myocardial infarction during follow-up in one case and rapid disease progression in the other case. A non-significant higher mortality rate was observed in patients with a histological diagnosis of HGGs. Although the mortality rate in the LGG group was higher than expected compared to the HGG group for low-field iMRI procedures, our results yielded no statistically significant evidence on differences in overall survival depending on the tumour type. We believe that this may have been due to the small sample size enrolled. Years later, in several articles, Senft et al. published the results of a randomised clinical trial30–32 which compared results of low-field iMRI-assisted tumour resection to results of conventional surgery. An analysis was performed of 49 patients (24 patients randomly assigned to the iMRI group and 25 patients randomly assigned to conventional management). A higher percentage of tumour CR was obtained in the iMRI group (23 [96%] of 24 patients) than in the control group (17 [68%] of 25 patients) (p = 0.023). There were no significant differences in the rate of new neurological deficits in the intraoperative MRI group (3 [13%] of 24) versus the control group (2 [8%] of 25, p = 1). None of the patients with additional resection showed postoperative functional decline. Hirschl et al.33 retrospectively studied the degree of consis® tency between intraoperative imaging (PoleStar N-20 iMRI) and postoperative high-field MRI (1.5 T in the first 48 h). The SEN of iMRI for detecting tumour remnants was 0.74 (95% CI 0.58–0.86), and the SPE was 0.95 (95% CI 0.59–0.94). For tumours of glial origin, SEN increased to 0.82 (95% CI 0.59–0.94) and SPE was 0.95 (95% CI 0.73–1). With these data, together with the pre-existing literature, it can be concluded that iMRI is at least very helpful when identifying tumour remnants and making decisions during surgery. Our results are consistent with those of Hirschl et al., who found an overall SEN of 0.71 (95% CI 0.45–0.88) and a SPE of 0.94 (95% CI 0.83–0.98) for detecting tumour remnants. Our data for HGGs were slightly higher but still comparable. The increase in ST is among the top concerns in operating theatres with iMRI. Haydon et al.34 found that in adults with brain gliomas who undergo initial surgery, average ST in an iMRI suite did not differ from average ST in a standard operating theatre if the device was not used. However, average ST increased significantly when an intraoperative examination was performed and increased even more if the examination led to additional resection. The specific analysis of times showed that efficiency increased and times decreased as experience using the device was accumulated. This demonstrated
that, like other newly added instruments, the use of iMRI in the neurosurgery operating theatre has its own learning curve. An increase in ST is inevitable in cases of iMRI-assisted procedures. Our results were consistent with that observed in the literature and demonstrated this increase in ST with iMRI use and a gradual trend towards a decrease in ST as cumulative experience increased. Obviously, the training of the nursing and auxiliary staff, the experience of the neurosurgeon and neuroanaesthesiologist, the specialisation of anaesthesia and the proper selection of imaging sequences contributed to the improvement in times compared to the initial cases. One of the most controversial factors analysed in evaluations of intraoperative image-guided surgery is related to cost analysis and its relationship to clinical variables. This technology is still very expensive. Given the current economic, social and healthcare paradigm in Spain, this matter is of capital importance. Makary et al.35 analysed the cost–effectiveness of implementing low-field iMRI (0.15-T) in neuro-oncology. They compared it to both conventional surgery and high-field iMRI (1.5-T). As a general conclusion, they did not find in low-field iMRI the benefit reported in high-field iMRI. They warned that there is no justification for widespread implementation of low-field iMRI in its current state of development. This has not been duly studied or demonstrated overall in the studies included. With our experience accumulated in this period, we were unable to settle this matter one way or another. Even so, we believe that the potential benefit of such technology would be clearer if the disease were centralised and those cases likely to benefit from advanced techniques were referred to centres that boast the latest diagnostic and therapeutic advances. The limitations of the study included the limited sample size, the heterogeneity of the sample and the follow-up time, which was particularly inadequate for low-grade variants. However, it does offer a prospective series of a cohort of brain tumours, taking into account preoperative, intraoperative and postoperative radiological, pathological and clinical variables.
Conclusions The use of iMRI allows for a high DoTR compared to the use of conventional techniques previously found in the literature. In addition, thanks to intraoperative monitoring and compensation for brain shift, we observed that the iMRI use reduces the risk of postoperative morbidity in brain glioma surgery. Although the results obtained were not statistically significant, and the follow-up period was not sufficient to obtain solid data in this regard, it can be intuited that iMRI use could have positive repercussions for the prognosis of patients in both the short term and the long term in terms of both overall survival and progression-free survival. Although the quality of images obtained with low-field iMRI is lower than the quality of images obtained with highfield iMRI, we obtained some high-quality images showing that iMRI has high SEN and SPE when detecting tumour remnants for resection. ST was markedly longer when iMRI was used compared to conventional procedures. iMRI ST and PT, as well as workflow, improved as experience using the technology increased,
ARTICLE IN PRESS n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
and were markedly lower in more recent surgical procedures. There was a learning curve comparable to that of any newly added technology. There seemed to be a non-significantly shorter hospital stay in patients who underwent iMRI-assisted surgery. This would correspond to cost–effectiveness benefits in using this device. We found no complications in the postoperative period associated with its use. This translates to a better cost/benefit ratio and a reduction in hospital charges and costs overall. Longer-term studies enrolling a higher number of patients are required to reliably demonstrate the extent of the positive repercussions of this technology for the different types of brain glioma. The efforts to come should be aimed at improving implementation of the technology, improving indications, decreasing ST, minimising resulting complications and lowering direct and indirect costs, thereby helping improve the overall efficiency and effectiveness of the procedure.
Conflicts of interest The authors declare that they have no conflicts of interest.
references
1. Hill DL, Maurer CR Jr, Maciunas RJ, Barwise JA, Fitzpatrick MJ, Wang MY. Measurement of intra-operative brain surface deformation under a craniotomy. Neurosurgery. 1998;43:514–26. 2. Maurer CR Jr, Hill DL, Martin AJ, Liu H, McCue M, Rueckert D, et al. Investigation of intraoperative brain deformation using a 1.5-T interventional MR system: Preliminary results. IEEE Trans Med Imaging. 1998;17:817–25. 3. Hata N, Nabavi A, Wells WM 3rd. Three-dimensional optical flow method for measurement of volumetric brain deformation from intra-operative MR images. J Comput Assist Tomogr. 2000;24:531–8. 4. Nabavi A, Black PM, Gering DT. Serial intra-operative magnetic resonance imaging of brain shift. Neurosurgery. 2001;48:787–97. 5. Nimsky C, Ganslandt O, Hastreiter P, Fahlbusch R. Intraoperative compensation for brain shift. Surg Neurol. 2001;56:357–64. 6. Reinges MHT, Nguyen HH, Krings T, Hutter BO, Rohde V, Gilsbach JM. Course of brain shift during microsurgical resection of supratentorial cerebral lesions: Limits of conventional neuronavigation. Acta Neurochir (Wien). 2004;146:369. 7. Roldan P. Experiencia con el uso de la resonancia magnética Intraoperatoria de bajo campo en Neurocirugía para el tratamiento de neoplasias cerebrales [tesis]. Valencia: Facultad de Medicina. Universitat de Valéncia; 2016. 8. Unsgaard G, Rygh OM, Selbekk T, Müller TB, Kolstad F, Lindseth F, et al. Intraoperative 3D ultrasound in neurosurgery. Acta Neurochir (Wien). 2006;148:235–53. 9. Black PM, Moriarty T, Alexander E 3rd. Development and implementation of intraoperative magnetic resonance imaging and its neurosurgical applications. Neurosurgery. 1997;41:831–45. 10. Gluch L, Walker DG. Intraoperative magnetic resonance: The future of surgery. ANZ J Surg. 2002;72:426–36.
7
11. Lewin JS, Metzger A, Selman WR. Intraoperative magnetic resonance image guidance in neurosurgery. J Magn Reson Imaging. 2000;12:512–24. 12. Jolesz FA. Interventional and intraoperative MRI: A general overview of the field. J Magn Reson Imaging. 1998;8:3–7. 13. Tummala RP, Chu RM, Liu H, Truwit CL, Hall WA. Optimizing brain tumor resection. High-field interventional MR imaging. Neuroimaging Clin N Am. 2001;11:673–83. 14. Berkenstadt H, Perel A, Ram Z, Feldman Z, Nahtomi-Shick O, Hadani M. Anesthesia for magnetic resonance guided neurosurgery: initial experience with a new open magnetic resonance imaging system. J Neurosurg Anesthesiol. 2001;13:158–62. 15. Hall W, Truwit C. Intraoperative MR-guided neurosurgery. J Magn Reson Imag. 2008;27:368–75. 16. Lipson AC, Gargollo PC, Black PM. Intraoperative magnetic resonance imaging: considerations for the operating room of the future. J Clin Neurosci. 2001;8:305–10. 17. Lacroix M, Abi-Said D, Fourney DR, Gokaslan ZL, Shi W, DeMonte F, et al. A multivariate analysis of 416 patients with glioblastoma multiforme: prognosis, extent of resection, and survival. J Neurosurg. 2001;95:190–8. 18. Stummer W, Reulen HJ, Meinel T. Extent of resection and survival in glioblastoma multiforme: identification of and adjustment for bias. Neurosurgery. 2008;62:564–76. 19. Chaichana KL, Halthore AN, Parker SL. Factors involved in maintaining prolonged functional independence following supratentorial glioblastoma resection. J Neurosurg. 2011;114:604–12. 20. Brown PD, Maurer MJ, Rummans TA. A prospective study of quality of life in adults with newly diagnosed high-grade gliomas: The impact of the extent of resection on quality of life and survival. Neurosurgery. 2005;57:495–504. ˜ 21. Brell M, Roldán P, González E, Llinàs P, Ibánez J. Implantación de la primera resonancia magnética intraoperatoria en un ˜ hospital de la red sanitaria pública espanola: experiencia inicial, viabilidad y dificultades en nuestro entorno. Neurocirugía. 2013;24:11–21. 22. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A, et al. The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol. 2007;114: 97–109. 23. Sanai N, Berger MS. Glioma extent of resection and its impact on patient outcome. Neurosurgery. 2008;62:753–64. 24. Dorward NL, Alberti O, Velani B. Postimaging brain distortion: magnitude, correlates, and impact on neuronavigation. J Neurosurg. 1998;88:656–62. 25. Roberts DW, Hartov A, Kennedy FE. Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery. 1998;43:749–60. 26. Barone D, Lawrie TA, Hart MG. Image guided surgery for the resection of brain tumours. Cochrane Database Syst Rev. 2014. 27. Nimsky C, Ganslandt O, Buvhfelder M, Fahlbusch R. Glioma surgery evaluated by intraoperative low-field magnetic resonance imaging. Acta Neurochir Suppl (Wien). 2003;85:55–63. 28. Senft C, Franz K, Ulrich CT, Bink A, Szelényi A, Gasser T, et al. Low field intraoperative MRI-guided surgery of gliomas: a single center experience. Clin Neurol Neurosurg. 2010;112:237–43. 29. Herrera R, Ledesma JL, Pomata H, Lambre J, Rojas H, Houssay A, et al. El desarrollo e implementación de la resonancia magnética intraoperatoria en neurocirugía (REMAIN) y su aplicación en 83 intervenciones en la Argentina. Rev Argent Neurocir. 2007:21. 30. Senft C, Bink A, Franz K, Vatter H, Gasser T, Seifert V. Intraoperative MRI guidance and extent of resection in glioma
ARTICLE IN PRESS 8
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
surgery: a randomised, controlled trial. Lancet Oncol. 2011;12:997–1003. 31. Senft C, Bink A, Franz K, Gasser T, Seifert V. Intraoperative MRI-Guided vs. conventional microsurgical brain tumor resection - results of a prospective randomized trial. J Neurosurg. 2011;115:602. 32. Senft C, Bink A, Heckelmann M, Gasser T, Seifert V. Glioma extent of resection and ultra-low-field iMRI: Interim analysis of a prospective randomized controlled trial. Acta Neurochir Suppl. 2011;109:49–53. 33. Hirschl RA, Wilson J, Miller B, Bergese S, Chiocca EA. The predictive value of low-field strength magnetic resonance
imaging for intraoperative residual tumor detection. J Neurosurg. 2009;111:252–7. 34. Haydon DH, Chicoine MR, Dacey RG. The impact of high-field-strength intraoperative magnetic resonance imaging on brain tumor management. Neurosurgery. 2013;60:92–7. 35. Makary M, Chiocca EA, Erminy N, Antor M, Bergese SD, Abdel-Rasoul M, et al. Clinical and economic outcomes of low-field intraoperative MRI-guided tumor resection neurosurgery. J Magn Reson Imaging. 2011;34:1022–30.
NEUROCIRUGÍA www.elsevier.es/neurocirugia
Clinical Research
Low field intra-operative magnetic resonance imaging for brain tumour surgery: Preliminary experience夽 Pedro Roldán a,∗ , Sergio García a , Josep González a , Luis Alberto Reyes a , Jorge Torales a , ˜ a Ricard Valero b , Laura Oleaga c , Joaquim Ensenat a b c
Servicio de Neurocirugía, Hospital Clínic de Barcelona, Barcelona, Spain Servicio de Anestesiología y Reanimación, Hospital Clínic de Barcelona, Barcelona, Spain Servicio de Radiodiagnóstico, Hospital Clínic de Barcelona, Barcelona, Spain
a r t i c l e
i n f o
a b s t r a c t
Article history:
Objectives: Intra-operative magnetic resonance imaging (iMRI) is a recently introduced tool
Received 12 June 2016
in the most advanced neurosurgical operating rooms worldwide. We present our preliminary ®
Accepted 10 August 2016
experience in brain tumour surgery with low field PoleStar N30
Available online xxx
its introduction in 2013 in the Barcelona Clinic Hospital.
Keywords:
on using iMRI and intention of complete removal up to October 2015. A record was made of
intraoperative MRI since
Material and methods: A prospective non-randomised study was conducted on cases operated Glioma
the data as regards surgical times, resection rates, histological diagnosis, hospital stay, and
Craniotomy
survival rates during follow-up.
Intra-operative magnetic resonance
Results: The study included 50 patients, with a mean age of 55 years (±13.7), a preoperative
imaging
mean Karnofsky of 92 (being 81 post-operatively), and a mean follow-up of 10.5 months
Neuronavigation
(±6.5). There were 26% re-operations due to recurrence. High-grade gliomas were reported in
Intra-operative neuroimaging
56%, low-grade gliomas in 24%, and 20% “Other” tumours. Overall hospital stay was 10 days
Neurosurgery
(±4.5). Depending on the histologiacl diagnosis, the “Others” group had a longer hospital stay. Overall, there were 52% complete removal, 18% of maximum removals, and 30% of partial removals. The overall survival rates during follow-up was 84%. Conclusions: iMRI is a safe and effective tool for brain tumour surgery. Its use allows an increase in resection rates, and minimises post-operative complications. Its implementation involves an increase in surgical time, which improves with the characteristic learning curve. More studies are needed to establish its role in the long-term survival of patients. ˜ ˜ S.L.U. All rights © 2016 Sociedad Espanola de Neurocirug´ıa. Published by Elsevier Espana, reserved.
夽 Please cite this article as: Roldán P, García S, González J, Reyes LA, Torales J, Valero R, et al. Resonancia magnética intraoperatoria de bajo campo para la cirugía de neoplasias cerebrales: experiencia preliminar. Neurocirugía. 2017. http://dx.doi.org/10.1016/j.neucir.2016.08.001 ∗ Corresponding author. E-mail address: [email protected] (P. Roldán). ˜ ˜ S.L.U. All rights reserved. 2529-8496/© 2016 Sociedad Espanola de Neurocirug´ıa. Published by Elsevier Espana,
NEUCIE-254; No. of Pages 8
ARTICLE IN PRESS 2
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
Resonancia magnética intraoperatoria de bajo campo para la cirugía de neoplasias cerebrales: experiencia preliminar r e s u m e n Palabras clave:
Objetivos: La resonancia magnética intraoperatoria (RMi) es una herramienta recientemente
Glioma
introducida en los quirófanos de neurocirugía más vanguardistas mundialmente. Presenta-
Craneotomía Resonancia magnética
mos nuestra experiencia clínica preliminar con relación al empleo de la RMi de bajo campo, PoleStar N30® , desde su implementación en 2013 en el Hospital Clínic de Barcelona, para
intraoperatoria
el tratamiento de neoplasias cerebrales.
Neuronavegación
Material y método: Se realizó un estudio prospectivo no aleatorizado incluyendo los casos
Neuroimagen intraoperatoria
intervenidos mediante RMi con intención de resección completa hasta octubre de 2015. Se
Neurocirugía
registraron los tiempos quirúrgicos así como los grados de resección, diagnóstico histológico, estancia hospitalaria y la supervivencia durante el seguimiento. ˜ Resultados: Se incluyeron 50 pacientes con edad media de 55 anos (±13,7), un Karnofsky preoperatorio de 92 (siendo el postoperatorio de 81); y un seguimiento medio de 10,5 meses (±6,5). Un 26% fueron reintervenciones por recidiva. Un 56% eran gliomas de alto grado, un 24% gliomas de bajo grado y un 20% otras neoplasias. La estancia hospitalaria global fue de 10 días (±4,5). Según el diagnóstico histológico el grupo «otras» fue el que mayor estancia hospitalaria presentaba. Globalmente, se lograron un 52% de resección completa, un 18% de resecciones parciales máximas y un 30% de resecciones parciales. La supervivencia durante el seguimiento fue del 84%. Conclusiones: La RMi es una herramienta segura y eficaz en la cirugía de neoplasias cerebrales. Su uso permite aumentar el grado de resección disminuyendo las complicaciones posquirúrgicas. Su empleo conlleva una prolongación del tiempo quirúrgico que mejora con la curva de aprendizaje característica. Más estudios son necesarios para poder establecer su papel en la supervivencia a largo plazo de los pacientes. ˜ ˜ S.L.U. Todos © 2016 Sociedad Espanola de Neurocirug´ıa. Publicado por Elsevier Espana, los derechos reservados.
Introduction The development of imaging techniques has had an unquestionable impact on the development of neurosurgery. Computed tomography, introduced in the 1970s, and magnetic resonance imaging (MRI), introduced in the 1990s, have been implemented in neuronavigation systems. Thanks to surgical planning, simulation and navigation, more complex procedures can be performed with ever-fewer invasive techniques. At present, the most widespread navigation systems are socalled frameless systems. Navigation can be affected by events occurring during initial recording, or by changes in intracranial tissues and fluids during surgery. Brain shift may render recorded preoperative data inaccurate. It is estimated that shifts of up to 1 cm per 1 h occur after opening the dura mater in more than half of patients studied. This shift is greater in cases of tumour volume resection or hydrocephaly, thereby drastically decreasing the usefulness of navigation during the procedure.1–6 Intraoperative imaging was developed to overcome these disadvantages. Current intraoperative imaging modalities include intraoperative computed tomography,7 intraoperative ultrasounds8 and intraoperative MRI (iMRI). The problems encountered in bringing MRI into the operating theatre included electromagnetic interference as well as device size, mobility and magnetic isolation. In 1997, Black et al.9 first
used a novel iMRI system allowing high-quality images to be acquired and used as a navigation system during neurosurgical procedures.10–13 The potential benefits of this technology would allow intraoperative confirmation of degree of tumour resection (DoTR). This would contribute to an increase in complete resection (CR) and a decrease in postoperative complications, morbidity and length of hospital stay, which would in turn contribute to an increase in quality of life and overall survival.14–20 Since this first publication in 1997,11 the development of iMRI has gradually increased in significance and sites worldwide have equipped their surgeons with this very useful yet very expensive technology. There are 3 major types of iMRI depending on magnetic field intensity–high-field (1.5–3 teslas [T]), medium-field (0.5–1.5 T) and low-field (0.12–0.5 T)–with advantages and disadvantages inherent to the performance of each. In this study we present preliminary experience with the use of low-field iMRI, PoleStar N30 Medtronic® , to assist neuro-oncology surgery at Hospital Clínic Universitari in Barcelona, since it was implemented in 2013 (Fig. 1).
Material and methods All patients who underwent surgery for brain tumours with the intention of CR and iMRI assistance were analysed
ARTICLE IN PRESS 3
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
®
Fig. 1 – Panoramic view of the “smart” operating theatre with PoleStar N30 low-field iMRI system at Hospital Clínic in Barcelona.
prospectively. The sample was recruited from October 2013, when the system was implemented at our site, to October 2015. All patients signed the informed consent form authorising their enrolment in the framework intraoperative imaging study, “SMART OR”, conducted at the site. The same inclusion and exclusion criteria as those established for conventional neuro-oncology surgery were used and the standard safety regulations for patients examined with MRI were applied. iMRI was performed in all cases of intra-axial brain lesions with the intention of CR where MRI use was not contraindicated and was permitted by patient physiognomy. The neurosurgeon selected the imaging modality on a case-by-case basis taking into account the best technique in each case. The surgical technique for neuro-oncology surgery assisted by low-field iMRI has already been previously reported by other authors7,21 and is not discussed in this study. We evaluated parameters such as surgery time (ST), placement time (PT), DoTR, postoperative morbidity, length of hospital stay and survival. To measure DoTR, preoperative images (3 T) were compared to early postoperative images (up to 72 h after surgery). To evaluate DoTR, 2 separate neuroradiologists with no knowledge of the surgical procedure performed were asked to classify an excision as complete resection when there was no residual tumour, maximum partial resection when less than 10% of the tumour remained and partial resection when less than 90% of the tumour had been resected. All tumours were examined by an experienced neuropathologist at the site and were classified according to the categories on brain tumours established by the World Health Organisation.22 The clinical variables analysed included form of presentation (seizure, intracranial hypertension, focal neurological signs) and overall status through the Karnofsky scale both preoperatively and postoperatively after 6 months. The intraoperative and delayed-onset complications that occurred within the observation period and could be attributed to the process analysed in this study were gathered prospectively. The times were recorded on the operating sheet. PT starts with the securing of the MRI-compatible Mayfield head clamp and obtaining the first intraoperative images before beginning the procedure. A 24-s e-steady sequence to locate the injury, and 7–11 min sequences potentiated in T1 with contrast for gadolinium-enhancing lesions, or T2 or FLAIR for all others, were performed, as reported in previous studies.23 ST starts
and the i7 integrated navigation
when PT concludes and ends when the procedure is finished. To evaluate postoperative morbidity, preoperative and postoperative signs and symptoms were recorded and follow-up was performed for at least 3 months after surgery. To evaluate survival, regular clinical follow-up was performed, which in the case of this study was at least 6 months from the surgical procedure. We studied the correlation between immediate postoperative imaging with low-field iMRI and early high-field MRI, and analysed sensitivity (SEN), specificity (SPE), positive predictive values and negative predictive values to detect tumour remnants. For the statistical analysis, categorical variables were reported using frequencies and percentages, and continuous variables were reported using mean and standard deviation or median and interquartile range (25th percentile-75th percentile) depending on their distribution. The general strategy for analysis was established as follows: to compare categorical variables, Fisher’s exact test was used between groups and McNemar’s test was used within groups; to compare continuous variables, Student’s t test was used between 2 groups (for dependent or independent data as applicable) and ANOVA was used for more than 2 groups. If applicability assumptions were not met, non-parametric methods were used: Mann–Whitney U tests (2 groups) or Kruskal–Wallis tests (more than 2 groups) for independent data, and Wilcoxon’s test (2 groups). Pearson’s correlation coefficient, or Spearman’s correlation coefficient if parametric applicability requirements were not met, was used to study the correlation between variables. The level of statistical significance was set at 5% on either side and statistical analysis was performed with the statistical package SPSS version 20.0 (SPSS Inc., Chicago, IL, United States).
Results Table 1 shows the baseline characteristics of the patients who underwent surgery. Of them, 37 patients (74%) underwent initial tumour surgery. The remaining 13 patients underwent surgery due to tumour recurrence after CR, and had previously undergone surgery in operating theatres without iMRI. Average ST was 196 min (median 180 ± 71 min) and average PT was 51 min (±19 m). Fig. 2 features a scatter plot which
ARTICLE IN PRESS 4
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
Degrees of tumour resection
Table 1 – Characteristics of the study sample. Patients (No.)
50
Male Female
56% 44%
100% 90%
20.0 28.6
80%
Age (years) KPS
55 (±13.7) PreOP PostOP 6 m
First procedure (No. %) Follow-up (m)
37; 74% 10.5 (±4.5)
Histological type (No. %)
HGGs LGGs Other
28; 56% 12; 24% 10; 20%
Placement Surgery
51 (±19) 196 (±71)
10.1 (±4.5)
HGGs LGGs Other
92.2 (±10.3) 80.9 (±22)
30.0 0.0
41.7
70% 60%
18.0 28.6
8.3
50% 40%
Times (min)
Hospital stay (d)
20% 10% 0% High-gr. g. (%)
9.8 (±5.2) 9.6 (±4.1) 11.6 (±5)
illustrates how PT evolved from the first case in October 2013 to the last case recorded, in October 2015, from an average of 1.5 h to an average of 0.5 h. Fig. 3 summarises DoTR. We found no statistically significant evidence that tumour type influences DoTR (p = 0.17). Partial resections were due to the eloquence of the operated area with language mapping (4 cases); neurophysiological recording which showed proximity to a motor area (6 cases), thalamic remnants (3 cases) confirmed through intraoperative neuroimaging or intraoperative risk criteria (extreme bradycardia, 2 cases). In the remaining cases the resection planned preoperatively was achieved. In 26 cases (52%), tumour remnants were observed after initial resection, and in 15 of these cases (30% of the total), this initial resection could be extended. There were false negatives, wherein iMRI showed CR and
1.5
Placement time (hours)
52.0
50.0 42.9
d: days; HGGs: high-grade gliomas; KPS: Karnofsky performance status; LGGs: low-grade gliomas; m: months; No.: number; OP: operative.
1
Low-gr. g. (%)
Other (%)
Total (%)
Histological diagnosis PR (n = 15)
MPR (n = 9)
CR (n = 26)
Fig. 3 – Degrees of tumour resection overall and according to tumour type. CR: complete resection; MPR: maximum partial resection; PR: partial resection.
high-field postoperative MRI showed tumour remnants, in 2 cases of partial resection and one case of maximum partial resection. These false negatives accounted for 6% of the total. None of the patients who had to undergo additional resection when tumour remnants were found on iMRI showed functional decline in the postoperative period. None of the patients died during the immediate postoperative period. During follow-up, overall mortality amounted to 8 cases (16%). The mortality rates in the follow-up period were 18% for a histological diagnosis of high-grade glioma (HGG), 17% for a histological diagnosis of low-grade glioma (LGG) and 10% for a different histological diagnosis. Fig. 4 shows the tables for Kaplan–Meier survival estimates at 25 months overall and by tumour type. The Karnofsky scale after 6 months of follow-up was 80.9 (±22) overall, 78 (±19) for HGGs, 77 (±35) for LGGs and 91 (±10) for other. No statistically significant evidence was found that tumour type influences overall survival. Table 1 shows average age. Table 2 shows the results for SEN, SPE, positive predictive values and negative predictive values.
Discussion
.5
0 01Oct2013
80.0
30%
01Apr2014
01Oct2014
01Apr2015
01Oct2015
Date of procedure
Fig. 2 – Scatter plot which illustrates the evolution of placement times. The arrow indicates the trend whereby placement time decreased as experience in use increased, from the first cases, in which placement time was 1.5 h, to the last cases, in which placement time was around 0.5 h.
With the series presented, we attempted to assess the use of iMRI technology in treating potentially resectable brain tumours. iMRI has demonstrated its usefulness as a technique to monitor degree of tumour resection and as a guide during surgery. In general, iMRI enabled maintenance of a high DoTR while minimising the rate of complications and gradually decreasing the ST. These results supported the belief that iMRI is a useful technological resource for brain tumour surgery. In 2008, Sanai and Berger,23 conscious of the lack of a general consensus on the role of DoTR in improving patient outcomes, reviewed all significant clinical publications since 1990 addressing this subject. Their most notable results showed that DoTR is among the most significant prognostic
ARTICLE IN PRESS 5
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
A
Kaplan–Meier survival estimate
1.00
B
Kaplan–Meier survival estimates
1.00
0.75
0.75 0.50 0.50 0.25
0.25
0.00
0.00 0
5
10
15
20
25
0
5
10
Months
15
20
25
Months High-gr. g.
Low-gr. g.
OTHER
Fig. 4 – Curves for Kaplan–Meier survival estimates, overall (A) and for each tumour type (B).
factors, as it has a substantial impact on disease-free survival and overall survival in both patients with LGGs and patients with HGGs. Moreover, brain shift is known to potentially decrease the reliability of neuronavigation and preoperative planning as well as negatively influence DoTR and thus overall survival and postoperative morbidity.24 iMRI allows images to be updated throughout the procedure by compensating for different degrees of brain shift and distortion which could not be corrected with other neuroimaging modalities.24,25 Thus, iMRI assistance allows more precise, less invasive approaches to be performed while simultaneously permitting maximum resection to be confirmed and intraoperative complications such as haemorrhages and infarcts to be detected. Since P. Black implemented the first iMRI in Boston in ® 1996 (Signa SP, General Electric Medical Systems), technological advancement has overcome initial difficulties: poor image definition, limited surgical access, instruments, size of equip® ® ment, etc. PoleStar (Medtronic ) iMRI, with a 0.12-T field (version N10) or a 0.15-T field (version N20/N30) was developed as a portable device with an image quality comparable to that of high-field systems which avoided patient transfers and changes in material, and reduced surgery time. Thus, some of the main disadvantages for widespread use of this technology were overcome. The scientific evidence of the benefit of image-guided surgery is controversial. Recently, Cochrane commissioned a literature review that compiled the most significant publications referring to the advantages, in terms of DoTR, of image-guided surgery versus conventional surgery.26 It also attempted to analyse whether any tool or technology was more effective. It concluded that the currently existing evidence was of low or very low quality with respect to the
notion that iMRI-guided surgery, 5-aminolevulinic acid or DTIneuronavigation increases the proportion of CR in patients with HGGs. Neither beneficial effects nor non-beneficial effects of image-guided surgery on survival and quality of life could be confirmed. However, the review made it possible to rule out, through the studies included, the prospect that iMRI use will cause greater postoperative sequelae than other surgical treatment modalities. Our study on the prospectively analysed series contributed data that supported the existing results in the literature. A review by Nimsky et al.27 on the usefulness of iMRI (0.2T) in glioma surgery found no complications associated with iMRI. Intraoperative imaging found residual tumour tissue in 63% of cases and allowed for extensive resection in 26% of all cases, especially in low-grade tumours. The researchers concluded that iMRI achieves a higher DoTR in the same surgical session. In their study on 103 patients with gliomas who underwent low-field iMRI-assisted surgery, Senft et al.28 confirmed that intraoperative imaging allowed residual tumour tissue to be identified in 51 individuals (49.5%) and allowed DoTR to be improved in 31 of them (30.1%). Our results are coherent with these studies since iMRI contributed to continuing resection in 30% of cases. In all other cases, resection could not be extended due to neurophysiological limitations or morphological eloquence. In line with the conclusions of Nimsky et al.27 and Herrera et al.,29 we believe that iMRI helps increase the extent of resection and monitor potential intraoperative events. Like them, we did not find that the rates of surgical morbidity and mortality were negatively affected by iMRI use. For Senft et al.28 , Nimsky et al.27 and Herrera et al.29 , more extensive resection did not correspond to a higher
Table 2 – Values for sensitivity (SEN), specificity (SPE), positive predictive value (PPV) and negative predictive value (NPV) of iMRI in our study. Values in parentheses indicate 95% confidence intervals. Histological diagnosis Overall HGGs LGGs Other
SEN 90.9% (76.4–96.9%) 93.8% (71.7–98.9%) 83.3% (43.6–97%) 50% (9.5–90.5%)
SPE 100% (81.6–100%) 100% (75.7–100%) 100% (61–100%) 100% (67.6–100%)
PPV 100% (88.6–100%) 100% (79.6–100%) 100% (56.6–100%) 100% (20.7–100%)
NPV 85% (64–94.8%) 92.3% (66.7–98.6%) 85.7% (48.7–97.4%) 88.9% (56.5–98%)
ARTICLE IN PRESS 6
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
rate of neurological sequelae. In addition, when the survival of patients with HGGs was analysed, it was found that the subjects who underwent complete tumour resection had a significantly higher survival rate than the patients with residual tumour (a 50% survival rate after 57.8 weeks versus 33.8 weeks, p = 0.003). In their study, age had no effect on survival (p = 0.12). Given these results, the authors of the study concluded that low-field iMRI was a useful device for brain tumour resection with a maximum safety level. Its use leads to a better prognosis for patients with malignant tumours. Our results showed a mortality rate of 16% overall, 18% for patients with a histological diagnosis of HGG, 17% for patients with a histological diagnosis of LGG and 10% for patients with a different histological diagnosis. Mortality in the LGG group was due to myocardial infarction during follow-up in one case and rapid disease progression in the other case. A non-significant higher mortality rate was observed in patients with a histological diagnosis of HGGs. Although the mortality rate in the LGG group was higher than expected compared to the HGG group for low-field iMRI procedures, our results yielded no statistically significant evidence on differences in overall survival depending on the tumour type. We believe that this may have been due to the small sample size enrolled. Years later, in several articles, Senft et al. published the results of a randomised clinical trial30–32 which compared results of low-field iMRI-assisted tumour resection to results of conventional surgery. An analysis was performed of 49 patients (24 patients randomly assigned to the iMRI group and 25 patients randomly assigned to conventional management). A higher percentage of tumour CR was obtained in the iMRI group (23 [96%] of 24 patients) than in the control group (17 [68%] of 25 patients) (p = 0.023). There were no significant differences in the rate of new neurological deficits in the intraoperative MRI group (3 [13%] of 24) versus the control group (2 [8%] of 25, p = 1). None of the patients with additional resection showed postoperative functional decline. Hirschl et al.33 retrospectively studied the degree of consis® tency between intraoperative imaging (PoleStar N-20 iMRI) and postoperative high-field MRI (1.5 T in the first 48 h). The SEN of iMRI for detecting tumour remnants was 0.74 (95% CI 0.58–0.86), and the SPE was 0.95 (95% CI 0.59–0.94). For tumours of glial origin, SEN increased to 0.82 (95% CI 0.59–0.94) and SPE was 0.95 (95% CI 0.73–1). With these data, together with the pre-existing literature, it can be concluded that iMRI is at least very helpful when identifying tumour remnants and making decisions during surgery. Our results are consistent with those of Hirschl et al., who found an overall SEN of 0.71 (95% CI 0.45–0.88) and a SPE of 0.94 (95% CI 0.83–0.98) for detecting tumour remnants. Our data for HGGs were slightly higher but still comparable. The increase in ST is among the top concerns in operating theatres with iMRI. Haydon et al.34 found that in adults with brain gliomas who undergo initial surgery, average ST in an iMRI suite did not differ from average ST in a standard operating theatre if the device was not used. However, average ST increased significantly when an intraoperative examination was performed and increased even more if the examination led to additional resection. The specific analysis of times showed that efficiency increased and times decreased as experience using the device was accumulated. This demonstrated
that, like other newly added instruments, the use of iMRI in the neurosurgery operating theatre has its own learning curve. An increase in ST is inevitable in cases of iMRI-assisted procedures. Our results were consistent with that observed in the literature and demonstrated this increase in ST with iMRI use and a gradual trend towards a decrease in ST as cumulative experience increased. Obviously, the training of the nursing and auxiliary staff, the experience of the neurosurgeon and neuroanaesthesiologist, the specialisation of anaesthesia and the proper selection of imaging sequences contributed to the improvement in times compared to the initial cases. One of the most controversial factors analysed in evaluations of intraoperative image-guided surgery is related to cost analysis and its relationship to clinical variables. This technology is still very expensive. Given the current economic, social and healthcare paradigm in Spain, this matter is of capital importance. Makary et al.35 analysed the cost–effectiveness of implementing low-field iMRI (0.15-T) in neuro-oncology. They compared it to both conventional surgery and high-field iMRI (1.5-T). As a general conclusion, they did not find in low-field iMRI the benefit reported in high-field iMRI. They warned that there is no justification for widespread implementation of low-field iMRI in its current state of development. This has not been duly studied or demonstrated overall in the studies included. With our experience accumulated in this period, we were unable to settle this matter one way or another. Even so, we believe that the potential benefit of such technology would be clearer if the disease were centralised and those cases likely to benefit from advanced techniques were referred to centres that boast the latest diagnostic and therapeutic advances. The limitations of the study included the limited sample size, the heterogeneity of the sample and the follow-up time, which was particularly inadequate for low-grade variants. However, it does offer a prospective series of a cohort of brain tumours, taking into account preoperative, intraoperative and postoperative radiological, pathological and clinical variables.
Conclusions The use of iMRI allows for a high DoTR compared to the use of conventional techniques previously found in the literature. In addition, thanks to intraoperative monitoring and compensation for brain shift, we observed that the iMRI use reduces the risk of postoperative morbidity in brain glioma surgery. Although the results obtained were not statistically significant, and the follow-up period was not sufficient to obtain solid data in this regard, it can be intuited that iMRI use could have positive repercussions for the prognosis of patients in both the short term and the long term in terms of both overall survival and progression-free survival. Although the quality of images obtained with low-field iMRI is lower than the quality of images obtained with highfield iMRI, we obtained some high-quality images showing that iMRI has high SEN and SPE when detecting tumour remnants for resection. ST was markedly longer when iMRI was used compared to conventional procedures. iMRI ST and PT, as well as workflow, improved as experience using the technology increased,
ARTICLE IN PRESS n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
and were markedly lower in more recent surgical procedures. There was a learning curve comparable to that of any newly added technology. There seemed to be a non-significantly shorter hospital stay in patients who underwent iMRI-assisted surgery. This would correspond to cost–effectiveness benefits in using this device. We found no complications in the postoperative period associated with its use. This translates to a better cost/benefit ratio and a reduction in hospital charges and costs overall. Longer-term studies enrolling a higher number of patients are required to reliably demonstrate the extent of the positive repercussions of this technology for the different types of brain glioma. The efforts to come should be aimed at improving implementation of the technology, improving indications, decreasing ST, minimising resulting complications and lowering direct and indirect costs, thereby helping improve the overall efficiency and effectiveness of the procedure.
Conflicts of interest The authors declare that they have no conflicts of interest.
references
1. Hill DL, Maurer CR Jr, Maciunas RJ, Barwise JA, Fitzpatrick MJ, Wang MY. Measurement of intra-operative brain surface deformation under a craniotomy. Neurosurgery. 1998;43:514–26. 2. Maurer CR Jr, Hill DL, Martin AJ, Liu H, McCue M, Rueckert D, et al. Investigation of intraoperative brain deformation using a 1.5-T interventional MR system: Preliminary results. IEEE Trans Med Imaging. 1998;17:817–25. 3. Hata N, Nabavi A, Wells WM 3rd. Three-dimensional optical flow method for measurement of volumetric brain deformation from intra-operative MR images. J Comput Assist Tomogr. 2000;24:531–8. 4. Nabavi A, Black PM, Gering DT. Serial intra-operative magnetic resonance imaging of brain shift. Neurosurgery. 2001;48:787–97. 5. Nimsky C, Ganslandt O, Hastreiter P, Fahlbusch R. Intraoperative compensation for brain shift. Surg Neurol. 2001;56:357–64. 6. Reinges MHT, Nguyen HH, Krings T, Hutter BO, Rohde V, Gilsbach JM. Course of brain shift during microsurgical resection of supratentorial cerebral lesions: Limits of conventional neuronavigation. Acta Neurochir (Wien). 2004;146:369. 7. Roldan P. Experiencia con el uso de la resonancia magnética Intraoperatoria de bajo campo en Neurocirugía para el tratamiento de neoplasias cerebrales [tesis]. Valencia: Facultad de Medicina. Universitat de Valéncia; 2016. 8. Unsgaard G, Rygh OM, Selbekk T, Müller TB, Kolstad F, Lindseth F, et al. Intraoperative 3D ultrasound in neurosurgery. Acta Neurochir (Wien). 2006;148:235–53. 9. Black PM, Moriarty T, Alexander E 3rd. Development and implementation of intraoperative magnetic resonance imaging and its neurosurgical applications. Neurosurgery. 1997;41:831–45. 10. Gluch L, Walker DG. Intraoperative magnetic resonance: The future of surgery. ANZ J Surg. 2002;72:426–36.
7
11. Lewin JS, Metzger A, Selman WR. Intraoperative magnetic resonance image guidance in neurosurgery. J Magn Reson Imaging. 2000;12:512–24. 12. Jolesz FA. Interventional and intraoperative MRI: A general overview of the field. J Magn Reson Imaging. 1998;8:3–7. 13. Tummala RP, Chu RM, Liu H, Truwit CL, Hall WA. Optimizing brain tumor resection. High-field interventional MR imaging. Neuroimaging Clin N Am. 2001;11:673–83. 14. Berkenstadt H, Perel A, Ram Z, Feldman Z, Nahtomi-Shick O, Hadani M. Anesthesia for magnetic resonance guided neurosurgery: initial experience with a new open magnetic resonance imaging system. J Neurosurg Anesthesiol. 2001;13:158–62. 15. Hall W, Truwit C. Intraoperative MR-guided neurosurgery. J Magn Reson Imag. 2008;27:368–75. 16. Lipson AC, Gargollo PC, Black PM. Intraoperative magnetic resonance imaging: considerations for the operating room of the future. J Clin Neurosci. 2001;8:305–10. 17. Lacroix M, Abi-Said D, Fourney DR, Gokaslan ZL, Shi W, DeMonte F, et al. A multivariate analysis of 416 patients with glioblastoma multiforme: prognosis, extent of resection, and survival. J Neurosurg. 2001;95:190–8. 18. Stummer W, Reulen HJ, Meinel T. Extent of resection and survival in glioblastoma multiforme: identification of and adjustment for bias. Neurosurgery. 2008;62:564–76. 19. Chaichana KL, Halthore AN, Parker SL. Factors involved in maintaining prolonged functional independence following supratentorial glioblastoma resection. J Neurosurg. 2011;114:604–12. 20. Brown PD, Maurer MJ, Rummans TA. A prospective study of quality of life in adults with newly diagnosed high-grade gliomas: The impact of the extent of resection on quality of life and survival. Neurosurgery. 2005;57:495–504. ˜ 21. Brell M, Roldán P, González E, Llinàs P, Ibánez J. Implantación de la primera resonancia magnética intraoperatoria en un ˜ hospital de la red sanitaria pública espanola: experiencia inicial, viabilidad y dificultades en nuestro entorno. Neurocirugía. 2013;24:11–21. 22. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A, et al. The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol. 2007;114: 97–109. 23. Sanai N, Berger MS. Glioma extent of resection and its impact on patient outcome. Neurosurgery. 2008;62:753–64. 24. Dorward NL, Alberti O, Velani B. Postimaging brain distortion: magnitude, correlates, and impact on neuronavigation. J Neurosurg. 1998;88:656–62. 25. Roberts DW, Hartov A, Kennedy FE. Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery. 1998;43:749–60. 26. Barone D, Lawrie TA, Hart MG. Image guided surgery for the resection of brain tumours. Cochrane Database Syst Rev. 2014. 27. Nimsky C, Ganslandt O, Buvhfelder M, Fahlbusch R. Glioma surgery evaluated by intraoperative low-field magnetic resonance imaging. Acta Neurochir Suppl (Wien). 2003;85:55–63. 28. Senft C, Franz K, Ulrich CT, Bink A, Szelényi A, Gasser T, et al. Low field intraoperative MRI-guided surgery of gliomas: a single center experience. Clin Neurol Neurosurg. 2010;112:237–43. 29. Herrera R, Ledesma JL, Pomata H, Lambre J, Rojas H, Houssay A, et al. El desarrollo e implementación de la resonancia magnética intraoperatoria en neurocirugía (REMAIN) y su aplicación en 83 intervenciones en la Argentina. Rev Argent Neurocir. 2007:21. 30. Senft C, Bink A, Franz K, Vatter H, Gasser T, Seifert V. Intraoperative MRI guidance and extent of resection in glioma
ARTICLE IN PRESS 8
n e u r o c i r u g i a . 2 0 1 7;x x x(x x):xxx–xxx
surgery: a randomised, controlled trial. Lancet Oncol. 2011;12:997–1003. 31. Senft C, Bink A, Franz K, Gasser T, Seifert V. Intraoperative MRI-Guided vs. conventional microsurgical brain tumor resection - results of a prospective randomized trial. J Neurosurg. 2011;115:602. 32. Senft C, Bink A, Heckelmann M, Gasser T, Seifert V. Glioma extent of resection and ultra-low-field iMRI: Interim analysis of a prospective randomized controlled trial. Acta Neurochir Suppl. 2011;109:49–53. 33. Hirschl RA, Wilson J, Miller B, Bergese S, Chiocca EA. The predictive value of low-field strength magnetic resonance
imaging for intraoperative residual tumor detection. J Neurosurg. 2009;111:252–7. 34. Haydon DH, Chicoine MR, Dacey RG. The impact of high-field-strength intraoperative magnetic resonance imaging on brain tumor management. Neurosurgery. 2013;60:92–7. 35. Makary M, Chiocca EA, Erminy N, Antor M, Bergese SD, Abdel-Rasoul M, et al. Clinical and economic outcomes of low-field intraoperative MRI-guided tumor resection neurosurgery. J Magn Reson Imaging. 2011;34:1022–30.