Mapping and Monitoring in Glioma Surgery
Mapping and Monitoring in Glioma Surgery
Neurosurgical Use. Although initially used primarily for scientific purposes, fMRI was quickly adopted for clinical purposes and has become a widely available clinical application for presurgical evaluation of functional areas prior to brain tumor surgery. In patients with tumors in eloquent brain regions, fMRI has been routinely used for many years as a noninvasive brain-mapping tool to guide neurosurgical treatment decisions (Fig. 1).
(Enlarge Image)
Figure 1.
Axial fMRI study showing a motor region dorsolateral to the LGG in the hand knob of the precentral gyrus.
Technical Details. Technically, fMRI detects a surrogate parameter of neuronal activation, a blood oxygenation level–dependent effect from activation-induced perfusion-related changes in the blood oxygen level from neurovascular coupling. Thereby, fMRI depicts functional networks involved in an investigated function such as a motor or language task. These networks are not necessarily required or critical for the assessed function in its completeness, because neuronal activity measured indirectly by blood oxygen levels is globally assessed and a differentiation of essential versus nonessential areas for function is not possible.
Current Evidence. Several studies have investigated the accuracy of fMRI, presenting promising results concerning sensitivity and specificity to adequately predict motor function compared with DCS. In these studies, motor fMRI has been proven to be a reliable method to localize motor function, which facilitates surgical planning and reduces the time needed for intraoperative mapping. However, there are also contradictory studies showing a large deviation of fMRI-depicted areas compared with areas detected by electrophysiological methods such as DCS and nTMS, which should keep us thinking critically in terms of fMRI use for preoperative planning.
Limitations. Especially in the vicinity of tumors, vascular changes can lead to a neurovascular uncoupling instead of the regular coupling and thereby produce false-negative fMRI results, making fMRI unreliable for resective planning. These false-negative results from neurovascular uncoupling could lead to the misinterpretation of eloquent tissue being noneloquent, and to subsequent resection with the associated neurological sequelae.
Concerning fMRI language localization, a review by Giussani et al. summarized the available data and evaluated the reliability of presurgical fMRI language mapping compared with DCS from 9 different studies. Several of the studies that these authors summarized investigated the sensitivity and specificity of fMRI in mapping language function. Because different language tasks and different MRI machines, software, analysis paradigms, and algorithms were used, the specificity and sensitivity for presurgical language localization by fMRI was highly variable. Five studies provided sufficient statistical data showing that sensitivity ranged from 59% to 100% and that specificity ranged from 0% to 97% compared with intra-operative DCS mapping during awake surgery. These data show that fMRI in its present form cannot be used reliably to guide resections of tumors close to eloquent areas. It can certainly give a presurgical impression of functional organization; however, neurovascular uncoupling and potential false-negative results always have to be kept in mind.
At present, fMRI motor and language mapping is not able to detect critical functional areas reliably prior to surgery and cannot be recommended for planning of resective surgery of tumors in a potentially eloquent location. It can only serve as an adjunct to other methods, especially intraoperative electrical stimulation mapping.
Neurosurgical Use. So far, MEG has been applied for central sulcus localization, primary auditory and visual cortex delineation, language lateralization, and localization of the motor cortex. Several studies have assessed the feasibility of MEG motor or language mapping prior to surgery.
Technical Details. Magnetoencephalography is the detection of magnetic fields produced by bioelectric currents from neuronal activation, which means it is a direct measurement of cortical activity. To detect magnetic fields produced by cerebral electric activity at picoor femtotesla levels, supraconductive sensors and magnetically shielded environments are necessary. Use of MEG allows the detection of spontaneous activity or evoked activity time-locked to certain stimuli. The coregistration of MEG source localization with anatomical MRI allows the use of this technique in presurgical localization of activity and surgical planning. The MEG data obtained during motor tasks can be used to reconstruct spatiotemporal dynamics of brain sources.
Current Evidence. To assess MEG accuracy, some studies compared preoperative MEG motor mapping with intraoperative DCS in patients with brain tumors and found a reliable delineation of MEG motor areas in comparison with intraoperative DCS. Tarapore et al. compared nTMS and MEG motor mapping with intraoperative DCS in patients with brain tumor and found a deviation of nTMS, MEG, and DCS motor sites of 4.7–2.1 mm. The use of MEG for motor mapping confirmed functional activity within LGGs, a nd a ll p atients who h ad f unctional t issue located within the tumor and who underwent GTR despite this finding suffered from new neurological deficits after surgery, showing the predictive accuracy of MEG.
The experience with MEG for presurgical language mapping is limited. Only a few studies have assessed language prior to surgery and revealed a good agreement with intraoperative electrophysiological language mapping. Tarapore et al. recently published a study reporting their results for mapping language function via DCS, nTMS, and MEG. In 12 patients in whom language sites were outlined by MEG for verb generation and object naming, the sites correlated with nTMS sites in 5 of these patients and with DCS sites in 2.
Earlier MEG studies of language lateralization have demonstrated promising results. As discussed by Tarapore et al., although MEG lacks the accuracy of TMS, Findlay et al. have highlighted the use of MEG for a more global analysis of language lateralization that even predicted surgical outcome for patients with eloquent gliomas.
Concerning presurgical risk assessment based on MEG data for a neurological complication from lesion resection, a series of 119 patients was published. In this patient series 46% were not treated surgically because of tumor invasion of eloquent cortex as assessed by MEG, whereas 54% underwent resection on MEG mapping, with an associated neurological deterioration of 6%.
Limitations. Although MEG is efficient in terms of temporal and spatial resolution, the distribution of MEG mapping is still limited due to the high costs of the modality and as a consequence its limited availability. Therefore, the amount of data presently available for tumor resection is still quite scarce.
Neurosurgical Use. Transcranial magnetic stimulation is an older technique, which uses a transcranial magnetic field to elicit a cortical electrical field and thus neuronal activation or inhibition. The integration of TMS in an adjusted navigation system made it available for presurgical functional mapping, because this allows for an exact cortical representation of stimulation or inhibition by TMS and the induced or inhibited functional response. Single-pulse TMS is readily used for motor response stimulation, whereas repetitive TMS inhibits functional activation, leading to a so-called virtual lesion, which has been recently used for language mapping.
Technical Details. Navigated TMS is a unique method for mapping essential brain function due to a fundamental difference between TMS and other functional brain imaging tools such as fMRI and MEG. When stimulated or inhibited cortical areas evoke a measurable physiological response, these areas are mandatory; i.e., essential to the observed reaction. Other imaging methods such as fMRI and MEG detect and map all brain areas that participate in a given task or reaction; i.e., the entire network that is responsible for the reaction, without differentiating essential from nonessential areas.
Current Evidence. Navigated TMS has recently proven to be suitable for clinical mapping of the cortical motor areas and for the simultaneous assessment of the functional status of the motor tracts. For a detailed evaluation of their mapping accuracy, the noninvasive nTMS method and the DCS cortical map of motor function were compared by calculating the distances between the respective hot spots or centers of gravity of target muscles. Although there are various methodological flaws inherent to this approach—from the navigational error to the misconception of one hot spot or stable center of gravity for each muscle—all studies on nTMS mapping accuracy reported distances between both methods of 1.1 and 14.8 mm for the hot spot comparison of the adductor pollicis brevis muscle.
Due to the low expense of nTMS and the ease of use associated with the sufficient accuracy of this true electrophysiological method, an increasing number of centers use presurgical nTMS mapping when gliomas are located in or near eloquent areas.
Concerning presurgical planning, one study showed that brain mapping by nTMS influenced the surgical approach and the planned EOR, and even changed the indication in a small group of patients. Furthermore, it has been demonstrated that tumor-infiltrated eloquent tissue that prevented total tumor resection can become resectable due to functional reorganization over time as assessed by nTMS.
Recently, the first study addressing the impact of nTMS on the oncological and functional outcome of brain tumor surgery was published. T he s tudy c ompared t he o utcomes of 100 patients treated with preoperative nTMS examination to those of patients in a matched in-house historical pre-nTMS group and revealed better neurological outcomes combined with increased EOR in the nTMSmapped patient cohort. A similar second study revealed that these beneficial effects of preoperative nTMS also occurred in a subgroup of patients with LGGs. The authors provided data that nTMS caused a conversion of the treatment approach from biopsy or no surgery to surgery in 37 of 54 patients (68.5%).
Additionally, cortical nTMS mapping results can serve as a measure to standardize the visualization of subcortical motor fiber tracts. The cortical outline of essential motor areas by nTMS can be used as a seed region for initiation of a fiber tracking algorithm, leading to a more accurate and reproducible delineation of subcortical fibers than can be obtained with the standard tracking approach. This method has recently been shown to contribute to preoperative risk stratification by measuring the distance between subcortical tumors and the corticospinal tract (CST) as outlined by nTMS-based DTI-FT. The authors demonstrated that none of the 205 patients showed surgery-related paresis when the minimal distance between tumor and CST was larger than 10 mm.
Whereas the protocol for nTMS motor cortical mapping is well established and reliable, nTMS language mapping for neurosurgical patients performed using repetitive TMS stimulation is still evolving. Even so, the first published series including the first tested protocol has already reported a good overall correlation between repetitive nTMS and DCS, which was especially true for negatively mapped brain regions, resulting in a high negative predictive value. Nonetheless, the insufficient specificity in posterior perisylvian language areas demands further refinements of the mapping protocol, a difficulty that already has been partially overcome by modifications.
Navigated TMS language mapping has already been used as a tool for follow-up examinations prior to repeated awake surgery and to detect a shift of language function to the contralateral perisylvian region, and this method is therefore able to visualize language plasticity in patients with brain tumor.
Limitations. nTMS motor mapping has proven to be beneficial in treatment planning for the resection of tumors close to eloquent regions, but this technique has limitations. Although mapping results obtained prior to surgery are useful, choosing an approach using presurgical mapping does not allow the surgeon to waive intraoperative electrical stimulation mapping and monitoring. Moreover, although nTMS serves to estimate resectability, EOR, and surgical approaches, and to define starting points for intraoperative electrical stimulation, nTMS language mapping will not allow the general abandonment of awake language mapping. However, this method could reduce awake mapping time, and for only a small subgroup of patients unable to undergo mapping while awake or electrical stimulation while asleep, presurgical nTMS mapping can reduce the surgical risk.
Neurosurgical Use. Although fMRI, MEG, and TMS allow for cortical localization of neurological function, none of these techniques is able to delineate subcortical white matter tracts arising from or connecting relevant cortical areas. Again, the aim of presurgical functional localization is to assess lesion resectability and surgical risk prior to taking the patient to the operating room and to provide intraoperative orientation regarding when to expect relevant subcortical fibers and when to initiate subcortical electrical stimulation mapping during surgery.
Technical Details. Only a single technique—DTI-FT—is available to noninvasively depict subcortical white matter tracts, the preservation of which is also important to maintain neurological functions. However, the reconstruction of subcortical fiber tracts from diffusion tensor vectors is a purely anatomical imaging analysis that does not include true electrophysiological functional data.
Current Evidence. Particularly for preoperative mapping of the CST in relation to a tumor, DTI-FT is a commonly used technique. V arious s tudies h ave i nvestigated t he accuracy of pre- and intraoperative DTI-FT compared with intraoperative SCS mapping of the motor pathways, which revealed mostly good intraoperative correlations of DTI-FT and SCS, depending on intraoperative brain shift. The sensitivity of CST detection was 95% in a series by Bello et al., and the sensitivity was 93% at a specificity of 93% in a series by Zhu et al. Several reports have shown that especially in LGGs, fibers were frequently located inside the tumor, and DTI-FT was able to visualize these fibers. This aspect contributes to the estimation of resectability. In a series of 73 gliomas, Castellano et al. revealed that an infiltration or displacement of the DTI-FT CST was associated with a lower probability of total tumor resection. In a prospective randomized trial including 238 patients, the presurgical DTI-FT of the CST and inclusion in the neuronavigation system did lead to a larger EOR, an improved clinical outcome with regard to new deficits, and improved overall functional status in comparison with the control group without FT.
However, in addition to the limitation of solely anatomical fiber delineation, a major limitation of the intraoperative use of DTI-FT when integrated into neuronavigation is the brain shift, which has already happened when subcortical fibers are reached during a tumor resection. A study by Nimsky et al. revealed CST shifts at a range from −8 to 15 mm, where direction of shift was not predictable. Thus, DTI-FT is a valuable additional tool for preoperative planning, but it requires a strict control modality when used intraoperatively within the navigation systems during resection of tumors close to the CST; that control is intraoperative electrical stimulation mapping.
The introduction of nTMS in neurosurgical planning led to the fusion of the neurophysiologically based nTMS motor mapping and DTI-FT as a pure imaging technique by using the motor cortex as outlined by nTMS as a seed region for DTI-FT. Two recently reported studies have investigated this approach and both found a higher grade of standardization of DTI-FT when combined with nTMS. Moreover, this technique can be used to clarify highly impaired and unclear functional anatomy prior to surgery and enable the surgeon to get a better understanding of the essential structures, which have to be preserved (Fig. 2).
(Enlarge Image)
Figure 2.
Navigated TMS is able to clarify highly impaired and unclear functional anatomy prior to surgery and enables the surgeon to get a better understanding of the essential structures, which have to be preserved. Green = motor cortex; light blue = LGG; dark blue = nTMS-based DTI-FT of the CST; dark pink = language-involved cortex.
Apart from motor system tracking, DTI-FT can also be used to noninvasively visualize language tracts such as the arcuate fasciculus and the inferior frontooccipital fasciculus, and comparisons of this method to intraoperative electrical stimulation have been published. Similar to mapping of the CST, SCS for language tracts also correlated well with preoperative DTI-FT, with a sensitivity of 97%. As for motor tracts, a displacement or infiltration of subcortical language fiber tracts is predictive for a lower probability of a complete resection.
Limitations. It remains unclear whether DTI-FT of language tracts based on cortical nTMS language representation would also be beneficial. This approach is currently being evaluated in our department, but studies on the value of such an approach are still lacking. Figure 3 gives an impression of combined cortical and subcortical nTMS-based planning to determine areas of language and motor function prior to surgery.
(Enlarge Image)
Figure 3.
This neuronavigational 3D screenshot gives an impression of a combined cortical and subcortical nTMS-based planning image obtained prior to surgery for language (dark pink) and motor function (green) including nTMS-based DTI-FT of the CST (yellow) and language tracts such as the arcuate fascicle (blue).
Preoperative Functional Mapping
Functional MRI
Neurosurgical Use. Although initially used primarily for scientific purposes, fMRI was quickly adopted for clinical purposes and has become a widely available clinical application for presurgical evaluation of functional areas prior to brain tumor surgery. In patients with tumors in eloquent brain regions, fMRI has been routinely used for many years as a noninvasive brain-mapping tool to guide neurosurgical treatment decisions (Fig. 1).
(Enlarge Image)
Figure 1.
Axial fMRI study showing a motor region dorsolateral to the LGG in the hand knob of the precentral gyrus.
Technical Details. Technically, fMRI detects a surrogate parameter of neuronal activation, a blood oxygenation level–dependent effect from activation-induced perfusion-related changes in the blood oxygen level from neurovascular coupling. Thereby, fMRI depicts functional networks involved in an investigated function such as a motor or language task. These networks are not necessarily required or critical for the assessed function in its completeness, because neuronal activity measured indirectly by blood oxygen levels is globally assessed and a differentiation of essential versus nonessential areas for function is not possible.
Current Evidence. Several studies have investigated the accuracy of fMRI, presenting promising results concerning sensitivity and specificity to adequately predict motor function compared with DCS. In these studies, motor fMRI has been proven to be a reliable method to localize motor function, which facilitates surgical planning and reduces the time needed for intraoperative mapping. However, there are also contradictory studies showing a large deviation of fMRI-depicted areas compared with areas detected by electrophysiological methods such as DCS and nTMS, which should keep us thinking critically in terms of fMRI use for preoperative planning.
Limitations. Especially in the vicinity of tumors, vascular changes can lead to a neurovascular uncoupling instead of the regular coupling and thereby produce false-negative fMRI results, making fMRI unreliable for resective planning. These false-negative results from neurovascular uncoupling could lead to the misinterpretation of eloquent tissue being noneloquent, and to subsequent resection with the associated neurological sequelae.
Concerning fMRI language localization, a review by Giussani et al. summarized the available data and evaluated the reliability of presurgical fMRI language mapping compared with DCS from 9 different studies. Several of the studies that these authors summarized investigated the sensitivity and specificity of fMRI in mapping language function. Because different language tasks and different MRI machines, software, analysis paradigms, and algorithms were used, the specificity and sensitivity for presurgical language localization by fMRI was highly variable. Five studies provided sufficient statistical data showing that sensitivity ranged from 59% to 100% and that specificity ranged from 0% to 97% compared with intra-operative DCS mapping during awake surgery. These data show that fMRI in its present form cannot be used reliably to guide resections of tumors close to eloquent areas. It can certainly give a presurgical impression of functional organization; however, neurovascular uncoupling and potential false-negative results always have to be kept in mind.
At present, fMRI motor and language mapping is not able to detect critical functional areas reliably prior to surgery and cannot be recommended for planning of resective surgery of tumors in a potentially eloquent location. It can only serve as an adjunct to other methods, especially intraoperative electrical stimulation mapping.
Magnetoencephalography
Neurosurgical Use. So far, MEG has been applied for central sulcus localization, primary auditory and visual cortex delineation, language lateralization, and localization of the motor cortex. Several studies have assessed the feasibility of MEG motor or language mapping prior to surgery.
Technical Details. Magnetoencephalography is the detection of magnetic fields produced by bioelectric currents from neuronal activation, which means it is a direct measurement of cortical activity. To detect magnetic fields produced by cerebral electric activity at picoor femtotesla levels, supraconductive sensors and magnetically shielded environments are necessary. Use of MEG allows the detection of spontaneous activity or evoked activity time-locked to certain stimuli. The coregistration of MEG source localization with anatomical MRI allows the use of this technique in presurgical localization of activity and surgical planning. The MEG data obtained during motor tasks can be used to reconstruct spatiotemporal dynamics of brain sources.
Current Evidence. To assess MEG accuracy, some studies compared preoperative MEG motor mapping with intraoperative DCS in patients with brain tumors and found a reliable delineation of MEG motor areas in comparison with intraoperative DCS. Tarapore et al. compared nTMS and MEG motor mapping with intraoperative DCS in patients with brain tumor and found a deviation of nTMS, MEG, and DCS motor sites of 4.7–2.1 mm. The use of MEG for motor mapping confirmed functional activity within LGGs, a nd a ll p atients who h ad f unctional t issue located within the tumor and who underwent GTR despite this finding suffered from new neurological deficits after surgery, showing the predictive accuracy of MEG.
The experience with MEG for presurgical language mapping is limited. Only a few studies have assessed language prior to surgery and revealed a good agreement with intraoperative electrophysiological language mapping. Tarapore et al. recently published a study reporting their results for mapping language function via DCS, nTMS, and MEG. In 12 patients in whom language sites were outlined by MEG for verb generation and object naming, the sites correlated with nTMS sites in 5 of these patients and with DCS sites in 2.
Earlier MEG studies of language lateralization have demonstrated promising results. As discussed by Tarapore et al., although MEG lacks the accuracy of TMS, Findlay et al. have highlighted the use of MEG for a more global analysis of language lateralization that even predicted surgical outcome for patients with eloquent gliomas.
Concerning presurgical risk assessment based on MEG data for a neurological complication from lesion resection, a series of 119 patients was published. In this patient series 46% were not treated surgically because of tumor invasion of eloquent cortex as assessed by MEG, whereas 54% underwent resection on MEG mapping, with an associated neurological deterioration of 6%.
Limitations. Although MEG is efficient in terms of temporal and spatial resolution, the distribution of MEG mapping is still limited due to the high costs of the modality and as a consequence its limited availability. Therefore, the amount of data presently available for tumor resection is still quite scarce.
Navigated TMS
Neurosurgical Use. Transcranial magnetic stimulation is an older technique, which uses a transcranial magnetic field to elicit a cortical electrical field and thus neuronal activation or inhibition. The integration of TMS in an adjusted navigation system made it available for presurgical functional mapping, because this allows for an exact cortical representation of stimulation or inhibition by TMS and the induced or inhibited functional response. Single-pulse TMS is readily used for motor response stimulation, whereas repetitive TMS inhibits functional activation, leading to a so-called virtual lesion, which has been recently used for language mapping.
Technical Details. Navigated TMS is a unique method for mapping essential brain function due to a fundamental difference between TMS and other functional brain imaging tools such as fMRI and MEG. When stimulated or inhibited cortical areas evoke a measurable physiological response, these areas are mandatory; i.e., essential to the observed reaction. Other imaging methods such as fMRI and MEG detect and map all brain areas that participate in a given task or reaction; i.e., the entire network that is responsible for the reaction, without differentiating essential from nonessential areas.
Current Evidence. Navigated TMS has recently proven to be suitable for clinical mapping of the cortical motor areas and for the simultaneous assessment of the functional status of the motor tracts. For a detailed evaluation of their mapping accuracy, the noninvasive nTMS method and the DCS cortical map of motor function were compared by calculating the distances between the respective hot spots or centers of gravity of target muscles. Although there are various methodological flaws inherent to this approach—from the navigational error to the misconception of one hot spot or stable center of gravity for each muscle—all studies on nTMS mapping accuracy reported distances between both methods of 1.1 and 14.8 mm for the hot spot comparison of the adductor pollicis brevis muscle.
Due to the low expense of nTMS and the ease of use associated with the sufficient accuracy of this true electrophysiological method, an increasing number of centers use presurgical nTMS mapping when gliomas are located in or near eloquent areas.
Concerning presurgical planning, one study showed that brain mapping by nTMS influenced the surgical approach and the planned EOR, and even changed the indication in a small group of patients. Furthermore, it has been demonstrated that tumor-infiltrated eloquent tissue that prevented total tumor resection can become resectable due to functional reorganization over time as assessed by nTMS.
Recently, the first study addressing the impact of nTMS on the oncological and functional outcome of brain tumor surgery was published. T he s tudy c ompared t he o utcomes of 100 patients treated with preoperative nTMS examination to those of patients in a matched in-house historical pre-nTMS group and revealed better neurological outcomes combined with increased EOR in the nTMSmapped patient cohort. A similar second study revealed that these beneficial effects of preoperative nTMS also occurred in a subgroup of patients with LGGs. The authors provided data that nTMS caused a conversion of the treatment approach from biopsy or no surgery to surgery in 37 of 54 patients (68.5%).
Additionally, cortical nTMS mapping results can serve as a measure to standardize the visualization of subcortical motor fiber tracts. The cortical outline of essential motor areas by nTMS can be used as a seed region for initiation of a fiber tracking algorithm, leading to a more accurate and reproducible delineation of subcortical fibers than can be obtained with the standard tracking approach. This method has recently been shown to contribute to preoperative risk stratification by measuring the distance between subcortical tumors and the corticospinal tract (CST) as outlined by nTMS-based DTI-FT. The authors demonstrated that none of the 205 patients showed surgery-related paresis when the minimal distance between tumor and CST was larger than 10 mm.
Whereas the protocol for nTMS motor cortical mapping is well established and reliable, nTMS language mapping for neurosurgical patients performed using repetitive TMS stimulation is still evolving. Even so, the first published series including the first tested protocol has already reported a good overall correlation between repetitive nTMS and DCS, which was especially true for negatively mapped brain regions, resulting in a high negative predictive value. Nonetheless, the insufficient specificity in posterior perisylvian language areas demands further refinements of the mapping protocol, a difficulty that already has been partially overcome by modifications.
Navigated TMS language mapping has already been used as a tool for follow-up examinations prior to repeated awake surgery and to detect a shift of language function to the contralateral perisylvian region, and this method is therefore able to visualize language plasticity in patients with brain tumor.
Limitations. nTMS motor mapping has proven to be beneficial in treatment planning for the resection of tumors close to eloquent regions, but this technique has limitations. Although mapping results obtained prior to surgery are useful, choosing an approach using presurgical mapping does not allow the surgeon to waive intraoperative electrical stimulation mapping and monitoring. Moreover, although nTMS serves to estimate resectability, EOR, and surgical approaches, and to define starting points for intraoperative electrical stimulation, nTMS language mapping will not allow the general abandonment of awake language mapping. However, this method could reduce awake mapping time, and for only a small subgroup of patients unable to undergo mapping while awake or electrical stimulation while asleep, presurgical nTMS mapping can reduce the surgical risk.
Diffusion Tensor Imaging Fiber Tracking
Neurosurgical Use. Although fMRI, MEG, and TMS allow for cortical localization of neurological function, none of these techniques is able to delineate subcortical white matter tracts arising from or connecting relevant cortical areas. Again, the aim of presurgical functional localization is to assess lesion resectability and surgical risk prior to taking the patient to the operating room and to provide intraoperative orientation regarding when to expect relevant subcortical fibers and when to initiate subcortical electrical stimulation mapping during surgery.
Technical Details. Only a single technique—DTI-FT—is available to noninvasively depict subcortical white matter tracts, the preservation of which is also important to maintain neurological functions. However, the reconstruction of subcortical fiber tracts from diffusion tensor vectors is a purely anatomical imaging analysis that does not include true electrophysiological functional data.
Current Evidence. Particularly for preoperative mapping of the CST in relation to a tumor, DTI-FT is a commonly used technique. V arious s tudies h ave i nvestigated t he accuracy of pre- and intraoperative DTI-FT compared with intraoperative SCS mapping of the motor pathways, which revealed mostly good intraoperative correlations of DTI-FT and SCS, depending on intraoperative brain shift. The sensitivity of CST detection was 95% in a series by Bello et al., and the sensitivity was 93% at a specificity of 93% in a series by Zhu et al. Several reports have shown that especially in LGGs, fibers were frequently located inside the tumor, and DTI-FT was able to visualize these fibers. This aspect contributes to the estimation of resectability. In a series of 73 gliomas, Castellano et al. revealed that an infiltration or displacement of the DTI-FT CST was associated with a lower probability of total tumor resection. In a prospective randomized trial including 238 patients, the presurgical DTI-FT of the CST and inclusion in the neuronavigation system did lead to a larger EOR, an improved clinical outcome with regard to new deficits, and improved overall functional status in comparison with the control group without FT.
However, in addition to the limitation of solely anatomical fiber delineation, a major limitation of the intraoperative use of DTI-FT when integrated into neuronavigation is the brain shift, which has already happened when subcortical fibers are reached during a tumor resection. A study by Nimsky et al. revealed CST shifts at a range from −8 to 15 mm, where direction of shift was not predictable. Thus, DTI-FT is a valuable additional tool for preoperative planning, but it requires a strict control modality when used intraoperatively within the navigation systems during resection of tumors close to the CST; that control is intraoperative electrical stimulation mapping.
The introduction of nTMS in neurosurgical planning led to the fusion of the neurophysiologically based nTMS motor mapping and DTI-FT as a pure imaging technique by using the motor cortex as outlined by nTMS as a seed region for DTI-FT. Two recently reported studies have investigated this approach and both found a higher grade of standardization of DTI-FT when combined with nTMS. Moreover, this technique can be used to clarify highly impaired and unclear functional anatomy prior to surgery and enable the surgeon to get a better understanding of the essential structures, which have to be preserved (Fig. 2).
(Enlarge Image)
Figure 2.
Navigated TMS is able to clarify highly impaired and unclear functional anatomy prior to surgery and enables the surgeon to get a better understanding of the essential structures, which have to be preserved. Green = motor cortex; light blue = LGG; dark blue = nTMS-based DTI-FT of the CST; dark pink = language-involved cortex.
Apart from motor system tracking, DTI-FT can also be used to noninvasively visualize language tracts such as the arcuate fasciculus and the inferior frontooccipital fasciculus, and comparisons of this method to intraoperative electrical stimulation have been published. Similar to mapping of the CST, SCS for language tracts also correlated well with preoperative DTI-FT, with a sensitivity of 97%. As for motor tracts, a displacement or infiltration of subcortical language fiber tracts is predictive for a lower probability of a complete resection.
Limitations. It remains unclear whether DTI-FT of language tracts based on cortical nTMS language representation would also be beneficial. This approach is currently being evaluated in our department, but studies on the value of such an approach are still lacking. Figure 3 gives an impression of combined cortical and subcortical nTMS-based planning to determine areas of language and motor function prior to surgery.
(Enlarge Image)
Figure 3.
This neuronavigational 3D screenshot gives an impression of a combined cortical and subcortical nTMS-based planning image obtained prior to surgery for language (dark pink) and motor function (green) including nTMS-based DTI-FT of the CST (yellow) and language tracts such as the arcuate fascicle (blue).