**4.1 The traditional approach (Yasargil's lesson)**

Yasargil lesson's may be summarized as follows: 1) glioma is an infiltrative tumor only in its advanced stages of growth; 2) its growth pattern follows functional pathways and respects compartmental boundaries (neo-archipallial, central nuclei, white matter and ventricles); and 3) tumor debulking starts from the center and proceeds in a circumferential manner till the white matter is exposed ("the deflating ball and one window" techniques) (Yasargil, 1996b, 1996c; Yasargil et al., 2004).

This last point is justified by the risk of damage to adjacent brain areas and forms the basic principle of resection strategy. In light of recent pathological and functional knowledge, however, this defensive strategy for treating both LGGs and HGGs, which seeks foremost to avoid brain injury without considering the infiltrated brain or the functional characteristics of adjacent brain areas, now seems largely unwarranted. Nonetheless, it is not an absolutely safe technique, since HGGs are often highly vascularized tumors and bleeding may mask vascular or parenchymal structures and result in involuntary injury. Moreover, tumor manipulation may induce tissue swelling and physical modifications, altering the typical

Surgical Treatment of Supratentorial Glioma in Eloquent Areas 305

infiltrated brain and less from the tumor. Since diffuse tumors are heterogeneous, it's better to see the phenotype of the single variant from the beginning and possibly to stay around the tumor, creating a corridor between the infiltrated brain and the normal brain. Starting from the corridor created around the tumor, additional space is obtained by removing or retracting the tumor and leaving the brain untouched. The rule is: expose the normal brain, map and stay away from it. Working very closely to normal brain tissue implies avoiding any pressure and retraction that could cause brain injury. As a rule, use of the self-retaining spatula is unnecessary and dangerous. This strategy provides us with a better anatomical orientation in a context which is still similar to the anatomo-functional picture as acquired

Following large, randomized prospective studies in Europe, several countries have already approved, and are now using, 5-ALA for improved quality of resection in HGGs, aimed at

The standard for subcortical mapping for language is not yet defined: current amplitude, functional significance of fasciculi, safe distance from fasciculi or tolerance of partial injury, tests to be used, false negative stimulation rate, definition of positive sites, method of validation of positive sites. In particular, continuous monitoring is affected by the learning

The limiting factor to total removal is not always the eloquent areas. In large and diffuse tumors we can state that a strategy aimed at exploring and removing safe areas first is a preliminary condition to continue this aggressive strategy in order to obtain maximal resection which justifies risks in eloquent areas. Safe exposure of deep structures is not always feasible; and facing increasing structural and functional uncertainties is preferable to stopping the operation before taking additional risks in eloquent areas, assuming that the degree of cytoreduction has limited value if total removal cannot be achieved (Lacroix et al., 2001). Vascular injury may be another limiting factor to tumor removal, especially in insular tumors where transit and terminal arteries, normal and pathological arteries, and venous

The approach to temporal and insular tumors may be trans-cortical or trans-sylvian. When corticectomy is done, the tumor is devascularized by subpial dissection, while trans-sylvian dissection makes it possible to follow the vessels, free the tumor, and then decide which vessels to preserve and which to divide. The lenticulostriate arteries, the parietal trunk of the middle cerebral artery bifurcation on the non-dominant hemisphere and both trunks in the dominant hemisphere and the long perforators running posteriorly are the critical vessels to preserve in order to avoid permanent language and motor deficits. As a rule, vessels smaller than 0.5 mm are pathological. In these tumors, temporal or frontal basal cortical areas are seldom the critical point of resection, and multiple windows can be created between vessels. Accordingly, various sequential options may be weighed in order to start from safe areas and

Mapping techniques allow the neurosurgeon to recognize the internal capsule, which is the critical anatomical point, at least in a non dominant site, while additional information can be obtained from monitoring motor evoked potentials (MEPs) in the course of the operation since insular tumors carry a risk of vascular injury, higher than the risk of direct injury (Figure 1-3). Neuloh and Schramm highlight the semi-quantitative relation between MEP

to control arterial vessels (Duffau, 2009; Sanai et al., 2010, Simon et al., 2009).

assisting the surgeon according to the same resection strategy (Stummer et al., 2006).

effect, as described by Ojemann ( Ojemann J.G., et al., 2002).

drainage are ever-present challenges during the operation.

**4.2.1 Temporo-insular tumors** 

preoperatively with MRI.

characteristics of the normal brain, thus confounding the neurosurgeon's judgment, which basically relies on visual inspection.

The simple observation that, as a rule, recurrences originate from the margins of the operative cavity should underline that additional resection is a primary aim of surgery if it does not endanger functions that could impair patient performance, the accurate knowledge of which was the limiting factor of the strategy. Bearing this in mind, the rationale of the current approach becomes readily understandable: attack the tumor at its margins.

#### **4.2 The current approach according to mapping information**

As a preliminary consideration, the knowledge how a tumor grows may orientate the cortical approach. As the glioma grows in the gyrus, it infiltrates the short tracts (the u fibers) first and then the longer tracts, rather than the pia. Consequently, a direct or sulcal approach will be unsuitable for diffuse tumors in which resection of the involved gyrus and eventually the adjacent gyrus is preferable (Meyer et al., 2001). Corticectomy at this site, at the tumor margin, is the main surgical innovation of brain mapping. Cortical mapping highlights positive and negative areas before starting resection. This "negative mapping" strategy represents a paradigm shift in language and no-language mapping techniques by eliminating the neurosurgeon's reliance on positive site control in operative exposure, although validation for other related functions is still needed (Berger, 1994; Black et al, 1987; Keles & Berger, 2004; Matz et al., 1999; Sawaya et al., 1999; Toms et al., 1999).

Cortical LGGs may be nearly indistinguishable from normal brain tissue. The neuronavigator is therefore essential for precisely demarcating the superficial limits of the tumor. The arachnoid opening is also useful before deciding on the area of resection. The exact distinction acquired with the neuronavigation system may provide immediate visual feedback during the initial subcortical resection when the resection proceeds between different tissues.

The technique is to start from safe areas and where the vascular supply is expected: following the vessels, skeletonizing and preserving those serving eloquent areas and dividing the others, especially those supplying the tumor, in order to devascularize it. Subcortical pathways perpendicular to positive sites can be followed within the infiltrated brain as far as apparent normal brain is exposed or positive mapping is detected. This is not always possible in language areas, since the closer we come to positive sites, the easier it will be to have intraoperative impairment which will prevent further mapping. Consequently, in language mapping, DTI may be of help in orientating where to map on the tumor periphery in order to stimulate the least possible. DTI may be not considered as more than a imaginary guide.

This strategy may be changed in some situations, for instance, with less cooperative patients and during long operations, by adopting the following sequence: mapping, starting resection close to positive sites and following subcortical pathways, and then approach the vascular supply and safe areas. Repeating the mapping subcortically 2-3 times is usually enough, even when anatomical distortion occurs.

Starting from safe areas enables us to follow the tumor into the transition zone, from the margins to the infiltrated brain till the normal brain. First sight inspection means starting resection from the tumor periphery before causing possible injury, which can alter tissue color. Here, the microscope is essential for magnifying the region in order to differentiate normal brain from infiltrated brain. This major challenge can be successfully addressed only by experienced neurosurgeons. The corridor measures about 1 cm taken mostly from the

characteristics of the normal brain, thus confounding the neurosurgeon's judgment, which

The simple observation that, as a rule, recurrences originate from the margins of the operative cavity should underline that additional resection is a primary aim of surgery if it does not endanger functions that could impair patient performance, the accurate knowledge of which was the limiting factor of the strategy. Bearing this in mind, the rationale of the

As a preliminary consideration, the knowledge how a tumor grows may orientate the cortical approach. As the glioma grows in the gyrus, it infiltrates the short tracts (the u fibers) first and then the longer tracts, rather than the pia. Consequently, a direct or sulcal approach will be unsuitable for diffuse tumors in which resection of the involved gyrus and eventually the adjacent gyrus is preferable (Meyer et al., 2001). Corticectomy at this site, at the tumor margin, is the main surgical innovation of brain mapping. Cortical mapping highlights positive and negative areas before starting resection. This "negative mapping" strategy represents a paradigm shift in language and no-language mapping techniques by eliminating the neurosurgeon's reliance on positive site control in operative exposure, although validation for other related functions is still needed (Berger, 1994; Black et al, 1987;

Cortical LGGs may be nearly indistinguishable from normal brain tissue. The neuronavigator is therefore essential for precisely demarcating the superficial limits of the tumor. The arachnoid opening is also useful before deciding on the area of resection. The exact distinction acquired with the neuronavigation system may provide immediate visual feedback during the initial subcortical resection when the resection proceeds between

The technique is to start from safe areas and where the vascular supply is expected: following the vessels, skeletonizing and preserving those serving eloquent areas and dividing the others, especially those supplying the tumor, in order to devascularize it. Subcortical pathways perpendicular to positive sites can be followed within the infiltrated brain as far as apparent normal brain is exposed or positive mapping is detected. This is not always possible in language areas, since the closer we come to positive sites, the easier it will be to have intraoperative impairment which will prevent further mapping. Consequently, in language mapping, DTI may be of help in orientating where to map on the tumor periphery in order to stimulate the least possible. DTI may be not considered as more than a imaginary guide. This strategy may be changed in some situations, for instance, with less cooperative patients and during long operations, by adopting the following sequence: mapping, starting resection close to positive sites and following subcortical pathways, and then approach the vascular supply and safe areas. Repeating the mapping subcortically 2-3 times is usually

Starting from safe areas enables us to follow the tumor into the transition zone, from the margins to the infiltrated brain till the normal brain. First sight inspection means starting resection from the tumor periphery before causing possible injury, which can alter tissue color. Here, the microscope is essential for magnifying the region in order to differentiate normal brain from infiltrated brain. This major challenge can be successfully addressed only by experienced neurosurgeons. The corridor measures about 1 cm taken mostly from the

current approach becomes readily understandable: attack the tumor at its margins.

Keles & Berger, 2004; Matz et al., 1999; Sawaya et al., 1999; Toms et al., 1999).

**4.2 The current approach according to mapping information** 

basically relies on visual inspection.

different tissues.

enough, even when anatomical distortion occurs.

infiltrated brain and less from the tumor. Since diffuse tumors are heterogeneous, it's better to see the phenotype of the single variant from the beginning and possibly to stay around the tumor, creating a corridor between the infiltrated brain and the normal brain. Starting from the corridor created around the tumor, additional space is obtained by removing or retracting the tumor and leaving the brain untouched. The rule is: expose the normal brain, map and stay away from it. Working very closely to normal brain tissue implies avoiding any pressure and retraction that could cause brain injury. As a rule, use of the self-retaining spatula is unnecessary and dangerous. This strategy provides us with a better anatomical orientation in a context which is still similar to the anatomo-functional picture as acquired preoperatively with MRI.

Following large, randomized prospective studies in Europe, several countries have already approved, and are now using, 5-ALA for improved quality of resection in HGGs, aimed at assisting the surgeon according to the same resection strategy (Stummer et al., 2006).

The standard for subcortical mapping for language is not yet defined: current amplitude, functional significance of fasciculi, safe distance from fasciculi or tolerance of partial injury, tests to be used, false negative stimulation rate, definition of positive sites, method of validation of positive sites. In particular, continuous monitoring is affected by the learning effect, as described by Ojemann ( Ojemann J.G., et al., 2002).

The limiting factor to total removal is not always the eloquent areas. In large and diffuse tumors we can state that a strategy aimed at exploring and removing safe areas first is a preliminary condition to continue this aggressive strategy in order to obtain maximal resection which justifies risks in eloquent areas. Safe exposure of deep structures is not always feasible; and facing increasing structural and functional uncertainties is preferable to stopping the operation before taking additional risks in eloquent areas, assuming that the degree of cytoreduction has limited value if total removal cannot be achieved (Lacroix et al., 2001).

Vascular injury may be another limiting factor to tumor removal, especially in insular tumors where transit and terminal arteries, normal and pathological arteries, and venous drainage are ever-present challenges during the operation.

#### **4.2.1 Temporo-insular tumors**

The approach to temporal and insular tumors may be trans-cortical or trans-sylvian. When corticectomy is done, the tumor is devascularized by subpial dissection, while trans-sylvian dissection makes it possible to follow the vessels, free the tumor, and then decide which vessels to preserve and which to divide. The lenticulostriate arteries, the parietal trunk of the middle cerebral artery bifurcation on the non-dominant hemisphere and both trunks in the dominant hemisphere and the long perforators running posteriorly are the critical vessels to preserve in order to avoid permanent language and motor deficits. As a rule, vessels smaller than 0.5 mm are pathological. In these tumors, temporal or frontal basal cortical areas are seldom the critical point of resection, and multiple windows can be created between vessels. Accordingly, various sequential options may be weighed in order to start from safe areas and to control arterial vessels (Duffau, 2009; Sanai et al., 2010, Simon et al., 2009).

Mapping techniques allow the neurosurgeon to recognize the internal capsule, which is the critical anatomical point, at least in a non dominant site, while additional information can be obtained from monitoring motor evoked potentials (MEPs) in the course of the operation since insular tumors carry a risk of vascular injury, higher than the risk of direct injury (Figure 1-3). Neuloh and Schramm highlight the semi-quantitative relation between MEP

Surgical Treatment of Supratentorial Glioma in Eloquent Areas 307

Fig. 3. Right insular fibrillary astrocitoma. Trans-sylvian approach was used and tumor removal was achieved under motor monitoring (preoperative on the left side, postoperative

findings and clinical outcome. An amplitude reduction between 25 and 50% increases probability of a postoperative permanent deficit, as well as reversible loss and irreversible loss of MEPs (Neluoh & Schramm, 2004). Typically, MEPs deterioration will allow the neurosurgeon or anesthesiologist to react promptly, with temporary or definitive discontinuation of resection in the critical target area, readjustment or temporary release of retractors if used, application of papaverine to spastic vessels and increased blood pressure.

The well-known corticospinal tract may be difficult to identify in pathological situations where brain distortion alters the normal anatomy. This is the main concern in the surgical approach to fronto-parietal tumors, the supplementary motor area (SMA) being secondary in importance. Image- and neurophysiologic-assisted surgery may be effective as well (Talacchi et al., 2010a). Vascular supply derives from the distal branch of the pericallosal artery and the fronto-parietal distal branches of the middle cerebral artery. Phase reversal, MEPs, cortical and subcortical mapping allow identification of the central sulcus, motor area and descendent pathways, respectively. The closest safe and convenient distance is 5 mA for diffuse tumors and 1-2 mA for circumscribed lesions. This is the site with the best surgical

In this area the great advantage of intraoperative mapping was to render tumors operable which were otherwise not amenable to surgery because of an unpredictable risk of aphasia. The language-mapping technique described in detail by Ojemann remains the mainstay of this procedure and may be combined with motor mapping in the fronto-parietal area, for both the periphery in cortical tumors and for trajectory planning in subcortical tumors

results among eloquent areas (Talacchi et al., 2010a) (Figure 4).

on the right side).

**4.2.2 Rolandic tumors** 

**4.2.3 Peri-sylvian dominant tumors** 

Fig. 1. Pilocytic astrocytoma of the right posterior insula. The tumor was removed via a trans-sylvian approach assisted by intraoperative motor monitoring (preoperative on the left side; postoperative on the right side)

Fig. 2. Left posterior temporo-insular glioblastoma. The tumor was operated on under awake surgery for cortical language mapping and subcortical motor monitoring. Posterior perisylvian cortical sub-cortical approach was used for tumor removal (preoperative on the left side; postoperative on the right side)

Fig. 1. Pilocytic astrocytoma of the right posterior insula. The tumor was removed via a trans-sylvian approach assisted by intraoperative motor monitoring (preoperative on the

Fig. 2. Left posterior temporo-insular glioblastoma. The tumor was operated on under awake surgery for cortical language mapping and subcortical motor monitoring. Posterior perisylvian cortical sub-cortical approach was used for tumor removal (preoperative on the

left side; postoperative on the right side)

left side; postoperative on the right side)

Fig. 3. Right insular fibrillary astrocitoma. Trans-sylvian approach was used and tumor removal was achieved under motor monitoring (preoperative on the left side, postoperative on the right side).

findings and clinical outcome. An amplitude reduction between 25 and 50% increases probability of a postoperative permanent deficit, as well as reversible loss and irreversible loss of MEPs (Neluoh & Schramm, 2004). Typically, MEPs deterioration will allow the neurosurgeon or anesthesiologist to react promptly, with temporary or definitive discontinuation of resection in the critical target area, readjustment or temporary release of retractors if used, application of papaverine to spastic vessels and increased blood pressure.

#### **4.2.2 Rolandic tumors**

The well-known corticospinal tract may be difficult to identify in pathological situations where brain distortion alters the normal anatomy. This is the main concern in the surgical approach to fronto-parietal tumors, the supplementary motor area (SMA) being secondary in importance. Image- and neurophysiologic-assisted surgery may be effective as well (Talacchi et al., 2010a). Vascular supply derives from the distal branch of the pericallosal artery and the fronto-parietal distal branches of the middle cerebral artery. Phase reversal, MEPs, cortical and subcortical mapping allow identification of the central sulcus, motor area and descendent pathways, respectively. The closest safe and convenient distance is 5 mA for diffuse tumors and 1-2 mA for circumscribed lesions. This is the site with the best surgical results among eloquent areas (Talacchi et al., 2010a) (Figure 4).

#### **4.2.3 Peri-sylvian dominant tumors**

In this area the great advantage of intraoperative mapping was to render tumors operable which were otherwise not amenable to surgery because of an unpredictable risk of aphasia. The language-mapping technique described in detail by Ojemann remains the mainstay of this procedure and may be combined with motor mapping in the fronto-parietal area, for both the periphery in cortical tumors and for trajectory planning in subcortical tumors

Surgical Treatment of Supratentorial Glioma in Eloquent Areas 309

Fig. 6. Huge left frontal oligodendroglioma with intraventricular invasion. Notwithstanding intracranial hypertension awake surgery was used in order to perform a safe perisylvian trans-cortical approach (preoperative on the left side; postoperative on the right side).

While the initial assumption was that no electrically identified areas should be removed if postsurgical language complications are to be avoided, it was later increasingly assumed that postsurgical language deficits would not occur only if cortex that did not result in language deficits with electrical stimulation was removed (Sanai et al., 2008). This indirect message is gaining strength, although most studies lack pre- and postoperative global assessment and objective determination of cognitive complications. Furthermore, the original assumption that resection of any essential language areas will result in postoperative aphasia has not yet been definitively confirmed nor has the assumption that sparing positive sites for naming task will result in sparing of other language functions (Hamberger, et al., 2005; (Peraud, et al., 2004; Petrovich Brennan, et al., 2007; Seek, et al.,

Moving from intraoperative naming-assisted surgical resection to other language and cognitive tasks, before relying on new protocols we need a multi-staged system of evidence for the potential and limitations of AS for cognitive mapping, its clinical validity for a single task or battery of tasks and technical standardization (intentionally addressed in the present article). Meanwhile, patient safety must be guaranteed by accurate comparative assessment, which should be discussed and defined (Lacroix et al., 2001; Vives & Piepmeier, 1999).

Microsurgical resection remains a critical therapeutic modality for all gliomas (Black, 1998; Guthrie & Laws, 1990; Keles et al., 1999; Yasargil et al., 2004). For all gliomas, the

**5. Outcome assessment** 

2006; Whittle, et al., 2003, 2005).

**5.1 The value of glioma extent of resection** 

(Ojemann G.A. et al., 1989). Methodological validation for subcortical language mapping is still lacking, and landmarks for presumed significant fascicles are not well defined and often altered by anatomical distortion. In awake patients, there is an approximately 10% risk of finding a positive site within the tumor, which remains the main limiting factor to removal (Ojemann G.A. et al.,1989) (Figure 5-6).

Fig. 4. Left rolandic glioblastoma. The tumor was removed using motor monitoring (preoperative on the left side; postoperative on the right side).

Fig. 5. Left frontal oligodendroglioma in the Broca area. The tumor was removed under awake surgery (preoperative on the left side; postoperative on the right side).

(Ojemann G.A. et al., 1989). Methodological validation for subcortical language mapping is still lacking, and landmarks for presumed significant fascicles are not well defined and often altered by anatomical distortion. In awake patients, there is an approximately 10% risk of finding a positive site within the tumor, which remains the main limiting factor to removal

Fig. 4. Left rolandic glioblastoma. The tumor was removed using motor monitoring

Fig. 5. Left frontal oligodendroglioma in the Broca area. The tumor was removed under

awake surgery (preoperative on the left side; postoperative on the right side).

(preoperative on the left side; postoperative on the right side).

(Ojemann G.A. et al.,1989) (Figure 5-6).

Fig. 6. Huge left frontal oligodendroglioma with intraventricular invasion. Notwithstanding intracranial hypertension awake surgery was used in order to perform a safe perisylvian trans-cortical approach (preoperative on the left side; postoperative on the right side).
