CLINICAL STUDY – PATIENT STUDY

Surgical treatment of high-grade gliomas in motor areas.The impact of different supportive technologies: a 171-patientseries

Andrea Talacchi • Sergio Turazzi • Francesca Locatelli • Francesco Sala •

Alberto Beltramello • Franco Alessandrini • Paolo Manganotti •

Paola Lanteri • Roberta Gambin • Mario Ganau • Vincenzo Tramontano •

Barbara Santini • Massimo Gerosa

Received: 30 October 2009 / Accepted: 13 April 2010 / Published online: 14 May 2010

� Springer Science+Business Media, LLC. 2010

Abstract In the last few years much has been published

to validate new technology in brain mapping for clinical

purposes, but there have been few clinical results. In this

report we describe our five-year experience in the surgical

management of malignant gliomas around motor areas with

an evaluation of the impact of functional magnetic reso-

nance imaging (fMRI) plus navigator and intraoperative

neurophysiology (IN). End-points were extent of removal,

morbidity, and survival. Variables describing patient and

tumor characteristics and treatment modalities were sta-

tistically weighted in relation to treatment outcome. Tumor

depth (P = 0.01), midline shift C1 cm. (P = 0.05), and

insular location (P = 0.001) negatively affected extent of

removal, whereas IN (P \ 0.001) and fMRI plus navigator

(P = 0.02) contributed to increasing the rate of total

removal (73%, 71% vs. 40%). Postoperative motor

impairment was mild and transient in a minority of cases

(20%). General complications, as defined by the Glioma

Outcome Project, occurred in 23% of cases. IN was the

only factor associated with acute postoperative motor

deterioration (P \ 0.001). IN and age [65 years (P =

0.01) were associated with the occurrence of complica-

tions. Overall survival was significantly higher in patients

operated with IN or fMRI plus navigator (P \ 0.01).

Comparing different surgical strategies used in the same

period, we observed that supportive technologies in glioma

surgery have their primary impact on the quality of

resection and survival. IN led to transient motor impair-

ment and some additional complications which did not

affect functional outcome.

Keywords Brain tumors � Surgical treatment �Motor pathways � Functional magnetic resonance �Intraoperative neurophysiology

Introduction

Most clinical series on gliomas seek to establish prognostic

factors, especially for survival [1–4]. Location has seldom

been considered a variable. Few studies have investigated

specific tumor locations, in line with the hypothesis that,

because eloquent areas are a risk factor for higher mor-

bidity, the extent of removal will be necessarily limited

[5–8]. Although what precisely constitutes residual tumor

is still controversial [9, 10], thus restricting the possibility

A. Talacchi � F. Sala � M. Ganau � V. Tramontano �B. Santini � M. Gerosa

Section of Neurosurgery, Department of Neurological Sciences

and Vision, University Hospital, Verona, Italy

S. Turazzi � R. Gambin

Division of Neurosurgery, University Hospital, Verona, Italy

F. Locatelli

Section of Epidemiology and Medical Statistics, Department of

Hygiene and Public Health, University of Verona, Verona, Italy

P. Manganotti

Section of Neurology, Department of Neurological Sciences

and Vision, University Hospital, Verona, Italy

A. Beltramello � F. Alessandrini

Division of Neuroradiology, University Hospital, Verona, Italy

P. Lanteri

Division of Neurology, ‘‘Sacro Cuore’’ Hospital, Negrar,

Verona, Italy

A. Talacchi (&)

Section of Neurosurgery, Department of Neurological Sciences

and Vision, University Hospital, P. Stefani 1, 37126 Verona,

Italy

e-mail: andrea.talacchi@univr.it

123

J Neurooncol (2010) 100:417–426

DOI 10.1007/s11060-010-0193-x

of defining the role of removal, every effort should be

undertaken to differentiate and properly define subgroups

of cases in which prognosis could be significantly

improved if specific surgical treatment were given.

Surgical planning is evolving continuously with advan-

ces in surgical technology (high magnetic field MRI, MR

spectroscopy, IN including awake surgery, intraoperative

MRI, etc.) applied to current treatment modalities [11–14].

And although numerous recent papers have addressed the

validation of these techniques, their impact on clinical

practice is yet to be determined [15–17].

We report our experience with the surgical treatment of

malignant gliomas around supratentorial motor areas. The

objectives of this study were to determine the results of

treatment and the role of the supportive technology—IN

and fMRI-guided surgery. Because the extent of tumor

removal and functional outcome are so closely interrelated,

we considered a pool of outcome variables—extent of

removal, postoperative change in motor grading, compli-

cations, and follow-up—all of which, because of their

reciprocal influence, help to give complete results.

Materials and methods

We retrospectively reviewed all supratentorial malignant

gliomas operated on since January 2002 at the Department

of Neurosurgery, University of Verona, shortly after the

introduction of IN: 171 patients with preoperative motor

impairment or with tumors 1 cm or less from motor path-

ways as determined at MRI. Patients with recurrences or

multifocal tumors were excluded from the study.

Records included clinical history, neurological exami-

nation, MRI, and operative notes in all cases. fMRI was

usually performed as an alternative to intraoperative map-

ping, with or without the aid of frameless stereotactic image-

guided surgery. Preoperative and postoperative MRI studies

were performed within 48 h of the operation in all but 18

cases, in which computed tomography (CT) was carried out

postoperatively with the same timing. The latest technology

for brain surgery, including microscopy and cavitron, were

available. All four surgeons who operated had done more

than 100 operations for gliomas before this series. Surgical

strategy was discussed by the team, but the choice and

application of additional technology to surgical planning was

an individual decision. Follow-up (four-month, functionally

independent survival, and overall survival) was possible for

most patients. Functionally independent survival was the

period of survival with Karnofski score (KPS) C80. All

patients underwent radiotherapy and most (n = 131)

received chemotherapy with temozolamide (TMZ). TMZ

treatment schedules varied as did the number of cycles and

combination with radiotherapy (concomitant or adjuvant).

Patient characteristics

Table 1 lists age, sex, motor grading (MRC scale), symp-

toms and signs, KPS, and symptom duration. Note that our

interpretation of the MRC scale emphasized even slight

modifications in strength. Accordingly, we considered 4?, 4,

and 4- to be single points on the scale, in addition to 1–3 and

5, because of the functional gap between 3 (movements

against gravity) and 5 (normal). Thus, patients who were able

to maintain a static position for a few seconds only were

classified as 4-, those with slow drooping of one arm as 4,

and those with some impairment evident only against resis-

tance as 4?.

Tumor characteristics

Table 1 reports side, size, location, MRI T2-weighted

alterations, midline shift, tumor depth, and histology.

Characteristics of treatment

Retrospectively, three types of supportive treatment modal-

ity were recognized: null (a), fMRI and navigator (b), and

intraoperative neurophysiologic monitoring (c). Each

modality was used alone in 101, 19, and 51 cases, respec-

tively (Table 1) based on single-case evaluation and sur-

geon’s choice. Because of uncertainties about some of the

methodological issues in the intraoperative use of these

techniques, some surgeons do not trust them or prefer only

one of them.

fMRI and navigator

The MR imaging data were acquired with a Siemens 1.5

Tesla and 3.0 Tesla scanner (Siemens, Erlangen, Germany)

and a standard birdcage radiofrequency coil. The functional

images were acquired using gradient echo-planar imaging

(EPI) (TR/TE 3000/30; 36 axial slices; 128 9 128 matrix;

3.2-mm slice thickness with no gap, 220 mm FOV). With

two dumping scans, the total acquisition time for both tasks

(finger tapping and toe flexing contralateral to the affected

side) was 6 min and 36 s. Three-dimensional T1-weighted

anatomic images were acquired with a spoiled gradient recall

sequence (MPRAGE, TR/TE 2300/3.93; 256 9 256 image

matrix; 1-mm slice thickness with no gap; 230 mm FOV).

The fMRI data were analyzed using Brain Voyager QX

1.8 software. Data processing included slice time correc-

tion and 3D motion correction; 3D rigid-body registration

was used to align the reconstructed fMRI data against a

reference image, and spatial smoothing (Gaussian filtering:

full width of half maximum 6 mm) was applied. The vol-

ume of activation of the primary motor cortex was deter-

mined using a cross-correlation technique. A statistical

418 J Neurooncol (2010) 100:417–426

123

Table 1 Characteristics of patients, types of tumors, and treatments by outcome variables

Variables N (%) Outcomes

Removal Motor grading Complications

Subtot. (%) Tot. (%) P-value U/I (%) Worsen. (%) P-value No (%) Yes (%) P-value

Age

\50 64 (38) 28 (49) 29 (51) 52 (83) 11 (17) 55 (86) 9 (14)

50–65 60 (35) 26 (50) 26 (50) 45 (75) 15 (25) 47 (78) 13 (22)

[65 46 (27) 14 (35) 26 (65) NS 36 (82) 8 (18) NS 30 (65) 16 (35) 0.04

Sex

Male 97 (57) 37 (46) 44 (54) 76 (80) 19 (20) 77 (79) 20 (21)

Female 73 (43) 31 (46) 37 (54) NS 57 (79) 15 (21) NS 55 (75) 18 (25) NS

Sympt./signs

ICH 75 (44) 35 (26) 40 (74) 33 (49) 35 (51) 58 (77) 17 (23)

Epilepsy 18 (11) 8 (44) 10 (56) 7 (44) 9 (66) 14 (78) 14 (22)

Dysphasia 41 (24) 27 (70) 14 (30) 22 (63) 13 (37) 34 (83) 7 (17)

Others 36 (21) 15 (42) 21 (58) NS 14 (42) 19 (68) NS 26 (72) 10 (28) NS

Sympt. duration

B6 months 28 (17) 11 (44) 14 (56) 23 (82) 5 (18) 23 (82) 5 (18)

[6 months 138 (83) 55 (45) 66 (55) NS 107 (79) 28 (21) NS 106 (77) 32 (23) NS

Motor grading preop

5 92 (54) 44 (51) 43 (49) 77 (84) 15 (16) 74 (80) 18 (20)

4 63 (37) 20 (38) 32 (62) 46 (74) 16 (26) 49 (78) 14 (22)

B3 14 (9) 3 6 NS 10 3 NS 8 6 NS

KPS

C80 126 (74) 52 (45) 64 (55) 100 (80) 25 (20) 99 (79) 27 (21)

\80 44 (26) 16 (48) 17 (52) NS 33 (79) 9 (21) NS 33 (75) 11 (25) NS

Side

Right 80 (47) 31 (43) 41 (57) 58 (74) 20 (26) 60 (75) 20 (25)

Left 90 (53) 37 (48) 40 (52) NS 75 (84) 14 (16) NS 72 (80) 18 (20) NS

Tumor depth

Cortical 84 (56) 27 (38) 44 (62) 67 (82) 15 (18) 65 (77) 19 (23)

Subcortical 65 (44) 37 (60) 25 (40) 0.01 51 (80) 13 (20) NS 51 (78) 14 (22) NS

Location

Rolandic 93 (55) 15 (19) 62 (81) 68 (75) 23 (25) 74 (80) 19 (20)

Insular 77 (45) 53 (74) 19 (26) <0.001 65 (86) 11 (14) NS 58 (75) 19 (25) NS

Size

\2 cm 32 (23) 13 (48) 14 (52) 22 (73) 8 (27) 25 (78) 7 (22)

3–4 cm 58 (41) 19 (37) 33 (63) 47 (83) 10 (17) 45 (78) 13 (22)

[4 cm 50 (36) 24 (51) 23 (49) NS 40 (80) 10 (20) NS 36 (72) 14 (28) NS

MR T2-w. alteration

BTumor volume 63 (51) 27 (47) 31 (53) 46 (75) 15 (25) 49 (78) 14 (22)

[Tumor volume 61 (49) 24 (44) 31 (56) NS 53 (87) 8 (13) NS 48 (79) 13 (21) NS

Midline shift

\1 cm 82 (66) 30 (40) 45 (60) 65 (81) 15 (19) 66 (80) 16 (20)

C1 cm 43 (34) 23 (59) 16 (41) 0.05 34 (81) 8 (19) NS 33 (77) 10 (23) NS

Histology

Anaplastic oligo 17 (10) 6 (40) 9 (60) 12 (71) 5 (29) 11 (65) 6 (35)

Anaplastic astro 28 (16) 16 (60) 10 (40) 20 (74) 7 (26) 24 (86) 4 (14)

GBM 125 (74) 46 (43) 62 (57) NS 101 (82) 22 (18) NS 97 (78) 28 (22) NS

J Neurooncol (2010) 100:417–426 419

123

threshold of P-value \ 0.5 for each patient was considered

significant.

Anatomic and functional imaging were simply fused,

resliced with DICOM format and then transferred via fast

Ethernet to a navigation data set (StealthStation; Medtronic

Sofamor-Danek, Memphis, TN, USA).

Intraoperative neurophysiology

Neurophysiological techniques for motor-evoked potentials

(MEPs) are described elsewhere [18]. The same monopolar

multipulse stimulation technique used for MEPs monitoring

was also used for cortical and subcortical mapping. In this

case, the same stimulation conditions were used to map the

cortex with a handheld monopolar probe. Subsequently,

especially when approaching the tumor–brain interface,

functional motor mapping was repeated subcortically to

identify and spare the white matter motor tracts. Threshold

intensities up to 20 mA were reached in a stepwise manner,

starting from low thresholds, to obtain indications about

proximity to subcortical motor pathways. As far as we could

distinguish tumor from healthy tissue, we continued to

remove it, even if IN showed we were very closely

approaching the motor pathways (until 1–2 mA); we were

more cautious in cases of infiltrated brain and stopped

resection at 4–5 mA.

Outcome

Extent of tumor removal, morbidity (variation in motor

score and complications), four-month outcome, and sur-

vival were considered. Extent of tumor removal (total

versus subtotal and partial) was determined postoperatively

within 48 h. Total removal ([90%) was defined as the

absence of residual mass or ring-like enhancement at

postoperative MR (T1-weighted images). Removal

between 90% and 70% of the tumor was defined as sub-

total, and less than 70% as partial (Fig. 1).

Variation in motor score was indicated as unchanged,

worsened, or improved, where a change of one point or

more was graded as improvement or worsening (4-, 4, and

4? were regarded as one point each). Mild was defined as a

loss of one point, severe the loss of two or more points.

Complications were defined as systemic, regional, or neu-

rological in accordance with previous studies [19, 20].

Four-month outcome was classified according to chan-

ges versus the preoperative score. At this time, severe

impairment was differentiated from mild impairment if the

patient had lost his/her independence (from KPS [70 to

KPS B70).

Statistical analysis

Crude odds ratio (95% confidence interval (CI) and relative

P-value) was calculated using a univariate logistic regres-

sion model in order to evaluate the effect of treatment

modalities and the other potential prognostic factors on any

single surgical outcome. Subsequently, a multivariate

logistic regression model was used to simultaneously

estimate the effect of treatment modalities on outcomes,

adjusted for sex, age, and the prognostic factors that had

presented a P-value of B0.05 in the previous univariate

analysis. Then, a multinomial logistic regression model

that involved nominal response variables from more than

two categories was fitted to the data in order to estimate the

treatment effect on the different types of complications

(null vs. systemic, local and neurological).

Functionally independent survival and overall survival

Kaplan–Meier curves were estimated after recording the

type of treatment as a two-level variable (null vs. IN or

fMRI plus navigator) using the long-rank test to evaluate

the difference between the two levels [21].

Table 1 continued

Variables N (%) Outcomes

Removal Motor grading Complications

Subtot. (%) Tot. (%) P-value U/I (%) Worsen. (%) P-value No (%) Yes (%) P-value

Treatment

Null 102 (60) 50 (60) 33 (40) 88 (87) 13 (13) 83 (81) 19 (19)

fMR ? navigator 17 (10) 5 (29) 12 (71) 15 (88) 2 (12) 14 (82) 3 (18)

Monitoring 51 (30) 13 (27) 36 (73) <0.001 30 (61) 19 (39) 0.001 35 (69) 16 (31) NS

Removal

Total 81 (54) – – 65 (75) 20 (25) 64 (79) 17 (21)

Subtotal 56 (38) – – 46 (84) 9 (16) 42 (75) 14 (25)

Partial 112 (8) – – 10 (83) 2 (17) NS 9 (75) 3 (25) NS

Statistically significant results are indicated in bold

420 J Neurooncol (2010) 100:417–426

123

Statistical analysis was performed using STATA soft-

ware, release 7.0 (Stata Corporation, College Station, TX,

USA) [22].

Results

Extent of removal

Total removal was achieved in 54% of cases overall. Age,

sex, preoperative motor grading, KPS, symptoms and

signs, symptom duration, side, size, location, tumor depth,

MRI T2-weighted alterations, histology, and treatment

were entered into the univariate analysis in order to find

variables possibly associated with the extent of resection

(total vs. subtotal and partial) (Table 1).

Of all the variables assessed individually in the univariate

analysis, midline shift C1 cm. (P = 0.05), subcortical

position (P = 0.01), and insular location (P \ 0.001) cor-

related with incomplete removal. Treatment with fMRI plus

navigator (P = 0.02) and treatment with IN (P \ 0.001)

were prognostic factors for complete removal (Table 2),

achieved in 71% and 73% of patients, respectively, com-

pared with 40% of those treated without any technological

aids. Some tumor characteristics were negatively associated

with complete removal, but the type of treatment was still

effective in achieving complete removal, making this

decision crucial for oncological outcome.

Multivariate analysis confirmed the association of insular

location (P \ 0.001) and of treatment with IN (P = 0.003)

with the extent of resection (Table 2). This suggested that

tumor characteristics, midline shift, subcortical position,

and insular location were closely correlated, as was further

confirmed by direct comparison (P \ 0.001).

Postoperative motor grading

The preoperative and postoperative motor grading scores

remained unchanged in 61% of patients and improved in

19% (Table 3). Crude odds ratios (95% CI, P-value) were

calculated for the variables listed above, plus extent of

removal, in order to find predictors for worsening, which

Fig. 1 Comparison between

preoperative (a) and

postoperative (b) T1-weighted

enhanced MRI (\48 h.). A

glioblastoma just anterior to the

pre-rolandic gyrus was removed

completely without any

parenchymal alterations

Table 2 Crude and adjusted odds ratios of predictors for total removal

Crude Adjusteda

OR 95% C.I. P-value OR 95% C.I. P-value

Treatment (vs. null)

fMRI ? navigator 3.64 1.17–11.28 0.02 5.19 0.97–30.92 0.07

Monitoring 4.20 1.93–9.08 <0.001 5.75 1.81–18.27 0.003

Shift (vs. \1 cm)

[1 cm 0.46 0.21–1.02 0.05 1.63 0.53–4.96 0.39

Depth (vs. cortical)

Subcortical 0.41 0.21–0.83 0.01 0.74 0.26–2.12 0.57

Location (vs. rolandic)

Insular 0.09 0.04–1.87 <0.001 0.08 0.02–0.25 <0.001

Statistically significant results are indicated in bolda Also adjusted for age and sex

J Neurooncol (2010) 100:417–426 421

123

occurred in 20% of cases (mild in 13%, severe in 7%). IN

was the only variable that was significantly associated with

postoperative worsening (39% vs. 13%; P \ 0.001), which

was confirmed after adjusting for the other variables

(P \ 0.001) (Table 4).

Complications

Table 3 lists the prevalence of complications: 77% of

patients presented without complications. In univariate

analysis, age C65 years (P = 0.01) was associated with

the occurrence of postoperative complications (systemic in

8%, local in 6%, neurological in 9%), whereas no associ-

ation was found with IN (P = 0.08). Multivariate analysis

showed a significant correlation between development of

complications and age C65 (P = 0.01) and IN (P = 0.01)

(Table 5). In an attempt to assign a type of complication to

IN, we found a significant correlation only with neuro-

logical complications (P = 0.03) (Table 6): hemianopsia

(n = 3), dysphasia (n = 2), contralateral hypoesthesia

(n = 2), neglect (n = 1), and acalculia (n = 1).

Follow-up

The course of motor grading scores up to four-month fol-

low-up is summarized in Table 7. In 10/122 patients (8%),

Table 5 Crude and adjusted odds ratios of predictors for post-operative complications

Crude Adjusteda

OR 95% C.I. P-value OR 95% C.I. P-value

Treatment (vs. null)

fMRI ? navigator 0.94 0.24–3.58 0.92 1.45 0.35–6.06 0.61

Monitoring 2.00 0.92–4.33 0.08 7.74 1.18–6.38 0.01

Motor grading preop (vs. 5)

4 1.17 0.53–2.58 0.68 1.00 0.43–2.34 0.99

B3 3.08 0.95–10.0 0.06 2.49 0.69–8.99 0.16

Age (vs. \50)

50–65 1.69 0.66–4.30 0.27 1.90 0.68–5.31 0.22

[65 3.25 1.28–8.26 0.01 3.80 1.36–10.65 0.01

Statistically significant results are indicated in bolda Also adjusted for sex

Table 3 Outcome groups and results

No (%)

Removal

Total 81 54.4

Sub-total 56 37.6

Partial 12 8.0

Motor grading post

Unchanged 102 61.0

Improved 31 18.6

Mildly worsened 22 13.2

Severely worsened 12 7.2

Complications

Null 132 77.6

Systemic 13 7.7

Local 10 5.9

Neurological 15 8.8

Mortality 3 1.5

Table 4 Crude and adjusted odds ratios of predictors for post-operative motor grading worsening

Crude Adjusteda

OR 95% C.I. P-value OR 95% C.I. P-value

Treatment (vs. null)

fMRI ? navigator 0.90 0.18–4.41 0.90 1.05 0.20–5.43 0.96

Monitoring 4.29 1.90–9.72 <0.001 5.13 2.13–12.37 <0.001

Location (vs. rolandic)

Insular 0.50 0.22–1.11 0.08 0.46 0.20–1.09 0.08

Statistically significant results are indicated in bolda Also adjusted for age and sex

422 J Neurooncol (2010) 100:417–426

123

motor grading score worsened and should therefore be

considered to reflect persistent deficits. Of these 10

patients, six had received treatment without any additional

technical aids, one with the aid of fMRI plus navigator, and

three with IN. Of these 10 patients, two presented with

severe impairment, because they changed functional per-

formance status group (from KPS [70 to KPS B70); both

had been operated on without any additional technical aids.

The remaining eight patients had mild deficits which did

not affect their functional status.

Neurological deficits due to complications subsided

within a few weeks in most cases (6/9 cases) and did not

change functional outcome. Later in the follow-up, 15% of

patients underwent reoperation for tumor recurrence.

Globally, 106 (62%) patients were traced for the follow-

up. No difference was found between traced and untraced

subjects with regard to histology, age, KPS, reoperation, or

TMZ treatment, but the rate of total removal differed sig-

nificantly (62% vs. 40%, respectively; P \ 0.01) as did

treatment modality (IN or fMRI plus navigator in 36% vs.

20%, respectively; P \ 0.05)

Overall, the univariate analysis showed that treatment

modality was significantly associated with overall survival

(P \ 0.01) but not with functionally independent survival

(Fig. 2). No further statistical analysis of the currently

available population can be made.

Discussion

With improvements in surgical techniques, nervous and

vascular structures can be safely manipulated and ever

greater proportions of tumor removed. The choice of

approach to subcortical diffuse glioma and the decision to

resect the infiltrated brain surrounding an enhanced mass

are the cornerstones of modern aggressive surgical strategy

and require knowledge of brain functions at the tumor

margin in individual cases [17, 23–25]. In malignant gli-

omas, because of the patient’s short life expectancy, the

cost–benefit balance is even more unfavorable than in any

other tumor. The risk of postoperative neurological dete-

rioration is present for all eloquent areas, but correlative

investigation of anatomy and function in motor areas has

provided the most reliable test to date, opening an oppor-

tunity for informed decisions about the implementation of

surgical treatment and a rationale for further extension of

these techniques to other eloquent areas [13, 26]. Intraop-

erative neurophysiological mapping and monitoring is the

gold standard for this purpose, and fMRI with the aid of

frameless stereotactic image-guided devices is almost as

reliable, especially for superficial lesions [11, 14, 27–29].

These techniques can have additional and unexplored value

when combined with techniques which improve the intra-

operative definition of tumor margins (5-ALA, infrared

spectroscopic imaging, etc.) [30, 31].

Extent of removal

In our experience, some tumor characteristics (insular

location, midline shift C1 cm and subcortical position) still

limit the extent of removal, but IN, and fMRI plus navi-

gator to a lesser extent, have an important impact on the

extent of total removal. Insular gliomas have infiltrative

behavior and intimate cohabitation with brain vascular

supply, making intraoperative mapping only a minor part

of a strategy aimed at safe removal. Similarly, deep tumors

may present additional problems: lack of functional infor-

mation about non-motor areas, paucity of anatomical

Table 7 Time course of motor grading in the postoperative period

Motor grading pre-op Motor grading post-op Motor grading 4 months

= : ; = : ;

5 53 (82%) 0 12(18%) 62 (95%) 0 3 (5%)

4 18 (38%) 14 (30%) 15(32%) 12 (25%) 30 (64%) 5 (11%)

B3 6 2 2 4 4 2

Total 77 (63%) 16 (13%) 29 (24%) 78 (64%) 34 (28%) 10a (8%)

a Treatment null n = 6 (10%); treatment fMRI ? navigator n = 1 (11%); treatment monitoring n = 3 (10%)

: Improved, ; worsened, = unchanged, compared with pre-op

Table 6 The effect of treatment, univariate multinomial odds ratios

(mOR), on type of complications

Complications Treatment mOR 95% CI P-value

Systemic fMRI ? navigator NE* NE* NE*

Monitoring 1.05 0.30–3.65 0.93

Local fMRI ? navigator 1.13 0.12–10.40 0.91

Monitoring 1.34 0.30–5.90 0.70

Neurological fMRI ? navigator 2.37 0.42–13.44 0.33

Monitoring 3.80 1.16–12.41 0.03

* No subjects with systemic complications in treatment with

fMRI ? navigator

Statistically significant results are indicated in bold

J Neurooncol (2010) 100:417–426 423

123

landmarks, and risk of overlooking hidden tumor portions.

Overall, these factors represent a limitation of these tech-

niques as regards extent of tumor removal [32–34].

Intraoperative neurophysiology

The first clinical paper on IN applied to glioma surgery

appeared in 1996, when Skirboll et al. described the

functional brain within the periphery of the tumor, dem-

onstrating how difficult surgical strategy can be for these

tumors. Low radical resection (43%) and high postopera-

tive deterioration (68%) marked early experience with IN

applied to glioma surgery [35]. At the time, although

Berger et al. [36, 37] first presented technical refinements

in the late 1980s, there were no clinical data about surgery

around motor areas. Nevertheless, in the same year, Sch-

ramm’s group reported the first clinical series using a dif-

ferent, modern technique which yielded encouraging

results in terms of reliability and outcome (17% of per-

manent deficits) [15]. Several later papers discussed the

intraoperative criteria that correlated with postoperative

conditions and with variable results (see ‘‘Morbidity’’).

In 2003 Reithmeier et al. [7] first demonstrated a clear

advantage of IN for increasing the rate of radical resection

by comparing two series of patients who had been operated

on at the same institution and for similar lesions (in part

gliomas). In 2005 Duffau et al., focusing on low-grade

gliomas, found a significant increase in the quality of

resection when IN was used, plus a significant increase in

surgical indications. Our series is the first to report on

malignant gliomas: present findings on the use of IN are the

most striking in this study, confirming a real advantage of

this technique (complete removal in 73% vs. 40% of cases)

[29].

fMRI and navigator

This is the first time that use of MRI has been found to

correlate with extent of resection and with favorable results

(complete removal in 71% vs. 40%). This is the conse-

quence of evidence of the reliability of fMRI in the pre-

operative planning of surgery around motor areas [14, 17,

27].

Morbidity

The Glioma Outcome Project established common criteria

for evaluation of clinical studies [3, 19]. In our series, the

overall complication rate was 23%, with less than 10% for

each type of complication (local, systemic, neurological).

Motor impairment was distinguished from neurological

complications for analytical purposes. Of note is that the

overall complication rate was consistent with the better

results of published studies, whatever the operated area

[16, 19, 20, 38]. At four-month follow-up, the percentage

of patients with motor impairment decreased from 20% to

8%; motor impairment was severe in only the 2% of

patients who had not received treatment with monitoring.

In our experience, IN was the only factor associated with

postoperative motor worsening; together with age

C65 years, this variable was associated with the occur-

rence of overall complications.

Intraoperative neurophysiology

There is ample evidence that the effect of IN on morbidity

is linked to its ability to prevent permanent deficits rather

than to predict early outcome, yet some issues remain open.

In the white matter, the effective advantage of MEPs

monitoring over mapping is controversial. While mapping

intensity is likely to correlate with distance from the cor-

ticospinal tract, monitoring of evoked potentials does not

produce exact correlations between neurophysiological

alterations and the presence and degree of neurological

deterioration. The amplitude criteria used in brain surgery

differ from the ‘‘all or nothing’’ criteria used in spinal cord

surgery, where only potential loss indicates a significant

Fig. 2 Estimated Kaplan–Meier plot of functionally independent

survival (a) and overall survival (b). Both were better in patients who

underwent surgical removal with the use of functional mapping aids

than in those who did not. Of the two, overall survival reached

statistical significance (P \ 0.01)

424 J Neurooncol (2010) 100:417–426

123

new permanent paresis. This difference suggests a more

complex pathophysiology of injury to the motor system

during brain surgery as compared with spinal cord surgery,

for which there is no a full explanation [28].

Nevertheless, continuous MEPs monitoring remains a

valid indicator of motor pathway function and can detect

ischemic derangements which would be likely to go

unrecognized with mapping techniques alone [39]. In ret-

rospective studies, patients who showed intraoperative

deterioration of muscle MEPs (e.g. amplitude decreased by

50% or more and/or transient loss of potentials) presented

with postoperative motor deficits [26, 28]. Other authors,

using the bipolar 60-Hz technique, showed a correlation

between positive subcortical mapping and a higher risk of

postoperative motor deficit. In all these series, most motor

deficits were transient and patients recovered within

3–4 months. With either modality, persistent deficits were

rare but remained in around 10% of cases in all clinical

series [5, 40].

Follow-up

The reversibility of most postoperative motor deficits at

follow-up is critical, because it suggests that better tumor

resection can be achieved without leaving the patients with

permanent neurological deterioration. On the other hand,

contrary to the often-repeated claim that IN prevents per-

manent impairment, we did not find a significant difference

between the functional outcomes in patients operated on

using IN, fMRI plus navigator, or nothing, at early follow-

up. At least in this preliminary analysis, these results were

confirmed at long-term follow-up: there was no difference

in functionally independent survival, whereas overall sur-

vival was significantly better for the patients operated on

with functional mapping aids. This finding compares

favorably with prospective studies which showed that

survival is prolonged by improving the quality of resection

[41, 42]. Moreover, these results may be further improved

when surgical treatment is combined with traditional and

new chemotherapeutic agents (i.e. signal transduction

inhibitors) [43].

Conclusions

Although our study has the limitations of any retrospective

study, some important findings emerged about malignant

gliomas, which have never been studied using uniform

criteria in large clinical series in modern intraoperative

settings. While our perioperative data are comprehensive,

including statistical analysis, the number of patients

decreased over the follow-up period, preventing further

analysis. Outcome analysis showed that both IN and fMRI

plus navigator had an important effect on the surgical

treatment of high-grade gliomas in motor areas. Both these

techniques significantly improved the quality of resection,

which is the main reason why these techniques are rec-

ommended. IN was associated with a higher risk of mild

transient postoperative motor deficits and other types of

neurological deficits, although their incidence and severity

compared favorably with those reported in other published

series. Further prospective studies are needed to confirm

the effect of these techniques, in combination with che-

motherapeutic agents, on survival.

References

1. Berger MS, Deliganis AV, Dobbins J et al (1994) The effect of

extent of resection on recurrence in patients with low grade

cerebral hemisphere gliomas. Cancer 74:1784–1791

2. Lacroix M, Abi-Said D, Fourney DR et al (2001) A multivariate

analysis of 416 patients with glioblastoma multiforme: prognosis,

extent of resection, and survival. J Neurosurg 95:190–198

3. Laws ER, Parney IF, Huang W, The Gliomas Outcome Investi-

gators et al (2003) Survival following surgery and prognostic

factors for recently diagnosed malignant glioma: data from the

glioma outcomes project. J Neurosurg 99:467–473

4. Quingley MR, Maroon JC (1991) The relationship between sur-

vival and the extent of the resection in patients with supratentorial

malignant gliomas. Neurosurgery 29(3):385–389

5. Duffau H, Capelle L, Denvil D et al (2003) Usefulness of intra-

operative electrical subcortical mapping during surgery for low-

grade gliomas located within eloquent brain regions: functional

result in a consecutive series of 103 patients. J Neurosurg

98:764–778

6. Ebeling U, Schmid UD, Ying H et al (1992) Safe surgery of

lesions near the motor cortex using intra-operative mapping

techniques: a report on 50 patients. Acta Neurochir 119:23–28

7. Reithmeier T, Krammer M, Gumprecth H et al (2003) Neuro-

navigation combined with electrophysiological monitoring for

surgery of lesions in eloquent brain areas in 42 cases: a retro-

spective comparison of the neurological outcome and the quality

of resection with a control group with a similar lesions. Minim

Invasive Neurosurg 46:65–71

8. Russell SM, Elliott R, Forshaw D et al (2005) Resection of parietal

lobe gliomas: incidence and evolution of neurological deficits in 28

consecutive patients correlated to the location and morphological

characteristics of the tumor. J Neurosurg 103:1010–1017

9. Dumas-Duport C, Monsaingeon V, Szikla G (1982) Serial ste-

reotactic biopsies: a double histological code of gliomas

according to malignancy and 3-D configuration, as an aid to

therapeutic decision and assessment of results. Appl Neuro-

physiol 45:431–437

10. Kelly PJ, Daumas-Duport C, Scheithauer BW et al (1987) Ste-

reotactic histological correlations of computed tomography- and

magnetic resonance imaging-defined abnormalities inpatients

with glial neoplasms. Mayo Clin Proc 62:450–459

11. Gumprecht H, Ebel GK, Auer DP et al (2002) Neuronavigation

and functional MRI for surgery in patients with lesion in eloquent

brain areas. Minim Invasive Neurosurg 45:151–153

12. Kamada K, Todo T, Masutani J et al (2005) Combined use of

tractography-integrated functional neuronavigation and direct

fiber stimulation. J Neurosurg 102:664–672

J Neurooncol (2010) 100:417–426 425

123

13. King RB, Schell GR (1987) Cortical localization and monitoring

during cerebral operations. J Neurosurg 67:210–219

14. Roessler K, Donat M, Lanzerberger R et al (2005) Evaluation of

preoperative high magnetic field motor functional MRI (3 Tesla)

in Glioma patients by navigated electrocortical stimulation and

postoperative outcome. J Neurol Neurosurg Psychiatry 76:1152–

1157

15. Cedzich C, Taniguchi M, Schafer S et al (1996) Somatosensory

evoked potential phase reversal and direct motor cortex stimu-

lation during surgery in and around the central region. Neuro-

surgery 38(5):962–970

16. Fadul C, Wood J, Thaler H et al (1988) Morbidity and mortality

of craniotomy for excision of supratentorial gliomas. Neurology

38:1374–1379

17. Lehericy S, Duffau H, Cornu P et al (2000) Correspondence

between functional magnetic resonance imaging somatotopic and

individual brain anatomy of the central region: comparison with

intraoperative stimulation in patients with brain tumors. J Neu-

rosurg 92:589–598

18. Sala F, Lanteri P (2003) Brain surgery in motor areas: the

invaluable assistance of intraoperative neurophysiological moni-

toring. J Neurosurg Sci 47:79–88

19. Chang SM, Parney IF, McDermott M, The Gliomas Outcome

Investigators et al (2003) Perioperative complications and neu-

rological outcomes of first and second craniotomies among

patients enrolled in the glioma outcome project. J Neurosurg

98:1175–1181

20. Sawaya R, Hammoud M, Schoppa D et al (1998) Neurological

outcomes in a modern series of 400 craniotomies for treatment of

parenchymal tumors. Neurosurgery 42:1044–1056

21. Kaplan EL, Meier P (1958) Non-parametric estimation from

incomplete observations. J Am Stat Assoc 53:457–481

22. McCullagh P, Nelder JA (1989) Generalized linear models.

Chapman and Hall, London

23. Black PM, Ronner SF (1987) Cortical mapping for defining the

limits of tumor resection. Neurosurgery 20(6):914–919

24. Keles GE, Berger MS (2004) Advances in neurosurgical tech-

nique in the current management of brain tumors. Semin Oncol

31:659–665

25. Meyer FB, Bates LM, Goerss SJ et al (2001) Awake craniotomy

for aggressive resection of primary gliomas located in eloquent

brain. Mayo Clin Proc 76:677–687

26. Kombos T, Suess O, Ciklatekerlio O et al (2001) Monitoring of

intraoperative motor evoked potentials to increase the safety of

surgery in and around the motor cortex. J Neurosurg 95:608–614

27. Krishnan R, Raabe A, Hattingen E et al (2004) Functional

magnetic resonance imaging- integrated Neuronavigation: cor-

relation between lesion-to-motor cortex distance and outcome.

Neurosurgery 55(4):904–915

28. Neuloh G, Pechstein U, Cedzich C et al (2004) Motor evoked

potential monitoring with supratentorial surgery. Neurosurgery

54:1061–1072

29. Duffau H, Lopes M, Arthuis F et al (2005) Contribution of

intraoperative electrical stimulation in surgery of low gliomas: a

comparative study between two series without (1985–96) and

with (1996–2003) functional mapping in the same institution. J

Neurol Neurosurg Psychiatry 76:845–851

30. Sobottka SB, Geiger KD, Salzer R et al (2009) Suitability of

infrared spectroscopic imaging as an intraoperative tool in cere-

bral glioma surgery. Anal Bioanal Chem 393:187–195

31. Stummer W, Pichlmeier U, Meinel T et al (2006) Fluorescence-

guided surgery with 5-aminolevulinic acid for resection of

malignant glioma: a randomised controlled multicentre phase III

trial. Lancet Oncol 7:392–401

32. Jeremic B, Milicic B, Gruijcic D et al (2004) Clinical prognostic

factors in patients with malignant glioma treated with combined

modality approach. Am J Clin Oncol 27:195–204

33. Devaux BC, O’Fallon JR, Kelly PJ (1993) Resection, biopsy, and

survival in malignant glial neoplasms. A retrospective study of

clinical parameters, therapy, and outcome. J Neurosurg 78:767–

775

34. Vecht CJ, Avezaat CJJ, van Putten WLJ et al (1990) The influ-

ence of the extent of surgery on the neurological function and

survival in malignant glioma. A retrospective analysis in 243.

J Neurol Neurosurg Psychiatry 53:466–471

35. Skirboll SS, Ojemann GA, Berger MS et al (1996) Functional

cortex and subcortical white matter located within gliomas.

Neurosurgery 38(4):678–685

36. Berger M (1995) Functional mapping-guided resection of low-

grade gliomas. Clin Neurosurg 42:437–456

37. Berger MS, Ojemann GA (1991) Intraoperative brain mapping

techniques in neuro-oncology. Epilepsy 58:153–161

38. Brell M, Ibanez J, Caral L et al (2000) Factors influencing

complications of intra-axial brain tumours. Acta Neurochir 142:

739–750

39. Neuloh G, Pechstein U, Schramm J (2007) Motor tract moni-

toring during insular glioma surgery. J Neurosurg 106:582–592

40. Keles GE, Lundin DA, Lamborn KR et al (2004) Intraoperative

subcortical stimulation mapping for hemispherical periolandic

glioma located within or adjacent to the descending motor

pathways: evaluation of morbidity and assessment of functional

outcome in 294 patients. J Neurosurg 100:369–375

41. Mc Girt MJ, Chaichana KL, Gathinji M et al (2009) Independent

association of extent of resection with survival in patients with

malignant brain astrocytoma. J Neurosurg 110:156–162

42. Stummer W, Reulen HJ, Meinel T et al (2008) Extent of resection

and survival in glioblastoma multiforme: identification of and

adjustment for bias. Neurosurgery 62:564–576

43. Van den Bent MJ, Hegi ME, Stupp R (2006) Recent developments

in the use of chemotherapy in brain tumors. Eur J Cancer 42:

582–588

426 J Neurooncol (2010) 100:417–426

123