ORIGINAL ARTICLE

Operability of glioblastomas: ‘‘sins of action’’ versus‘‘sins of non-action’’

Paolo Ferroli • Marco Schiariti • Gaetano Finocchiaro • Andrea Salmaggi •

Melina Castiglione • Francesco Acerbi • Giovanni Tringali • Mariangela Farinotti •

Morgan Broggi • Cordella Roberto • Elio Maccagnano • Giovanni Broggi

Received: 20 September 2012 / Accepted: 22 February 2013 / Published online: 12 March 2013

� Springer-Verlag Italia 2013

Abstract Despite prognosis of glioblastomas is still poor,

mounting evidence suggests that more extensive surgical

resections are associated with longer life expectancy.

However, the surgical indications, at present, are far from

uniform and the concept of operability is extremely sur-

geon-dependant. The results of glioblastoma resection in

104 patients operated on between March 2005 and April

2011 were reviewed with the aim to shed some light on the

limits between ‘sins of action’ (operating upon complex

tumors causing a permanent severe deficit) and ‘sins of

non-action’ (considering inoperable tumors that can be

resected with good results). Fifty-five patients (54.4 %)

(Group 1) presented with a ‘disputable’ surgical indication

because of one or more of the following clinico-radiolog-

ical aspects: involvement of motor and language areas

(39.4 %), deep location (7.7 %), corpus callosum infiltra-

tion (13.4 %), or major vessels encasement (8.6 %). Forty-

six (42.5 %) patients (Group 2) presented with an ‘indis-

putable’ surgical indication (readily accessible tumors in

non-eloquent areas). Overall mortality was 2.9 %. The

mean overall survival was 19.8 months and not signifi-

cantly different in the two Groups (20.4 Group 2 and 19.5

months for Group 1; p = 0.7). Patients with GTR and\72

years had a longer survival (p = 0.004 and 0.03, respec-

tively). Seventy patients (69.3 %) showed an uneventful

post-operative course, without statistical significance dif-

ference between Group 1 and 2. The gross total removal of

glioblastoma with many complexities (Group 1) was found

to be feasible with acceptable mortality, morbidity and

long-term survival rates.

Keywords Operability � Glioblastoma � Quality of life �Survival � Neurosurgery

Introduction

Despite prognosis of glioblastomas is still poor, mounting

evidence suggests better outcomes (prolonged post-surgical

life expectancy) after extensive surgical resection [1–4].

For this reason in dedicated brain cancer centers, a

‘‘maximal safe resection’’ of all tumors is the standard of

care. However, patients affected by complex tumors where

surgery poses the threat of inflicting new invalidating

neurological deficits still receive very different opinions

about surgical indication and many surgeons still apply the

label of ‘‘inoperability’’ to some cases. The most common

reasons for that are: multifocality; contralateral extension;

infiltration of supposed eloquent areas; encasement of

major cerebral vessels, and deep location. In our series,

none of these features were considered as an absolute

contraindication and the envelope of surgical indication

was pushed far beyond conventional limits. In this study,

we present the results of glioblastoma resection in 104

P. Ferroli (&) � M. Schiariti � M. Castiglione � F. Acerbi �G. Tringali � M. Broggi � C. Roberto � G. Broggi

Department of Neurosurgery, Fondazione Istituto Neurologico

Carlo Besta, Via Celoria 11, 20133 Milan, Italy

e-mail: paolo.ferroli@istituto-besta.it

G. Finocchiaro � A. Salmaggi

Department of Neuro-oncology, Fondazione Istituto Neurologico

Carlo Besta, Via Celoria 11, 20133 Milan, Italy

M. Farinotti

Department of Neuro-epidemiology, Fondazione Istituto

Neurologico Carlo Besta, Via Celoria 11, 20133 Milan, Italy

E. Maccagnano

Department of Neuro-radiology, Fondazione Istituto

Neurologico Carlo Besta, Via Celoria 11, 20133 Milan, Italy

123

Neurol Sci (2013) 34:2107–2116

DOI 10.1007/s10072-013-1345-5

patients with the specific aim of shining some much light

on the question of ‘limits’ in the critical debate between

‘‘sins of action’’ (i.e. surgical procedures done on the

complex GBM patient that cause severe, permanent neu-

rological deterioration) and ‘‘sins of non-action’’ (i.e. sur-

geries not performed on ‘inoperable’ grounds that could

have resulted in a resection with positive outcomes).

Patients and methods

Between March 2005 and April 2011, 104 patients affected

by GBM were operated upon at the Department of Neu-

rosurgery of the C. Besta Institute of Milan by a single

surgeon (PF) with the philosophy of ‘‘maximal safe

resection’’ in all cases. The reasons to study the work of

this single surgeon, retrospectively, were to avoid:

1. any bias in selecting patients arising from different

attitudes and experiences;

2. any bias related to different surgical techniques.

All demographic and clinical features were obtained

from our hospital records. This retrospective study received

the approval of our Institutional Ethical Committee and the

condition of an ‘‘Informed Consent’’ was obtained in all of

the patients that were studied.

We made no attempt to recognize such factors as: age,

cardiopathies, diabetes, obesity, motor dysfunctions, and

speech disturbances as ‘absolute contraindications’ when it

was obvious that a surgical intervention was necessary to

prolong life, provided that the patient willingly accepted

the risk of possible complications. A Karnofsky score of

less than 60 was also bypassed as a contraindication if it

was related to an acute intracranial hypertension syndrome

that was likely to improve after surgery.

The intensity of motor deficit was rated using the fol-

lowing standardized motor scale: 0, no deficiency; 1, mild

deficiency (patient can use his or her limbs almost nor-

mally—i.e. walking is possible, but there is impairment of

fine movements of the upper limbs); 2, moderate deficiency

(movement is possible with the help of the examiner); and

3, severe deficiency (no spontaneous movement against

gravity) [5]; language disturbances were graded as mild,

moderate or severe according to NIHSS (National Institutes

of Health Stroke Scale) [6]. All patients were selected for

surgery after a complete contrast-enhanced brain MRI

examination. Functional MRI was used for studying

hemispheric dominance, language function and motor

function. DTI was used to track cortico-spinal pathways

and dorsal and ventral language pathways [7]. All data

were available for intraoperative navigation (StealthSta-

tion, Medtronic Inc.). Proton spectroscopy was used to help

making a differential diagnosis, when needed. Angio-RMs

or angio-TCs were acquired in the case of complex vas-

cular anatomical relationship among tumors, veins and

arteries.

The following clinico-radiological aspects were con-

sidered to increase surgical complexity:

1. Contrast enhancement/necrotic component involving

the primary motor cortex (clinically and fMR demon-

strated).

2. Contrast enhancement/necrotic component involving

the cortico-spinal tract (clinically and DTI demon-

strated).

3. Contrast enhancement/necrotic component involving

cortical language areas (clinically and fMR demon-

strated).

4. Contrast enhancement/necrotic component involving

language subcortical pathways (clinically and DTI

demonstrated).

5. Contrast enhancement/necrotic component involving

basal ganglia/thalamus/brainstem.

6. Contrast enhancement/necrotic component with con-

tralateral extension through midline connections.

7. Major cerebral vessels (ICA, ACA, MCA, PCA)

within the contrast-enhanced/necrotic component.

Some significant examples of these complexities are

illustrated in Fig. 1. The pre-operative characteristics of the

patients are summarized in Tables 1 and 2. This study

involved GBMs that were categorized as: primitive

(66.6 %); secondary (14.2 %); and recurrent (19.2 %). The

mean tumor diameter was 47.5 mm (median 47.8 mm;

range 18.1–80.2 mm). Patients were divided into 2 Groups

(Table 1): Group 1 (complex cases) consisted of 58

patients (55.7 %) in which the patients presented with one

or more of the above mentioned clinico-radiological

aspects that were considered to increase surgical com-

plexity. Group 2 consisted of the remaining 46 patients

(44.2 %) with readily accessible tumors in non-eloquent

areas.

Post-operative treatment and imaging

All primitive and secondary GBMs received combined

radio and chemotherapy (RCT) according to the Stupp

protocol (Temozolomide TMZ 75 mg/m2 1 h before stan-

dard focal radiotherapy (60 Gy) followed by additional

monthly cycle of TMZ 150/200 mg/m2 5/28) [8] and

completed the treatments. Fotemustine, procarbazine,

bevacizumab, irinotecan and cisplatin administration were

used as second line treatment for 35 patients.

Extent of resection based on residual tumor volume was

determined in all cases within 24–72 h after surgery by

comparison between preoperative and early post-operative

2108 Neurol Sci (2013) 34:2107–2116

123

T1-weighted contrast-enhanced MRI (within 72 h of sur-

gery) by a dedicated neuroradiologist (E.M.) following a

volumetric analysis. According to the classification system

reported by Berger [9] resection was considered ‘total’ in

cases where there was no residual contrast enhancement;

‘subtotal’ in cases with less than 10 cm3 of tumor residue;

and ‘partial’ in cases with more than 10 cm3 of tumor

residue.

Outcome measures and statistical analysis

Short term surgical outcome and post-operative

complications

The clinical and neurological status of the patient was

evaluated immediately after surgery with the intention of

revealing any possible new undesired neurological prob-

lem. The same grading scales used pre-operatively were

repeated in the immediate post-operative follow-up.

The results were graded as illustrated in Table 3. In

addition, any complication was recorded following Land-

riel-Ibanez classification [10].

Long-term results

Long-term results in terms of quality of life were assessed

by clinical follow-up and periodic phone survey. A neu-

rological examination was carried out. The QoL of these

patients was classified in four decreasing levels according

to Heros classification [11]. The ultimate data collection

and retrospective analysis of outcome were carried out by

an independent reviewer (MC). Survival rate came from

the Institutional Cancer Registry.

Statistical analysis

Epidemiological and clinical data were recorded. The inter-

group morbidity was compared with the Chi square test. Sur-

vival as a function of time was plotted using the Kaplan–Meier

method. The impact on survival rate of the following param-

eters: involvement of thalamus/basal ganglia/brainstem,

involvement of corpus callosum, major brain arteries encase-

ment, presence/absence of motor deficit, presence/absence of

language deficits, KPS score \or C80, tumor size \or C4

cm, age and tumor resection were evaluated. Univariate com-

parison of parameters was performed by log-rank analysis.

Fig. 1 Some significant examples of patients with a disputable surgical indication that were operated upon with good post-operative results

Neurol Sci (2013) 34:2107–2116 2109

123

Table 1 Characteristics of 101

patients [3 patients (2.88 %) in

Group 1 died in the first post-

operative week]: radiological

features and site, presenting

symptoms, pre-operative

examination results

sl slight deficit, mo moderate

deficit, se severe deficit

(according to NIHSS

classification)

Characteristics (no of patients)

No of cases and

radiological features

Presenting symptoms Pre-op exam results

Group 1 (55)

Motor and language areas (41) Seizures (13) Normal (10)

Reduction of ideation (2) Cognitive impairment (5)

Motor circuits involvement (17) Short term memory

impairment (3)

Motor deficits (sl = 4; mo = 12;

se = 2)

Language areas (13) Headache (6)

Both (11) Headache and vomiting (4) Speech deficits (sl = 9; mo = 12;

se = 1)Motor disturbances (6)

Language disturbances (7)

Deep location only (8) Headache (3) Normal (3)

Extrapyramidal symptoms (2) Cognitive impairment (2)

Thalamus (3) Motor disturbances (2) Motor deficits (sl = 0; mo = 2;

se = 3)

Brainstem (1) visual field defect (1)

Vertigo and dizziness (1)

Basal ganglia (4)

Corpus callosum (14) Seizures (2) Normal (5)

Corpus callosum only (4) Headache (4) Cognitive impairment (5)

Motor disturbances (3) Motor deficits (sl = 2; mo = 2;

se = 1)

Speech deficits (sl = 1; mo = 1;

se = 0)

Anterior (4) Language disturbances (1) visual field defect (2)

Median (1) Visual field defect (1)

Posterior (1)

Frontomesial (4) Ataxia (1)

Parietomesial (4)

Behavioral anomalies (2)

Vessels encasement (9) Headache (2) Normal (3)

Vessels encasement only (2) Motor disturbances (3) Motor deficits (sl = 2; mo = 0;

se = 2)

Speech deficits (sl = 3; mo = 2;

se = 1)

Carotid artery (1) Language disturbances (4)

Anterior cerebral artery (1)

Middle cerebral artery (6)

Posterior cerebral artery (1)

Group 2 (46) Seizures (14) Normal (32)

Frontal (19) Headache (13) Motor deficits (sl = 8; mo = 1;

se = 0)

Reduction of ideation (4) Speech deficits (sl = 2; mo = 1;

se = 0)

Temporal (11) Visual field defect (4) Visual field defect (7)

Parietal (11) Behavioral anomalies (2)

Occipital (5)

Ataxia (7)

Vertigo and dizziness (2)

2110 Neurol Sci (2013) 34:2107–2116

123

Results

Peri-operative and post-operative outcomes

Early post-operative mortality

Three patients (2.9 %), all in Group 1, died in the first post-

operative week (Ibanez grade 4, Tables 4, 5). One patient

died because of uncontrolled cerebellar swelling after a

supracerebellar infratentorial approach for the removal of a

thalamic GBM. Post-op CT scan showed a complete

removal with a clean surgical cavity along with an

impressing cerebellar swelling that was probably due to

supracerebellar vein coagulation or thrombosis of the Galen

vein complex. Two patients died of pulmonary embolism

due to deep venous thrombosis (one female, obese patient,

Table 2 Patient demographics

in 104 patients: sex age, KPS,

presence of co-morbidities,

tumor diameter, MGMT (O(6)-

methylguanine-DNA

methyltransferase) status,

recurrence

Variable No. of patients

Group 1 Group 2 Total

Sex

Male 37 26 63

Female 21 20 41

Age

Men (mean) 61 62 61

Female (mean) 57 54 56

Older age ([72 years) 12 6 18

KPS \ 80 20 5 25

Comorbidities

Cardiopathy 12 8 20

Diabetes 4 6 10

Obesity 2 6 8

COPD 2 2 4

Hypertension 8 10 18

Tumor diameter

C4 cm 39 20 59

MGMT status

Presence of methylation 14 4 18

Absence of methylation 17 10 27

Recurrence 22 11 33

No. of patients operated two times by

the same operator (FP)

10 4 14

No. of patients operated on only for the

recurrence by the operator FP

10 9 19

Table 3 Post-operative

clinical result after surgery in

101 patients

Tumor resection and post-operative morbidity Group 1 (55 cases) Group 2 (46 cases)

Good: total removal with clinical and neurological

improvement or absence of new neurological

deficit (60.39 %)

29/55 (52.7 %) 32/46 (69.5 %)

GTR 29 (52.7 %) GTR 32 (69.5 %)

Acceptable: subtotal removal or presence of new

non-invalidating neurological deficit or new

invalidating neurological deficits that improved

within 30 days (31.68 %)

22/55 (40 %) 10/46 (21.7 %)

GTR 4 (7.3 %) GTR 2 (4.3 %)

STR 17 (30.9 %) STR 8 (17.3 %)

PR 1 (1.8 %) PR 0 (0 %)

Poor: partial removal or disability (7.92 %) 4/55 (7.3 %) 4/46 (8.7 %)

GTR 0 (0 %) GTR 4 (8.7 %)

STR 4 (7.3 %) STR 0 (0 %)

PR 0 (0 %) PR 0 (0 %)

Neurol Sci (2013) 34:2107–2116 2111

123

and one male, hemiplegic after surgery) despite adequate

antithrombotic therapy (seleparin 0.4 ml). The early post-

operative mortality in Group 1 did not result in a statistically

significant difference from the one in Group 2 (0 %)

(p = 0.25) because of limited sample sizes. All three of

these patients were excluded from further analysis.

Morbidity

The detail of complications is illustrated in Tables 4 and 5.

Four cases required a reoperation for the treatment of a

surgical complication (2 wound revisions for infection, 1

sub-dural and 1 intracerebral hematoma evacuation). No

Table 4 Complication grades: Ibanez model

Grades Surgical complications Medical complications

Group 1 Group 2 Group 1 Group 2

Ia (9) T new speech disorder (2) Subgaleal CSF collection (1)

P new speech disorder and

new motor deficits (2)

P visual field defect (1)

P visual field defect (1)

T New motor deficits (1)

P new motor deficit (1)

Ib (21) Seizures (4) Seizures (2) Steroid diabetes (3)

DVT (1)

Urinary tract infection (2)

arrhythmia (1) cardiac

ischemia (1)Fever (1)

CSWS (1)

Post-operative brain edema (1)

Wound infection (1)

Thalamic pain (1)

Middle cerebral artery

infarction (1)

CFS infection (1)

IIa (2) Subgaleal CSF accumulation

(lumbar puncture) (1)

Subgaleal CSF collection

(lumbar puncture) (1)

IIb (2) Wound infection (1) Wound infection (1)

IIIa (1) Subdural hematoma (1)

IIIb (4) Intracerebral hematoma (1) DVT and pulmonary

embolism (2)

lung distress (1)

IV (3) Cerebellar swelling (1) DVT and pulmonary

embolism (2)

T transient, P permanent, DVT deep vein thrombosis, CSF cerebrospinal fluid, CSWS cerebral salt wasting syndrome

Table 5 Neurosurgical complication following Landriel-Ibanez classification

Grade I Any non-life-threatening deviation from normal postoperative course, not requiring invasive treatment

Grade Ia Complication requiring no drug treatment

Grade Ib Complication requiring drug treatment

Grade II Complication requiring invasive treatment such as surgical, endoscopic, or endovascular interventions

Grade IIa Complication requiring intervention without general anesthesia

Grade IIb Complication requiring intervention with general anesthesia

Grade III Life-threatening complications requiring management in ICU

Grade IIIa Complication involving single organ failure

Grade IIIb Complication involving multiple organ failure

Grade IV Complication resulting in death

Surgical complications Adverse events that are directly related to surgery or surgical technique

Medical complications Adverse events that are not directly related to surgery or surgical technique

2112 Neurol Sci (2013) 34:2107–2116

123

statistically significant difference was found between

Groups 1 and 2 even though there was a definite trend

toward a higher percentage of new post-operative neuro-

logical deficits in Group 1. Age, sex and co-morbidities

also did not appear to have any effect on short-term sur-

gical outcomes in a statistically significant way. The short

term results of surgery on 101 patients could be graded as

‘good’ in 61 cases (60.4 %), ‘acceptable’ in 32 (31.6 %)

and ‘poor’ in 8 cases (7.9 %) (Table 3). There was no

statistically significant difference between Group 1 and

Group 2, even though there was a trend towards better post-

operative results in Group 2 (p = 0.14).

Survival

The mean overall survival time in 84 patients operated on the

first time by FP was 19.8 months (range 1.7–49.3 months;

median 17.7 months). Of these 84 patients, 16 were re-

operated on a second time for their recurrence (also by FP).

Overall survival rates at 6, 12, 18 and 24 months were 86.9,

65.4, 47.6 and 32 %, respectively. Patients in Group 2

showed a trend towards longer overall survival times (20.4

vs. 19.5 months) although this trend lacked any statistical

significance (p = 0.69) (Fig. 2).

In a univariate analysis after controlling for pre-opera-

tive factors known to be associated with survival, none of

the following parameters were found to be associated with

decreased survival: (1) presence/absence of motor deficits;

(2) presence/absence of language deficits (3) involvement

of thalamus/basal ganglia/brainstem; (4) involvement of

corpus callosum; (5) encasement of major brain arteries;

(6) KPS score \ or C80; (7) tumor size \ or C4 cm; (8)

MGMT status. The only parameter that seemed to be sig-

nificantly associated with increased survival time was the

extent of tumor excision. Patients with a total tumor

removal had a mean survival that was 9 months longer

than those with a partial excision (24 vs. 15.1 months,

p = 0.004). Patient age only showed statistical significance

at a threshold of 72 years (mean survival in patients

younger than 72: 21.2 months vs. 12.7 months in the

Group of patients older than 72, p = 0.03).

Recurrent GBMs

In this series 31 patients underwent surgery for a recurrent

tumor. Of these, 14 received their first surgery by the

author (FP) and 17 were selected for surgery when they

came from other hospitals. Of these, 19 were included in

Group 1 and 12 in Group 2. The mean overall survival time

was 13.6 months after the second surgery (Group 1:

11.8 months; Group 2: 17.5 months; p = 0.4).

The short-term results of surgery in these 31 patients

could be graded as ‘good’ in 12 cases (38.7 %); ‘accept-

able’ in 12 cases (38.7 %); and as ‘poor’ in 7 cases

(22.6 %). There was no statistically significant difference

between Group 1 and Group 2 (p [ 0.05).

Quality of life

Quality of life questionnaires were recorded in 93 out of

the 101 patients/families that were studied. In the Group 1,

in 55 % of cases the long-term results, in terms of quality

of life, were graded as either ‘excellent’ or ‘good’ (Fig. 3).

In Group 2 the percentage was higher (72 %), although it

lacked any statistical significance (p = 0.09) (Fig. 3).

Out of 93 patients, 89.2 % of these patients reported

positive experiences from their surgery. Only 10.8 %

reported that on the basis of their post-surgery status they

would not go through surgery again.

As far as recurrent cases are concerned, out of 31 patients,

most of them (74.2 %) found the second operation useful

and gave a positive opinion after that second surgery.

Fig. 2 Probability of survival in patients in Group 1 versus patients

in Group 2 in relation to the number of months after the first

operation. Rates were calculated with the Kaplan–Meier method. The

average survival in 84 patients was 19.5 (median 14.1 month) and

20.4 (median 19 months) months in Group 1 and in Group 2,

respectively (p = 0.69). Seventeen patients were excluded because

operated on the first time in another hospital. Out of these 84 patients

12 were re-operated a second time for their recurrence by the same

surgeon; 9 patients in Group 1 and 3 patients in Group 2

Neurol Sci (2013) 34:2107–2116 2113

123

Discussion

At present the surgical indications for high grade gliomas

are far from being uniform and are not based upon sound

evidence-based criteria [3, 12–15]. The significant pre-

operative prognostic factors that exist are: age (cut-off at

72 years in this series); Karnofsky score; multifocality and

tumor volume [14, 16–18].

Surgery is obviously one of the weapons that are available

to improve QoL and to slow the fatal evolution of the disease.

This is doubtlessly the reason why patients and their relatives

see surgery as a source of great hope. But aggressive surgery

remains debated and is not always offered in high risk glio-

mas because of number of reasons: poor prognosis, inability

to tolerate long surgical times, poor physiological and

neurophysiological compliance, increased risk of compli-

cations (venous thrombosis and arterial thrombosis of major

cerebral vessels) [14, 16, 17, 19].

Two of the criteria that are used by some surgeons to

define the ‘‘operability’’ of gliomas are the site of the tumor

and the degree of its extension [20–22]. Tumors that we

classified as complex in this series are often labeled as

‘inoperable’ because of the supposed much higher risk of

post-operative complications and invalidating neurological

deficits. The results of our retrospective analysis challenge

this idea and show that most of these patients benefit from a

gross total removal of the tumor (Fig. 1) with, in our

opinion, acceptable levels of mortality and morbidity.

Indeed, in our series, 68 % of the cases in Group 1 had a

post-operative course that revealed no new neurological

deficits. Although the difference between Group 1 and

Group 2 did not result to be statistically significant, the

number of patients with temporary neurological deficits in

the post-operative period was noticeably higher (18 vs. 8

pts) in Group 1 and appeared to be directly related to the

surgical manipulation of eloquent brain. Since the purpose

of surgery is to improve survival without affecting the level

of autonomy, what should be taken into account is not the

presence of temporary deficits but the absence of difference

between the two Groups in terms of permanent neurolog-

ical deficits (7.3 vs. 8.7 %) with loss of autonomy and need

for assistance.

In addition, it should be taken into account a specific

flaw of this study coming from the fact that the long-term

evaluation of the clinical results was influenced not only by

surgery, but also by the effects of radiotherapy, chemo-

therapy and possible disease progression.

As far as the difference in death rate (3 in Group 1 and 0

in Group 2) is concerned, we have to consider that only one

patient died because of the surgical procedure (post-oper-

ative cerebellar swelling and brainstem compression). The

other two patients died following a medical complication

(pulmonary embolism) that can occur following any neu-

rosurgical procedure for GBM removal, regardless of its

complexity.

The survival times of simple cases were found to be not

at all statistically significantly different from the survival

times observed in complex cases (Group 1:19.51 months

vs. Group 2: 20.4 months). This confirms the idea that

tumor removal has the potential to help patients wherever

the tumor is located. We also observed that a gross total

tumor removal was found to help patients regardless of

having an advanced age [18, 19, 23].

The current knowledge today is that among all patients

with GBM, those with unresectable tumors are considered

to have the worst prognosis. The treatment strategy for

these patients is, however, poorly documented in the lit-

erature, in spite of the fact that this subgroup of patients

can represent up to 35–40 % of all GBM patients. The need

to provide optimal treatment to these patients is evident.

We know that surgical techniques have significantly

improved in the last two decades. Kelly observed in 2004

that even though the survival of GBM patients is not rad-

ically altered by surgery, today’s surgical methods are far

less likely to ‘hurt the patient’ [24]. These circumstances

provide evidence that a better definition of surgical results

and overall survival prognostic factors is needed to refine

surgical indication rather than considering tumors simply

Fig. 3 Data of long-term

results obtained in 93 patients/

relatives after phone interview.

a Group 1; b Group 2

2114 Neurol Sci (2013) 34:2107–2116

123

operable or inoperable. Our results would appear to be

completely aligned with Kelly’s opinion. While it is clear

that surgery cannot ‘cure’ glioblastoma because of that

tumor’s intrinsic biological features, a complete removal

does seem to be associated with a longer disease-free

period and enhanced survival [4, 23, 25, 26]. In our view,

therefore, these complex glioblastoma cases are worthy of

every technical effort that promises to remove as much of

the tumor as possible without inflicting any new neuro-

logical deficits that could influence the QoL. There are now

many new strategies that can make this surgical possibility

a reality. These include: intraoperative image guidance;

inside-out tumor removal fluorescence guided surgery; ICG

videoangiography with special attention to the microsur-

gical respect of ‘passing’ arterial vessels; the use of new

hemostatic agents; neurophysiological mapping; and awake

surgery [25, 27–29]. All of these advances offer significant

improvements in helping the surgeon to maximize resec-

tion while minimizing the risks of damage to the healthy

residual brain. And in our estimation they all conspire to

invite surgeons to push the limits of neurosurgery beyond

the boundary conditions that, in the past, were used to

define tumors as ‘inoperable’ [15]. Even if we did not find

statistical significance, is noticeable that the patients in

Group 2 present a better result in terms of QoL and a higher

percentage (55 % in Group 1 vs. 72 % in Group 2) of

patients was able to work full time at the follow-up (grade

excellent and good according to Heros classification), thus

proving that also radio- and chemotherapy influence the

QoL.

Though a larger number of observations would obvi-

ously be much more desirable our findings do seem to be

making an obvious point. Supporting that point is the fact

89.2 % of our patients and their relatives gave a positive

opinion regarding surgery. These opinions, along with the

fact that the results of the entire series compare well with

the results that are currently found in the literature, provide

a rationale to encourage us to continue this investigation of

the legitimate and valid limits of modern surgery [30, 31].

In addition, we believe that these results, in spite of the

intrinsic limits of this analysis (retrospective study, limited

sample with small subgroups, single surgeon in a single

institution, lack of a control group), support the idea that

a patient with a ‘complex tumor’ should be evaluated

in high-volume modern oncologic centers before being

labeled as ‘inoperable’.

Conclusions

On the basis of our findings, we can argue that some

clinical and radiological aspects ‘per se’ do not constitute a

valid or useful reason to label a patient as harboring an

‘inoperable’ tumor. Further prospective studies that are

specifically aimed at a systematic stratification of the risks

of GBM surgery are needed, and hopefully, will be

conducted.

Acknowledgments We thank heartily Dr. Allen Fertziger (Adjunct

Professor, Honors Department, University of Maryland) for assistance

in revising the manuscript.

Conflict of interest We have disclosed potential conflicts of

interest.

References

1. Stummer W, Meinel T, Ewelt C et al (2012) Prospective cohort

study of radiotherapy with concomitant and adjuvant temozolo-

mide chemotherapy for glioblastoma patients with no or minimal

residual enhancing tumor load after surgery. J Neuro oncol (Epub

ahead of print)

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(2):190–198

3. Stummer W, van den Bent MJ, Westphal M (2011) Cytoreductive

surgery of glioblastoma as the key to successful adjuvant thera-

pies: new arguments in an old discussion. Acta Neurochir (Wien)

153(6):1211–1218 (Epub 2011 Apr 9)

4. McGirt 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(1):156–162

5. Cote R, Battista RN, Wolfson C et al (1989) The Canadian

Neurological Scale: validation and reliability assessment. Neu-

rology 39:638–643

6. Brott T, Adams HP Jr, Olinger CP et al (1989) Measurements of

acute cerebral infarction: a clinical examination scale. Stroke

20(7):864–870

7. Bizzi A (2009) Presurgical mapping of verbal language in brain

tumors with functional MR imaging and MR tractography.

Neuroimaging Clin N Am 19(4):573–596

8. Stupp R, Mason WP, Van den Bent MJ et al (2005) Radiotherapy

plus concomitant and adjuvant temozolomide for glioblastoma.

N Engl J Med 352:987–996

9. Berger MS (1994) Lesions in functional (‘‘eloquent’’) cortex and

subcortical white matter. Clin Neurosurg 41:444–463

10. Landriel Ibanez FA, Hem S, Ajler P et al (2011) A new classi-

fication of complications in neurosurgery. World Neurosurg

75(5):709–715 (discussion 604–11)

11. Heros RC, Korosue K, Diebold PM (1990) Surgical excision of

cerebral arteriovenous malformations: late results. Neurosurgery

26:570–578

12. Ashby LS, Ryken TC (2006) Management of malignant glioma:

steady progress with multimodal approaches. Neurosurg Focus

20:E3

13. 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

14. Vuorinen V, Hinkka S, Farkkila M et al (2003) Debulking or

biopsy of malignant glioma in elderly people—a randomised

study. Acta Neurochir (Wien) 145:5–10

15. Ferroli P, Acerbi F, Franzini A (2011) The dawn of the hodotopic

era in neurosurgery: is there a need to upgrade the operability

criteria for brain tumors? World Neurosurgery 75(5–6):571–572

16. Chaichana K, Parker S, Olivi A et al (2010) A proposed classi-

fication system that projects outcomes based on preoperative

Neurol Sci (2013) 34:2107–2116 2115

123

variables for adult patients with glioblastoma multiforme. J Neu-

rosurg 112(5):997–1004

17. Burger PC, Green SB (1987) Patient age, histologic features, and

length of survival in patients with glioblastoma multiforme.

Cancer 59(9):1617–1625

18. Ewelt C, Goeppert M, Rapp M et al (2011) Glioblastoma mul-

tiforme of the elderly: the prognostic effect of resection on sur-

vival. J Neurooncol 103(3):611–618 (Epub 2010 Oct 16)

19. Gulati S, Jakola AS, Johannesen TB, Solheim O (2012) Survival

and treatment patterns of glioblastoma in the elderly: a popula-

tion-based study. World Neurosurg 78(5):518–526

20. Teo C, Broggi M (2010) Surgical outcome of patients considered

to have ‘‘inoperable’’ tumors by specialized pediatric neuro-

oncological multidisciplinary teams. Childs Nerv Syst 26(9):

1219–1225 (Epub 2010 Jun 19)

21. Lange OF, Haase KD, Scheef W (1987) Simultaneous radio- and

chemotherapy of inoperable brain tumours. Radiother Oncol

8(4):309–314

22. Fazeny-Dorner B, Wenzel C, Veitl M et al (2003) Survival and

prognostic factors of patients with unresectable glioblastoma

multiforme. Anticancer Drugs 14(4):305–312

23. Brandes AA, Compostella A, Blatt V et al (2006) Glioblastoma in

the elderly: current and future trends. Crit Rev Oncol Hematol

60(3):256–266 (Epub 2006 Oct 5)

24. Kelly PJ (2004) Technology in the resection of gliomas and the

definition of madness. J Neurosurg 101(2):284–286

25. 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

26. Sanai N, Polley MY, Berger MS et al (2011) An extent of

resection threshold for newly diagnosed glioblastomas. Neuro-

surg 115(1):3–8 (Epub 2011 Mar 18)

27. Ferroli P, Tringali G, Acerbi F et al (2010) Brain surgery in a

stereoscopic virtual reality environment: a single institution’s

experience with 100 cases. Neurosurgery 67(3 Suppl Operative):

ons79–84 (discussion ons84)

28. Ferroli P, Acerbi F, Albanese E et al (2011) Application of

intraoperative indocyanine green angiography for CNS tumors:

results on the first 100 cases. Acta Neurochir Suppl 109:251–257

29. Leclercq D, Duffau H, Delmaire C et al (2010) Comparison of

diffusion tensor imaging tractography of language tracts and intra-

operative subcortical stimulations. J Neurosurg 112(3):503–511

30. Olson JJ, Fadul CE, Brat DJ et al (2009) Management of newly

diagnosed glioblastoma: guidelines development, value and

application. J Neurooncol 93(1):1–23 (Epub 2009 May 9)

31. Olson JJ, Ryken T (2008) Guidelines for the treatment of newly

diagnosed glioblastoma: introduction. J Neurooncol 89(3):255–

258 (Epub 2008 May 23)

2116 Neurol Sci (2013) 34:2107–2116

123