Risk factors for glioblastoma therapy associated complications

Clinical Neurology and Neurosurgery 134 (2015) 55–59

Contents lists available at ScienceDirect

Clinical Neurology and Neurosurgery journal homepage: www.elsevier.com/locate/clineuro

Risk factors for glioblastoma therapy associated complications Genevieve Ening a,b,∗ , Fransiska Osterheld b , David Capper c,d , Kirsten Schmieder a,b , Christopher Brenke a,b a Department of Neurosurgery, Knappschafts-Krankenhaus Bochum-Langendreer, Ruhr-University of Bochum, In der Schornau 23–25, 44892 Bochum, Germany b Department of Neurosurgery, University Medical Center Mannheim, University of Heidelberg, Theodor-Kutzer-Ufer 1, 68167 Mannheim, Germany c Department of Neuropathology, Institute of Pathology, University Hospital Heidelberg, Im Neuenheimer Feld 224, 69120 Heidelberg, Germany d Clinical Cooperation Unit Neuropathology, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 224, 69120 Heidelberg, Germany

a r t i c l e

i n f o

Article history: Received 24 November 2014 Received in revised form 24 December 2014 Accepted 3 January 2015 Available online 9 January 2015 Keywords: Glioblastoma Survival Neurological Medical Surgical complications

a b s t r a c t Objective: Thromboembolic events, seizures, neurologic symptoms and adverse effects from corticosteroids and chemotherapies are frequent clinical complications seen in Glioblastoma (GB) patients. The exact impact these have on dismal patient outcome has not been fully elucidated. We aimed at assessing treatment associated complications, evaluating the impact on survival and defining risk factors. Methods: Two hundred and thirty three consecutive adult patients operated on for newly diagnosed GB at a single tertiary institution over a 5-year-period (2006–2011) were assessed. Demographic parameters (age, gender, comorbidity status quantified by the Charlson-comorbidity-index (CCI), functional status computed by the Karnofsky Performance Scale (KPS), tumor characteristics (size, location, IDH-1 mutation and MGMT-Promotor-methylation-status) and treatment parameters (volumetrically quantified extent of resection and adjuvant therapy) were retrospectively reviewed. Complications assessed were recorded as neurological (N), surgical (S) and medical (M). Independent risk factor analysis was performed by the univariate and multivariate logistic regression method. Survival analysis was plotted by the Kaplan–Meier-method, influence of complication occurrence was evaluated by the log-rank test. Results: One hundred and fifty nine (68.2%) patients had a total of 281 complications (90 N, 174 M and 17 S). Univariate analysis identified age (P = 0.003), KPS < 70 (P = 0.002), CCI > 3 (P = 0.03), eloquent tumor location (P = 0.001) and therapy other than the standard radio-chemotherapy with temozolomide therapy (P = 0.034) as risk factors for complications. Multivariate analysis extracted the eloquent tumor location (P = 0.007, odds ratio 1.94) as a significant predictor for complications. Having a complication significantly decreased patient survival (P = 0.015). Conclusions: Complications significantly decrease GB patient survival. Age, poor functional status, other than standard adjuvant therapy and eloquent tumor location proved as significant risk factors for encountering a therapy associated complication. Not extensive surgery or tumor size but surgery at eloquent locations impacts complication occurrence the strongest with a 2 fold increased complication occurrence risk. © 2015 Published by Elsevier B.V.

1. Introduction Glioblastoma (GB) constitutes the most malignant subtype of primary brain tumors and has a dismal 5-year patient survival rate of less than 5% (CBTRUS http://www.cbtrus.org). Despite the

∗ Corresponding author at: Department of Neurosurgery, Ruhr-University Bochum, In der Schornau 23–25, D-44892 Bochum, Germany. Tel.: +49 23429980255; fax: +49 2342993609. E-mail addresses: [email protected], [email protected] (G. Ening). http://dx.doi.org/10.1016/j.clineuro.2015.01.006 0303-8467/© 2015 Published by Elsevier B.V.

advances in diagnostics and multimodal therapeutic approaches instigated in the last decade, individual patient survival remains heterogeneous [1]. Although age and KPS are well-defined outcome prognosticators, further parameters with significant impact on GB disease treatment still remain to be described [2]. Current standard treatment comprises a combination of surgery (resection or biopsy), radiation and chemotherapy [2]. Radiation and chemotherapy are therapeutic modalities with significant proof of influencing survival [3]. The exact impact of surgery on outcome remains indistinct [4]. Current reviews state that patients with perioperative complications and surgically acquired deficits were less likely to receive adjuvant therapy, again emphasizing

56

G. Ening et al. / Clinical Neurology and Neurosurgery 134 (2015) 55–59

the unclear effect of mere cytoreductive debulking on outcome [4]. However, whether the poor GB patient outcome was partly an attribute of surgical complications or effect of natural disease course remains undefined [4]. Hence in order to differentiate between effects of therapy associated complications and ineffective therapy modalities, the former have to be better defined. Complications reported to be frequently associated with the treatment of GB patients include pulmonary infections, deep venous thrombosis, sepsis [5] cardiovascular disease [6] and seizures [7]. With this retrospective single-center-study, we aimed at assessing GB patient treatment associated complications, defining risk factors and evaluating the impact on patient outcome. 2. Methods 2.1. Patient selection Adult patients diagnosed of GB at a single academic tertiary-care institution (University Hospital Mannheim, Germany) between 2006 and 2011 were retrospectively identified by review of the neurosurgical department’s database. A total of 252 cases were retrieved. Diagnosis was determined by a senior neuropathologist according to the WHO classification of central nervous system tumors [8,9]. Molecular analysis was performed in the same neuropathological department. Patients with progression from previously diagnosed low-grade glioma, having undergone prior resection, previous adjuvant therapy (chemotherapy (CT) or radiation therapy (RT)) and/or needle biopsies were excluded to avoid confounding variables related to prior surgery and treatment. Patients with incomplete medical records deficient of data on clinical presentation, pre- and postoperative imaging (neither computer tomography nor magnetic resonance images (MRI) and/or adjuvant therapies were also excluded aiming at generating a uniform patient population. The final study cohort comprised 233 patients. Institutional review board approval was obtained for all aspects of the study and all patients consented to have their medical records accessible for the study. 2.2. Patient, tumor and clinical characteristics Demographic parameters were categorized as age, gender and date of surgery; marking the date of disease diagnosis. Based on an independent blinded to outcome notes evaluation (notes were obtained during the last preoperative clinical visit) during chart review and assessment, preoperative comorbidity was indexed by the Charlson-comorbidity index (CCI) as described elsewhere [10,11] and further dichotomized as being >3 or ≤3. The preoperative functional status was classified by the KPS-index and dichotomized as >70% or ≤70%. Contrast enhanced magnetic resonance images (MRIs) were obtained preoperatively, <48 h after resection and during follow-up. Two independent reviewers blinded to patient outcomes carried out assessments. The characteristics recorded were defined as described previously [12]. Multifocal lesions were defined as multiple separate enhancing lesions with or without connection. Eloquent brain was stated as the motor, sensory, speech, or visual cortex, the basal ganglia/internal capsule, the thalamus, or the brainstem. Volumetric analysis including the contrast-enhancing preoperative tumor volume (CE-PTV) and the extent of resection (EOR) were assessed as described by other publications [13], accordingly gross total resection (GTR), subtotal resection (STR) and partial resection (PR) was defined as 100%, >95% < 100% or <95% tumor resection, respectively. DNA was extracted by standard methods either from formalinfixed paraffin-embedded tumor tissue or from unfixed frozen tumor tissue samples. O6-methylguanine-DNA methyltransferase

(MGMT) promoter methylation analysis was carried out by methylation-specific PCR. Isocitrate dehydrogenase (IDH)-1 codon 132 was analyzed by direct sequencing. In case of ambiguous results, the IDH-1 sequences were amplified by a different set of primers [14,15]. 2.3. Treatment, complication assessment and follow-up Treatment comprised surgical resection (GTR, STR or PR) followed by various therapeutic regimes. Either standard concomitant chemo-radiotherapy followed by adjuvant or maintenance chemotherapy with temozolomide (TMZ/RT → TMZ) or other adjuvant therapy modalities; non-concomitant TMZ therapy, CT or sole RT were also recorded. Complications recorded were divided into three groups as suggested by the Glioma Outcome Project [16]: neurological (N), regional/surgical (S) and systemic/medical (M) complications. This classification was previously defined for patients undergoing craniotomy. Patients were followed up clinically and by MRI every 3 months. Survival data were collected from patients visits to the clinic or during the phone interview with patients and/or their relatives. Patients who were still alive at last follow-up were considered as a censored event in analysis. Primary study endpoint was the overall survival (OS), this was calculated from the date of surgery. In cases in which death could not be confirmed by any means, the patients were classified as lost to follow-up at the time of their last clinic visit. Complications or death occurring within 30 days after surgery was defined as a perioperative morbidity or mortality, respectively. 2.4. Statistical analysis Statistical analyses were performed with SPSS version 22.0 (SPSS Inc., Chicago, Illinois, USA). Descriptive data was presented as means ± SDs for parametric data and medians with ranges for nonparametric data. Intergroup frequencies were compared using the Fisher exact test/Mann–Whitney U-test for categorical variables and the Student t-test for continuous variables. Two-sided P values less than 0.05 were considered statistically significant. Factors independently associated with occurrence of complications were assessed by univariate analysis. The Student t-test was used for continuous variables and the chi-square for categorical variables. Independent variables proving significant here served as covariates for multivariate logistic regression analysis. The Kaplan–Meiermethod and log-rank analysis was used to assess the impact of complications on outcome. 3. Results 3.1. Complication assessment Table 1 summarizes all complications reviewed. 68% (159) of patients experienced a sum of 281 complications with 119 patients suffering 174 medical, 77 patients 90 neurological and 17 patients 17 surgical complications. Whereas urogenital and gastrointestinal problems with 17% and 16%, respectively, topped the list, meningitis and dermatological complications with 2% each bottomed the list of medical complications. The most frequently encountered neurological complications were seizures with more than 37 (41%) cases. In contrast, signs of raised intracranial pressure defined as nausea/vomiting/papilledema were least frequently observed and registered only in one case (1%). Surgical complications were very rare in our study cohort. The 17 cases comprised of 10 wound infections and 7 intracranial bleedings. Analysis of the time of complication occurrence was available for 159 patients and 213 cases. 55 cases were perioperative (<30 days

G. Ening et al. / Clinical Neurology and Neurosurgery 134 (2015) 55–59 Table 1 Summary of 281 complications recorded. (No.[%]) Neurological Headache Motor deficit Seizures Speech deficit Cranial nerve palsy Altered level of consciousness Psychotic disorders Raised cranial pressure signs (nausea/vomiting/papilledema) Total Medical Cardiac Pulmonal Gastointestinal Urogentital Hematological Vascular Systemic infection Meningitis Endocrinological Reduced general condition Dermatological Total Surgical Wound infections Intracranial bleeding Total

3 (3) 21 (23) 37 (41) 6 (7) 7 (8) 6 (7) 9 (10) 1 (1) 90 24 (14) 24 (14) 28 (16) 29 (17) 26 (15) 12 (7) 9 (5) 3 (2) 5 (3) 11 (6) 3 (2) 174 10 (59) 7 (41) 17

after surgery), 85 occurred after this time scope and 19 cases were recorded at both < and >30 days after surgery. Perioperative mortality rate of our study group was 8% (18 patients). These patients encountered a sum of 21 complications (18 M, 4 N and 3 S). 3.2. Univariate and multivariate risk factor analysis Depending on the occurrence of a complication or not, patients were grouped into two: (159) with and (74) without complications. Descriptive statistics with respect to 12 independent risk factors: age, sex, KPS < 70, CCI > 3, tumor side-, tumor lobe localization, eloquent localization, CE-PTV, MGMT-methylation, IDH-1 mutation, EOR and adjuvant therapy are depicted in Table 2. Using the univariate analysis method, risk factor analysis for the occurrence of a complication was performed for the 12 variables (Table 2). Age (P = 0.003), KPS < 70 (P = 0.002), CCI > 3 (P = 0.03), eloquent tumor location (P = 0.001) and adjuvant therapy (P = 0.034) proved significant risk factors for encountering a complication. Gender, tumor side, lobe localization, CE-PTV, MGMT-methylation, IDH-1 mutation and EOR were not significantly associated with complication occurrence. On multivariate logistic regression analysis, only eloquent tumor location proved as a significant complication risk factor (P < 0.007, OR 1.94, 95%CI: 1.20–3.13) (Table 2). 3.3. Patient outcome Survival data was available for 222 patients, median OS was 9.5 months (range 0–72) (95%CI: 11.24–14.41). We observed a 35% perioperative morbidity rate and 8% perioperative mortality rate. Occurrence of complications was significantly associated with decreased survival (P = 0.015) (Fig. 1).

57

of approximately 40% and 30%, respectively [17]. In an attempt to improve patient outcome, numerous studies spanning the last decade have named various patient associated prognostic and treatment related predictive factors influencing outcome [18]. Although thromboembolic events, seizures, fluctuations in neurologic symptoms, and adverse effects from corticosteroids and chemotherapies have been named as frequent clinical complications the exact impact on outcome is not known [19]. With our study we aimed at assessing risk factors for treatment related complications and evaluate the influence on outcome. 4.1. Complication analysis and correlations Of the 159 (68%) patients who manifested with complications, medical problems were most often encountered 174 medical complications by 119 (56%) patients, whereas 77 patients experienced 90 neurological complications. Complications actually relating to surgery were extremely few 17 patients with 17 surgical complications (10 cases of wound infection and 7 cases of intracranial bleeding). Also the perioperative lethality resulting from a surgical casualty was equally low with 4 intracranial hemorrhages. On the other hand, 14 patients encountered a medical mortality. The common terminology criteria for adverse events v3.0 (CTCAE v3.0) employed in many clinical trials is generally organized by organ system categories. Although it represents a comprehensive, multimodality grading system including both acute and late effects [20] it fails to address complications in relation to treatment measures. The classification system we used in our study as employed in the GO project categorizes complications associating to craniotomy more specifically into neurological, surgical and medical complications. Standardized reporting systems for postoperative complications is still not well defined for many surgical specialities, this makes interpreting literature and measuring surgical outcomes difficult [21]. Surgery, most often is the initial therapy of choice in GB treatment hence prospective studies assessing complications relating to GB treatment should employ a simple, reproducible, flexible, and applicable principle of complication assessment. We observed a 41% seizure frequency in our study, which correlates with previous reports of a 30–40% seizure incidence in malignant Glioma patients [22]. This emphasizes the need to adequately diagnose and treat this adverse event. Further antiepileptic monotherapy has proved to associate with higher compliance and less adverse effects [23]. Even though about 30% of malignant glioma patients develop deep venous thromboembolism (VT) at some point of disease course [24], there is no data or recommendation regarding prolonged thrombosis prophylaxis after patient ambulation. We recorded vascular complications specifically VT at a frequency of 7% of which 2 cases were lethal. Medical complications frequently encountered in our cohort were urogenital and gastrointestinal problems with 5 lethal cases (3 cases of urosepsis and 2 cases of gastrointestinal bleeding). Hence these complications must not be underestimated. In our cohort most complications occurred outside the perioperative (<30 days after surgery) time scope. 54% of the cases strictly after this time point and 12% at both < and >30 days after surgery. In literature, the time of complication or adverse event record in relation to index surgery is heterogenous, some take the 6 week span [4,20] whereas others, the more economic 30 days [25]. Taking the last measure, we recorded 35% morbidity rate and 8% mortality rate. 4.2. Risk factor for complications and outcome

4. Discussion To date GB patient outcome remains dismal, patients with the most favorable combination of prognostic factors receiving optimal up-to-date therapy still have 2- and 5-year survival rates

Our study shows that encountering a treatment associated complication has a detrimental impact on OS (P = 0.015). Further the most consistently named factors associating with improved survival [26], age and poor preoperative functional status also

58

G. Ening et al. / Clinical Neurology and Neurosurgery 134 (2015) 55–59

Table 2 Univariate and multivariate analysis factors in relation to complication occurrence. Variable

No. of patients (%) Without complication

With complication

Age (mean ± SD)

57.9

63.9

Sex Male Female

39(33) 35(30)

78(67) 81(70)

KPS <70 >70

7(14) 67(36)

42(86) 117(64)

CCI ≤3 >3

47(38) 27(25)

78(62) 80(75)

Tumorside location Left Right

33(27) 41(37)

88(73) 71(63)

Tumor location Frontal Temporal Parietal Occipital Not localized to one site Others

24(32) 17(36) 9(26) 5(56) 18(28) 1(33)

51(68) 30(64) 25(74) 4(44) 47(72) 2(67)

Eloquenz Eloquent Near-eloquent Non-eloquent

8(22) 27(24) 38(51)

29(78) 88(76) 37(49)

Tumor volume CE-PTV CE-PTV (cm3 )

39.05

37.41

EOR <95% >95% < 100% 100%

16(31) 19(34) 21(39)

35(69) 36(66) 33(62)

MGMT Methylated Unmethylated

25(31) 31(31)

55(69) 68(69)

IDH1 Wildtype Mutant

56(32) 4(36)

121(68) 7(64)

Adjuvant therapy RT/TMZ → TMZ Other

56(38) 11(22)

92(62) 38(77)

*

Univariate analysis

Multivariate analysis

P value

P value

0.003*

0.282

Odds ratio

95%CI

1018

0.986–1.050

0.353

0.002*

0.29

0.589

0.221–1.571

0.03*

0.703

0.853

0.376–1.934

0.083

0.655

0.001*

0.007

34,335

1.202–3.13

0.794 0.42

0.123

0.488

0.034

Statistically significant (P < 0.05).

significantly associated with a higher risk for encountering a treatment associated complication (Table 2). The more preoperative comorbidities (>3) a patient had, significantly posed a risk for encountering a complication. As to how these factors confound each other on impacting outcome still remains subject of further studies. Recent studies describe the CE-PTV to be the more significant predictor of survival compared to EOR [27]. However, in our study group neither EOR nor CE-PTV proved as a significant predictor for complications. Eloquent tumor location on the other hand, significantly associated with complications at both the univariate and multivatiate levels, implying that not extensive surgery but surgery at eloquent locations impacts complication occurrence. This emphasizes the eminent need to keep postoperative complications at a minimum by implementing modern technologies including intraoperative electro-physiological motor, sensory, and speech mapping, as well as interactive image-guidance systems such as resection with 5-ALA and intra-operative MRI scanning. Therapy modalities other than standard treatment (RT/TMZ → TMZ) proved to be significantly associated with

more complications. This confers the good tolerance of this established adjuvant treatment regime and implies that in implementing more aggressive therapies with the aim to improve survival [22], therapy complication management should not be underestimated. However MGMT-promoter methylation, a proven outcome predictor in association with TMZ treatment [3] did not show significant association with complication occurrence in our study. This was also the case for IDH-1 mutational status. 4.3. Study strength and limitations Our study is to our knowledge the first to assess volumetric data and modern molecular parameters in relation to complication occurrence, however it is limited by the retrospective database. Also our outcome evaluation did not survey quality of life data, global health status parameters or neurocognitive data. End-of-life care and palliative care services can help reduce hospitalization by as much as 80%, and these services have also been shown to extend survival [25]. Hence future prospective complication evaluation studies have to incorporate these outcome measures.

G. Ening et al. / Clinical Neurology and Neurosurgery 134 (2015) 55–59

59

Fig. 1. Kaplan–Meier curve depicting significant occurrence of complications in associated with decreased survival (P = 0.015).

5. Conclusion Our comprehensive analysis showed risk factors for predicting GB therapy associated compilations to be age, poor functional status, adjuvant therapy and eloquent tumor location. The later proved to be the strongest predictor as with a 2 fold increased risk for postoperative complication occurrence upon tumor resection here”. Furthermore, we could prove that encountering a complication significantly decreases patient survival. This emphasizes the need to identify patients at risk for complication development and implement adequate treatment measures. References [1] Tait MJ, Petrik V, Loosemore A, et al. Survival of patients with glioblastoma multiforme has not improved between 1993 and 2004: analysis of 625 cases. Br J Neurosurg 2007;21:496–500. [2] Weller M, van den Bent M, Hopkins K, et al. EANO guideline for the diagnosis and treatment of anaplastic gliomas and glioblastoma. Lancet Oncol 2014;15:e395–403. [3] Stupp R, Brada M, van den Bent MJ, et al. High-grade glioma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2014;25(Suppl 3):iii93–101. [4] Gulati S, Jakola AS, Nerland US, et al. The risk of getting worse: surgically acquired deficits, perioperative complications, and functional outcomes after primary resection of glioblastoma. World Neurosurg 2011;76:572–9. [5] Rahman M, Neal D, Fargen KM, et al. Establishing standard performance measures for adult brain tumor patients: a Nationwide Inpatient Sample database study. Neuro Oncol 2013;15:1580–8. [6] Fisher JL, Palmisano S, Schwartzbaum JA, et al. Comorbid conditions associated with glioblastoma. J Neurooncol 2014;116:585–91. [7] Tremont-Lukats IW, Ratilal BO, Armstrong T, et al. Antiepileptic drugs for preventing seizures in people with brain tumors. Cochrane Database Syst Rev 2008;16(April (2)):CD004424. [8] Louis DN, Ohgaki H, Wiestler OD, et al. The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 2007;114:97–109. [9] Kleihues P, Louis DN, Scheithauer BW, et al. The WHO classification of tumors of the nervous system. J Neuropathol Exp Neurol 2002;61:215–25 (Discussion 226-219). [10] Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83.

[11] Balducci M, Fiorentino A, De Bonis P, et al. Impact of age and co-morbidities in patients with newly diagnosed glioblastoma: a pooled data analysis of three prospective mono-institutional phase II studies. Med Oncol 2012;29: 3478–83. [12] Lacroix M, Abi-Said D, Fourney DR, et al. A multivariate analysis of 416 patients with glioblastoma multiforme: prognosis, extent of resection, and survival. J Neurosurg 2001;95:190–8. [13] Shi WM, Wildrick DM, Sawaya R. Volumetric measurement of brain tumors from MR imaging. J Neurooncol 1998;37:87–93. [14] Hartmann C, Hentschel B, Simon M, et al. Long-term survival in primary glioblastoma with versus without isocitrate dehydrogenase mutations. Clin Cancer Res 2013;19:5146–57. [15] Pusch S, Sahm F, Meyer J, et al. Glioma IDH1 mutation patterns off the beaten track. Neuropathol Appl Neurobiol 2011;37:428–30. [16] Chang SM, Parney IF, McDermott M, et al. Perioperative complications and neurological outcomes of first and second craniotomies among patients enrolled in the Glioma Outcome Project. J Neurosurg 2003;98:1175–81. [17] Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 2005;352:987–96. [18] Weller M, Stupp R, Hegi M, et al. Individualized targeted therapy for glioblastoma: fact or fiction. Cancer J 2012;18:40–4. [19] Omuro A, DeAngelis LM. Glioblastoma and other malignant gliomas: a clinical review. JAMA 2013;310:1842–50. [20] Trotti A, Colevas AD, Setser A, et al. CTCAE v3.0: development of a comprehensive grading system for the adverse effects of cancer treatment. Semin Radiat Oncol 2003;13:176–81. [21] Khan A, Palit V, Myatt A, et al. Assessment of Clavien–Dindo classification in patients >75 years undergoing nephrectomy/nephroureterectomy. Urol Ann 2013;5:18–22. [22] Preusser M, de Ribaupierre S, Wohrer A, et al. Current concepts and management of glioblastoma. Ann Neurol 2011;70:9–21. [23] St Louis EK, Rosenfeld WE, Bramley T. Antiepileptic drug monotherapy: the initial approach in epilepsy management. Curr Neuropharmacol 2009;7: 77–82. [24] Simanek R, Vormittag R, Hassler M, et al. Venous thromboembolism and survival in patients with high-grade glioma. Neuro Oncol 2007;9:89–95. [25] Nuno M, Ly D, Ortega A, et al. Does 30-day readmission affect long-term outcome among glioblastoma patients? Neurosurgery 2014;74:196–204 (discussion 204–195). [26] Chaichana KL, Pendleton C, Chambless L, et al. Multi-institutional validation of a preoperative scoring system which predicts survival for patients with glioblastoma. J Clin Neurosci 2013;20:1422–6. [27] Grabowski MM, Recinos PF, Nowacki AS, et al. Residual tumor volume versus extent of resection: predictors of survival after surgery for glioblastoma. J Neurosurg 2014;121(November (5)):1–9, http://dx.doi.org/10.3171/2014.7.JNS132449 [Epub 2014 Sep 5].