Sustained focal antitumor activity of bevacizumab in recurrent glioblastoma Oliver Bähr, Patrick N. Harter, Lutz M. Weise, et al. Neurology published online June 13, 2014 DOI 10.1212/WNL.0000000000000594 This information is current as of June 13, 2014
The online version of this article, along with updated information and services, is located on the World Wide Web at: http://www.neurology.org/content/early/2014/06/13/WNL.0000000000000594.full.html
Neurology ® is the official journal of the American Academy of Neurology. Published continuously since 1951, it is now a weekly with 48 issues per year. Copyright © 2014 American Academy of Neurology. All rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.
Published Ahead of Print on June 13, 2014 as 10.1212/WNL.0000000000000594
Sustained focal antitumor activity of bevacizumab in recurrent glioblastoma
Oliver Bähr, MD Patrick N. Harter, MD Lutz M. Weise, MD Se-Jong You, MD Michel Mittelbronn, MD Michael W. Ronellenfitsch, MD Johannes Rieger, MD Joachim P. Steinbach, MD Elke Hattingen, MD
Correspondence to Dr. Bähr: [email protected]. de
ABSTRACT
Objectives: To investigate the relevance of bevacizumab (BEV)-induced diffusion-restricted lesions and T1-hyperintense lesions in patients with recurrent glioblastoma.
Methods: We prospectively screened 74 BEV-treated patients with recurrent glioblastoma for (1) diffusion-restricted lesions and/or, (2) lesions with a hyperintense signal on precontrast T1-weighted images. We further evaluated overall survival (OS), histopathology of the lesions, and patterns of progression. Results: Twenty-five of 74 patients (34%) developed T1-hyperintense lesions, whereas diffusionrestricted lesions could be detected in 35 of 74 patients (47%). In 21 of 74 patients (28%), the lesions displayed both features (“double-positive”). OS for patients with double-positive lesions was 13.0 months; patients with neither of these lesions had an OS of 6.6 months (p , 0.005). Histologic evaluation of double-positive lesions revealed extensive calcified necrosis in 4 of 4 patients. Notably, these double-positive lesions were rarely involved in further tumor progression. However, they were associated with an increase in distant recurrences at BEV failure. Conclusions: BEV-induced double-positive MRI lesions are a predictive imaging marker associated with a substantial survival benefit and with improved local control in patients with recurrent glioblastoma. Our data suggest that these lesions are the result of a sustained focal antitumor activity of BEV. Neurology® 2014;83:1–8 GLOSSARY ADC 5 apparent diffusion coefficient; BEV 5 bevacizumab; DWI 5 diffusion-weighted imaging; MGMT 5 O6-methylguanineDNA methyltransferase; OS 5 overall survival; PFS 5 progression-free survival.
Supplemental data at Neurology.org
The impact of bevacizumab (BEV) in recurrent glioblastoma is debatable and no predictive biomarker exists. Moreover, there is concern that BEV may enhance infiltrative disease or remote relapses.1–4 Preclinical data supporting this have not been confirmed by clinical studies.5–9 Another unresolved issue is whether BEV, after initial vascular normalization, increases the extent and degree of hypoxia. Supporting this, preclinical and clinical studies have shown a BEV-induced increase in the expression of hypoxia inducible factor-1a or its target genes.1,10–14 Previously, we described the occurrence of stroke-like lesions on diffusion-weighted imaging (DWI) with corresponding decrease of the apparent diffusion coefficient (ADC).13,14 Our notion that this phenomenon is associated with improved survival was recently confirmed.15 In contrast, one report suggests that restricted diffusion could represent a dense cell pattern of glioblastoma recurrence under BEV treatment.16 Furthermore, we have previously shown that BEV-induced T1-hyperintense lesions on MRI are common in glioblastoma. Moreover, these lesions are predictive of a favorable outcome and correspond to tumor calcifications on CT scans.17 Beyond these findings, we noticed a coappearance of T1-hyperintense lesions and stroke-like lesions. Furthermore, these lesions were not involved in further tumor progression in several index patients. Therefore, we hypothesized that these “double-positive” lesions correspond to regions of focal antitumor activity of BEV. We screened a cohort of 74 BEV-treated patients with recurrent From the Dr. Senckenberg Institute of Neurooncology (O.B., M.W.R., J.R., J.P.S.), Institute of Neurology (Edinger-Institute) (P.N.H., M.M.), Department of Neurosurgery (L.M.W.), and Institute of Neuroradiology (S.-J.Y., E.H.), University Hospital Frankfurt, Goethe University, Frankfurt; and German Cancer Consortium (DKTK) and German Cancer Research Center (DKFZ) (O.B., P.N.H., L.M.W., S.-J.Y., M.M., M.W.R., J.R., J.P.S., E.H.), Heidelberg, Germany. Go to Neurology.org for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article. © 2014 American Academy of Neurology
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
1
Table
Patient characteristics, pretreatment, concomitant therapy, and patterns of progression
All patients (n 5 74)
Patients without doublepositive lesions (n 5 53)
Patients with doublepositive lesions (n 5 21)
Age, y, median (range)
53 (27–76)
54 (27–76)
52 (31–71)
Female, % (n)
37 (27)
36 (19)
38 (8)
KPS, %, median (range)
80 (50–100)
80 (50–100)
80 (50–100)
Primary glioblastoma
88 (65)
85 (45)
95 (20)
Secondary glioblastoma
7 (5)
8 (4)
5 (1)
Characteristics General
Histology, % (n)
Glioblastoma with oligod. component
4 (3)
6 (3)
0 (0)
Gliosarcoma
1 (1)
2 (1)
0 (0)
Not methylated
51 (38)
53 (28)
48 (10)
Methylated
28 (21)
25 (13)
38 (8)
Unknown
20 (15)
23 (12)
14 (3)
No. of resections, median (range)
1 (0–4)
1 (0–4)
1 (0–4)
Patients with 1 resection, % (n)
54 (40)
55 (29)
52 (11)
Patients with biopsy only, % (n)
14 (10)
11 (6)
19 (4)
No. of radiotherapies, median (range)
1 (0–3)
1 (0–3)
1 (1–2)
Patients with 1 radiotherapy, % (n)
66 (49)
70 (37)
57 (12)
Patients with 2 radiotherapies, % (n)
30 (22)
25 (13)
43 (9)
Patients without radiotherapy, % (n)
1 (1)a
2 (1)
0 (0)
MGMT status, % (n)
Surgery
Radiotherapy
Previous chemotherapy No. of chemotherapies, median (range)
2 (0–6)
2 (0–5)
2 (1–6)
Patients without chemotherapy, % (n)
1 (1)b
2 (1)
0 (0)
With 1 chemotherapy, % (n)
18 (13)
15 (8)
24 (5)
With 2 chemotherapies, % (n)
47 (35)
47 (25)
48 (10)
With 3 or more chemotherapies, % (n)
34 (25)
36 (19)
29 (6)
No. of recurrences, median (range)
3 (1–6)
3 (1–6)
3 (1–5)
Patients with 2 or 3 recurrences, % (n)
68 (50)
72 (38)
57 (12)
363 (34–1,703)
357 (34–1,703)
390 (132–1,510)
55 (41)
57 (30)
52 (11)
Recurrences
Time from diagnosis glioblastoma, d, median (range) Concomitant therapy, % (n) BEV monotherapy
c
Irinotecan
26 (19)
23 (12)
33 (7)
Radiotherapy
12 (9)c
15 (8)
5 (1)
Temozolomide
1 (1)
2 (1)
0 (0)
Lomustin
1 (1)
0 (0)
5 (1)
Other
5 (4)
6 (3)
5 (1)
Continued
2
Neurology 83
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Table
Continued
All patients (n 5 74)
Patients without doublepositive lesions (n 5 53)
Patients with doublepositive lesions (n 5 21)
Local
53 (32)
55 (24)
47 (8)
Diffuse
23 (14)
27 (12)
12 (2)
Distant
25 (15)
18 (8)
41 (7)
Characteristics Patterns of progression, % (n)
Abbreviations: BEV 5 bevacizumab; KPS 5 Karnofsky performance status; MGMT 5 O6-methylguanine-DNA methyltransferase; oligod. 5 oligodendroglial. Patient characteristics, pretreatments, concomitant therapy, and MRI pattern at progression at BEV failure for the whole cohort and for patients with and without double-positive lesions are shown. Because 9 patients without and 4 patients with the respective lesions had not yet progressed, the total number of patients for patterns of progression is lower than for the remainder of the table. a One patient without radiotherapy before BEV received radiotherapy as concomitant therapy with BEV. b One patient without chemotherapy before BEV had early progressive disease before the end of initial radiotherapy and received BEV as a rescue therapy. c One patient received radiotherapy and irinotecan.
glioblastoma for these MRI alterations. We evaluated overall survival (OS), histopathology, and patterns of progression. METHODS Study population. We prospectively enrolled patients with the neuropathologic diagnosis of a glioblastoma or gliosarcoma. The radiologic confirmation of recurrence, adequate laboratory values, scheduled initiation of a BEV-based therapy, and a full baseline MRI including DWI and ADC were mandatory. Because this was a noninterventional study, the treating physician defined the treatment plan and any concomitant therapy. To increase the number of evaluable patients, we screened our electronic patient records for patients fulfilling the same criteria. From January 2008 to December 2012, we treated 131 patients with BEV in our neurooncology outpatient unit. Eightyfour patients fulfilled the criteria as defined in our prospective study. Ten did not have sufficient radiologic follow-up. Finally, 74 patients were eligible. Forty-two patients were already prospectively included. Therefore, 32 additional patients were identified and extended our evaluable cohort.
Standard protocol approvals, registrations, and patient consents. Our institutional review board approved the prospective part of this study (University Hospital Frankfurt; ethics committee; reference number 4/09-SNO 01/09). The retrospective add-on approach was also approved (University Hospital Frankfurt; ethics committee; reference number 4/09). All patients gave their written informed consent.
Magnetic resonance imaging. MRI was done on a 3T scanner (Magnetom Trio; Siemens Medical AG, Erlangen, Germany) or a 1.5T scanner (Intera; Philips Medical Systems, Best, the Netherlands). MRI scans before therapy and until further tumor progression were mandatory. In addition, all patients had at least T1-weighted sequences before and after IV administration of gadolinium-containing contrast agent, T2-weighted sequences, and DWI sequences (mostly b 5 0, 500, 1,000 at 3 tesla, and b 5 0, 1,000 at 1.5 tesla) with calculated ADC maps. To determine the response to therapy, we used the updated response assessment criteria for high-grade gliomas (RANO [Revised Assessment in Neuro-Oncology]).18 The precontrast T1-weighted images, DWI, and ADC maps of each time point were jointly analyzed for unequivocal, therapyinduced lesions with a hyperintense signal on T1 and for new lesions with restricted diffusion (high DWI signal and
corresponding low ADC values) by one neuroradiologist (E.H.) and one neurooncologist (O.B.). Regarding T1-hyperintense lesions, MRI scans were rated with 0 for no T1-hyperintense lesions, 1 for punctate or faint T1-hyperintense lesions, and 2 for clear T1-hyperintense lesions. Only patients rated with 2 were regarded as patients with unequivocal T1-hyperintense lesions. For restricted diffusion, MRI scans were rated with 0 for no restricted diffusion and 1 for well-demarcated lesions of restricted diffusion. Lesions that displayed both features were defined as “double-positive.” To evaluate whether these lesions were involved in further tumor progression under BEV, E.H. and O.B. jointly examined follow-up MRIs. Particular attention was given to the positional relationship between double-positive lesions and the area of progressive tumor. To answer the question of whether the occurrence of these lesions was associated with a change in the patterns of progression, the pattern of recurrence at BEV failure was analyzed by an experienced neuroradiologist (S.-J.Y.) who was blinded for the scientific rationale of the analysis. We predefined modified categories as reported by Pope et al.8: 1. Local: Enhancing or nonenhancing tumor at or within 3 cm of the primary site. 2. Diffuse: Recurrent tumor extending more than 3 cm from the primary site with at least 50% of the margin of the recurrent tumor qualitatively assessed as poorly defined. 3. Distant: One or more new noncontiguous lesions (enhancing or not) with at least 1 cm distance to primary site.
Neuropathologic assessment. Tissue of double-positive lesions was available in 4 patients. Two patients had a stereotactic biopsy, and 2 patients donated their brain for scientific research as a voluntary option in this prospective noninterventional study. All samples were analyzed by 2 experienced neuropathologists (P.N.H., M.M.). We used standard hematoxylin & eosin stainings for morphologic analysis and alizarin red staining to detect mineralizations.19 Statistics. OS after start of BEV treatment was determined using the Kaplan-Meier method and the log-rank test (SPSS Statistics version 20.0; IBM Corp., Armonk, NY). The Fisher exact test was used to calculate the correlation between the occurrence of the described double-positive lesions and patterns of progression. RESULTS Patient characteristics and treatment. The table shows the characteristics of the whole cohort including previous treatments. The median number Neurology 83
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
3
of 3 recurrences before start of BEV is the major difference in comparison with the patient characteristics in the BRAIN Study, with approximately 80% of the patients being treated for their first relapse.20 BEV was administered at the standard dosing of 10 mg/kg body weight IV every other week to all of our patients. The table shows the concomitant therapy.
Figure 1
Radiologic findings
Occurrence of T1-hyperintense lesions and lesions with restricted diffusion. Twenty-five of 74 patients (34%)
showed new or significantly increased lesions with a hyperintense signal on T1 in the tumor area. In 35 of 74 patients (47%), new or significantly increased lesions with restricted diffusion (high DWI signal and corresponding low ADC values) became apparent. In 21 of 74 patients (28%), the lesions were double-positive (figure e-1 on the Neurology® Web site at Neurology.org). As described before, these alterations occurred early (2 to 3 months after start of BEV) and were stable with time.13–15,17 Follow-up MRI scans from 5 representative patients with newly emerged lesions under BEV treatment are presented in figure 1. Comparison of patient characteristics. Subsequently, we compared the patient characteristics of the 21 patients who developed double-positive lesions with the other 53 patients who did not. As shown in the table, there were no substantial differences in patient variables. Association with OS. We determined OS for patients
with or without T1-hyperintense lesions, patients with or without lesions with restricted diffusion, and for patients with or without double-positive lesions. As shown in figure 2A, the occurrence of lesions with restricted diffusion was associated with longer OS, confirming earlier reports of this radiographic phenomenon.13–15 The association between T1-hyperintense lesions and outcome was even more pronounced, as shown in figure 2B, and also confirms a previous report.17 Finally, the strongest association with OS was found for lesions that were double-positive (figure 2C). In addition, we analyzed median OS for the following subgroups: patients with no MRI alterations (n 5 35), patients with T1-hyperintense lesions only (n 5 4), patients with restricted diffusion only (n 5 14), and patients with double-positive lesions (n 5 21). Because of the small patient numbers in these groups, the value of these analyses is limited. Nonetheless, we did not find a relevant OS difference among patients with no MRI alterations (median OS 202 days), patients with T1-hyperintense lesions (median OS 235 days), and patients with restricted diffusion only (median OS 195 days). Only the group of patients with double-positive lesions showed a superior median OS of 394 days. Neuropathologic findings. We were able to analyze tissue
Follow-up MRI scans from 5 representative patients with newly emerged lesions under bevacizumab treatment. Lesions with a hyperintense signal on precontrast T1-weighted images are shown in the first column and areas of restricted diffusion are shown in the second and third columns (white arrows). ADC 5 apparent diffusion coefficient; DWI 5 diffusion-weighted imaging. 4
Neurology 83
specimens from double-positive lesions in 4 patients. Two patients underwent a stereotactic biopsy for suspected tumor progression and the stereotactic trajectory included lesions with hyperintense signal on T1-weighted images and restricted diffusion (figure e-2, patients 1 and 2). In
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Figure 2
Survival analysis
faint contrast enhancement. In this new lesion, we found vital tumor with strong Ki67 staining. Most of the double-positive lesion itself did not reveal vital tumor. At the time the biopsy was performed in patient 2 (MGMT status unknown), we still thought that lesions with restricted diffusion rather correspond to progressive tumor, but we only found extensive necrosis with calcifications in the specimens of this patient and mostly absence of vital tumor tissue. In 2 other patients, both with unmethylated MGMT promoter, we analyzed autopsy tissue (figure e-2, patients 3 and 4). As shown in figure 3, all lesions showed extensive necrosis and marked calcifications. Involvement in tumor progression. Next, we investigated the involvement of double-positive lesions in tumor progression at BEV failure. Therefore, we screened the follow-up imaging for the positional relationship between double-positive lesions and the area of progressive tumor. Of the 21 patients who developed double-positive lesions, 4 had not yet progressed. In 3 of the remaining 17 evaluable patients (18%), the lesion was clearly involved in tumor progression. In the other patients, tumor progression was either clearly distant from the described lesion or, in a few cases, close to the lesion but without involvement of the lesion itself. Figure e-3 shows 2 representative cases. Patterns of recurrence. Based on these findings, we asked whether patients with double-positive lesions had more distant recurrences. Therefore, we conducted a blinded analysis of the patterns of progression in the whole cohort. Intriguingly, we found a considerably higher occurrence of distant recurrences in the group of patients with doublepositive lesions (41% vs 18%), while local and diffuse recurrences were less frequently observed (table). However, a statistical comparison of distant vs nondistant (local 1 diffuse) recurrences in the 2 groups reached only borderline significance (2 3 2 contingency table, Fisher exact test, p 5 0.06). Patterns of progression are not associated with progressionfree survival or OS. Because the increased frequency of
(A) OS of patients without (red line) and with lesions with restricted diffusion (black line, DWI/ADC). (B) OS of patients without (red line) and with T1-hyperintense lesions (black line, T1). (C) OS of patients without (red line) and with double-positive lesions (black line). Tick marks indicate the time of patient censoring. ADC 5 apparent diffusion coefficient; DWI 5 diffusion-weighted imaging; OS 5 overall survival.
patient 1 (methylated O6-methylguanine-DNA methyltransferase [MGMT] promoter), the doublepositive lesion was located close to a new lesion with
distant recurrences could simply result from a longer progression-free survival (PFS) or OS in patients with occurrence of double-positive lesions, we analyzed clinical outcome data together with the type of progression. As shown in figure e-4, there was no association between the pattern of progression and PFS (A) or OS (B). This was also true for the subgroup analysis of patients with and without therapy-induced changes on MRI scans (data not shown). DISCUSSION BEV is widely used in the treatment of recurrent glioblastoma although its precise mode Neurology 83
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
5
Figure 3
Neuropathologic findings
H&E (first column) and alizarin red (second column) staining depicting extensive calcified necrotic areas in the tissue specimens of all 4 patients analyzed. The corresponding neuroradiologic areas of the examined tissue specimens are shown in figure e-2. H&E 5 hematoxylin & eosin.
of action and the impact on OS and tumor growth patterns remain uncertain. Our data are compatible with the hypothesis of a sustained focal antitumor activity of BEV in a subpopulation of patients with recurrent glioblastoma. In our cohort of 74 patients treated with BEV for recurrent glioblastoma, 21 (28%) developed doublepositive lesions with hyperintense signals on precontrast T1-weighted images and restricted diffusion (high DWI signal and corresponding low ADC values). This phenomenon was associated with a compelling 6
Neurology 83
survival benefit for these patients (13.0 vs 6.6 months, figure 2). Because patients with T1-hyperintense lesions only and patients with lesions with restricted diffusion only did not show longer survival, this benefit seems to depend on the presence of both MRI alterations. Notably, these patients were comparable regarding other patient variables (table). Histologic evaluation of these double-positive lesions revealed extensive calcified necrosis in 4 of 4 patients (figure 3). Furthermore, these lesions were involved in tumor progression in 3 of 17 recurrent patients only, suggesting that these nonvital
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
double-positive lesions cannot give rise to recurrent tumor. A blinded analysis of the patterns of progression at the time of BEV failure revealed a trend toward an increase in distant recurrences in the group of patients with these BEV-induced lesions (41% vs 18%, p 5 0.06, table). Because we did not find an association between the type of progression and PFS or OS (figure e-4, A and B), it is unlikely that this increase in distant recurrences is simply the result of prolonged PFS and OS in the group of patients with these MRI alterations. Therefore, a difference in tumor biology seems to be more likely responsible for the increase in distant relapses. Two aspects appear obvious: BEV-induced MRI changes could (1) be limited to tumors with a propensity for distant relapses, or (2) be the result of BEVinduced hypoxia driving infiltrative tumor growth and promoting distant relapses.6,7,21 Despite the probable increase in distant recurrences, the impressive survival benefit for patients with double-positive lesions emphasizes the value of BEV in the treatment of recurrent glioblastoma. Moreover, our radiologic and histologic findings suggest that BEV not only reduces vascular permeability, edema, and contrast enhancement but might also have a real and sustained antitumor effect in a relevant subgroup of patients. The tumor tissue in the areas where double-positive lesions develop might be particularly dependent on vascular endothelial growth factor and therefore more vulnerable for a BEV-induced depletion of vascular endothelial growth factor. This could facilitate a true antiangiogenic effect resulting in severe hypoxia and finally in extensive calcified necrosis characterized by the described double-positive lesions on MRI. In line with these considerations, we and others have reported reduced blood perfusion in lesions with restricted diffusion that developed under BEV.13,15 For clinical practice, our findings are important for the evaluation of follow-up MRI of patients with glioblastoma treated with BEV. It should be recognized that double-positive lesions with restricted diffusion and hyperintense signal on T1-weighted images corresponding to calcification on CT scans typically appear early under BEV treatment and are stable with time. They probably correspond to a focal antitumor activity of BEV and should not carelessly be interpreted as progressive tumor, cerebral ischemia, or microhemorrhages. Follow-up MRI after 4 weeks or additional CT can help to distinguish between favorable treatment effects and adverse events of BEV treatment. Thereby, our findings can prevent the unwarranted discontinuation of BEV treatment. Eventually, the question remains whether doublepositive lesions are specific for BEV treatment. A review of the literature did not reveal any study systematically investigating the appearance of T1
hyperintensities and stroke-like DWI lesions under other therapy regimens. From our daily practice, however, we are convinced that these alterations are regularly seen in patients on BEV only. In addition, none of our 74 evaluated, heavily pretreated patients showed double-positive lesions at baseline before treatment with BEV. Our study provides basic imaging markers predictive of a substantial survival benefit for BEV-treated patients with recurrent glioblastoma. The histologic evaluation of these lesions showed extensive calcified necrosis. Furthermore, BEV-induced lesions with restricted diffusion and hyperintense T1 signal are rarely the origin of progressive tumor and might lead to an increase of distant recurrences. We suggest that these lesions are the result of a sustained focal antitumor activity of BEV. AUTHOR CONTRIBUTIONS Oliver Bähr: design and conceptualization, analysis and interpretation, drafting and revising the manuscript. Patrick N. Harter, Lutz M. Weise, Se-Jong You, Michel Mittelbronn, Michael W. Ronellenfitsch, Johannes Rieger, and Joachim Steinbach: analysis and interpretation, drafting and revising the manuscript. Elke Hattingen: design and conceptualization, analysis and interpretation, drafting and revising the manuscript.
ACKNOWLEDGMENT The Dr. Senckenberg Institute of Neurooncology is supported by the Dr. Senckenberg Foundation and the Hertie Foundation. J.P.S. is Hertie Professor of Neurooncology.
STUDY FUNDING No targeted funding reported.
DISCLOSURE O. Bähr has served as a consultant for Roche and has received a travel grant from Roche, the European distributor of bevacizumab (Avastin). P. Harter reports no disclosures relevant to the manuscript. L. Weise has received speaker honoraria from Medtronic GmbH, Meerbusch, Germany. S. You, M. Mittelbronn, and M. Ronellenfitsch report no disclosures relevant to the manuscript. J. Rieger and J. Steinbach have served as consultants for Roche, the European distributor of bevacizumab (Avastin). E. Hattingen reports no disclosures relevant to the manuscript. Go to Neurology.org for full disclosures.
Received October 11, 2013. Accepted in final form April 11, 2014. REFERENCES 1. de Groot JF, Fuller G, Kumar AJ, et al. Tumor invasion after treatment of glioblastoma with bevacizumab: radiographic and pathologic correlation in humans and mice. Neuro Oncol 2010;12:233–242. 2. Iwamoto FM, Abrey LE, Beal K, et al. Patterns of relapse and prognosis after bevacizumab failure in recurrent glioblastoma. Neurology 2009;73:1200–1206. 3. Norden AD, Young GS, Setayesh K, et al. Bevacizumab for recurrent malignant gliomas: efficacy, toxicity, and patterns of recurrence. Neurology 2008;70:779–787. 4. Zuniga RM, Torcuator R, Jain R, et al. Efficacy, safety and patterns of response and recurrence in patients with recurrent high-grade gliomas treated with bevacizumab plus irinotecan. J Neurooncol 2009;91:329–336. Neurology 83
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
7
5. 6.
7.
8.
9.
10.
11.
12.
13.
14.
8
Chamberlain MC. Radiographic patterns of relapse in glioblastoma. J Neurooncol 2011;101:319–323. Ebos JM, Lee CR, Cruz-Munoz W, Bjarnason GA, Christensen JG, Kerbel RS. Accelerated metastasis after short-term treatment with a potent inhibitor of tumor angiogenesis. Cancer Cell 2009;15:232–239. Paez-Ribes M, Allen E, Hudock J, et al. Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis. Cancer Cell 2009;15:220–231. Pope WB, Xia Q, Paton VE, et al. Patterns of progression in patients with recurrent glioblastoma treated with bevacizumab. Neurology 2011;76:432–437. Wick A, Dorner N, Schafer N, et al. Bevacizumab does not increase the risk of remote relapse in malignant glioma. Ann Neurol 2011;69:586–592. DeLay M, Jahangiri A, Carbonell WS, et al. Microarray analysis verifies two distinct phenotypes of glioblastomas resistant to antiangiogenic therapy. Clin Cancer Res 2012; 18:2930–2942. Hu YL, DeLay M, Jahangiri A, et al. Hypoxia-induced autophagy promotes tumor cell survival and adaptation to antiangiogenic treatment in glioblastoma. Cancer Res 2012;72:1773–1783. Keunen O, Johansson M, Oudin A, et al. Anti-VEGF treatment reduces blood supply and increases tumor cell invasion in glioblastoma. Proc Natl Acad Sci USA 2011; 108:3749–3754. Rieger J, Bahr O, Muller K, Franz K, Steinbach J, Hattingen E. Bevacizumab-induced diffusion-restricted lesions in malignant glioma patients. J Neurooncol 2009;99:49–56. Rieger J, Bahr O, Ronellenfitsch MW, Steinbach J, Hattingen E. Bevacizumab-induced diffusion
Neurology 83
15.
16.
17.
18.
19.
20.
21.
restriction in patients with glioma: tumor progression or surrogate marker of hypoxia? J Clin Oncol 2010;28: e477. Mong S, Ellingson BM, Nghiemphu PL, et al. Persistent diffusion-restricted lesions in bevacizumab-treated malignant gliomas are associated with improved survival compared with matched controls. AJNR Am J Neuroradiol 2012;33:1763–1770. Gerstner ER, Frosch MP, Batchelor TT. Diffusion magnetic resonance imaging detects pathologically confirmed, nonenhancing tumor progression in a patient with recurrent glioblastoma receiving bevacizumab. J Clin Oncol 2010;28: e91–e93. Bähr O, Hattingen E, Rieger J, Steinbach JP. Bevacizumabinduced tumor calcifications as a surrogate marker of outcome in patients with glioblastoma. Neuro Oncol 2011;13: 1020–1029. Wen PY, Macdonald DR, Reardon DA, et al. Updated response assessment criteria for high-grade gliomas: Response Assessment in Neuro-Oncology Working Group. J Clin Oncol 2010;28:1963–1972. Meloan SN, Puchtler H, Valentine LS. Alkaline and acid alizarin red S stains for alkali-soluble and alkaliinsoluble calcium deposits. Arch Pathol 1972;93:190– 197. Friedman HS, Prados MD, Wen PY, et al. Bevacizumab alone and in combination with irinotecan in recurrent glioblastoma. J Clin Oncol 2009;27:4733–4740. Kunkel P, Ulbricht U, Bohlen P, et al. Inhibition of glioma angiogenesis and growth in vivo by systemic treatment with a monoclonal antibody against vascular endothelial growth factor receptor-2. Cancer Res 2001;61: 6624–6628.
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Sustained focal antitumor activity of bevacizumab in recurrent glioblastoma Oliver Bähr, Patrick N. Harter, Lutz M. Weise, et al. Neurology published online June 13, 2014 DOI 10.1212/WNL.0000000000000594 This information is current as of June 13, 2014 Updated Information & Services
including high resolution figures, can be found at: http://www.neurology.org/content/early/2014/06/13/WNL.00000 00000000594.full.html
Supplementary Material
Supplementary material can be found at: http://www.neurology.org/content/suppl/2014/06/13/WNL.00000 00000000594.DC1.html
Subspecialty Collections
This article, along with others on similar topics, appears in the following collection(s): All Cerebrovascular disease/Stroke http://www.neurology.org//cgi/collection/all_cerebrovascular_dis ease_stroke Risk factors in epidemiology http://www.neurology.org//cgi/collection/risk_factors_in_epidem iology Stroke prevention http://www.neurology.org//cgi/collection/stroke_prevention
Permissions & Licensing
Information about reproducing this article in parts (figures,tables) or in its entirety can be found online at: http://www.neurology.org/misc/about.xhtml#permissions
Reprints
Information about ordering reprints can be found online: http://www.neurology.org/misc/addir.xhtml#reprintsus
The online version of this article, along with updated information and services, is located on the World Wide Web at: http://www.neurology.org/content/early/2014/06/13/WNL.0000000000000594.full.html
Neurology ® is the official journal of the American Academy of Neurology. Published continuously since 1951, it is now a weekly with 48 issues per year. Copyright © 2014 American Academy of Neurology. All rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.
Published Ahead of Print on June 13, 2014 as 10.1212/WNL.0000000000000594
Sustained focal antitumor activity of bevacizumab in recurrent glioblastoma
Oliver Bähr, MD Patrick N. Harter, MD Lutz M. Weise, MD Se-Jong You, MD Michel Mittelbronn, MD Michael W. Ronellenfitsch, MD Johannes Rieger, MD Joachim P. Steinbach, MD Elke Hattingen, MD
Correspondence to Dr. Bähr: [email protected]. de
ABSTRACT
Objectives: To investigate the relevance of bevacizumab (BEV)-induced diffusion-restricted lesions and T1-hyperintense lesions in patients with recurrent glioblastoma.
Methods: We prospectively screened 74 BEV-treated patients with recurrent glioblastoma for (1) diffusion-restricted lesions and/or, (2) lesions with a hyperintense signal on precontrast T1-weighted images. We further evaluated overall survival (OS), histopathology of the lesions, and patterns of progression. Results: Twenty-five of 74 patients (34%) developed T1-hyperintense lesions, whereas diffusionrestricted lesions could be detected in 35 of 74 patients (47%). In 21 of 74 patients (28%), the lesions displayed both features (“double-positive”). OS for patients with double-positive lesions was 13.0 months; patients with neither of these lesions had an OS of 6.6 months (p , 0.005). Histologic evaluation of double-positive lesions revealed extensive calcified necrosis in 4 of 4 patients. Notably, these double-positive lesions were rarely involved in further tumor progression. However, they were associated with an increase in distant recurrences at BEV failure. Conclusions: BEV-induced double-positive MRI lesions are a predictive imaging marker associated with a substantial survival benefit and with improved local control in patients with recurrent glioblastoma. Our data suggest that these lesions are the result of a sustained focal antitumor activity of BEV. Neurology® 2014;83:1–8 GLOSSARY ADC 5 apparent diffusion coefficient; BEV 5 bevacizumab; DWI 5 diffusion-weighted imaging; MGMT 5 O6-methylguanineDNA methyltransferase; OS 5 overall survival; PFS 5 progression-free survival.
Supplemental data at Neurology.org
The impact of bevacizumab (BEV) in recurrent glioblastoma is debatable and no predictive biomarker exists. Moreover, there is concern that BEV may enhance infiltrative disease or remote relapses.1–4 Preclinical data supporting this have not been confirmed by clinical studies.5–9 Another unresolved issue is whether BEV, after initial vascular normalization, increases the extent and degree of hypoxia. Supporting this, preclinical and clinical studies have shown a BEV-induced increase in the expression of hypoxia inducible factor-1a or its target genes.1,10–14 Previously, we described the occurrence of stroke-like lesions on diffusion-weighted imaging (DWI) with corresponding decrease of the apparent diffusion coefficient (ADC).13,14 Our notion that this phenomenon is associated with improved survival was recently confirmed.15 In contrast, one report suggests that restricted diffusion could represent a dense cell pattern of glioblastoma recurrence under BEV treatment.16 Furthermore, we have previously shown that BEV-induced T1-hyperintense lesions on MRI are common in glioblastoma. Moreover, these lesions are predictive of a favorable outcome and correspond to tumor calcifications on CT scans.17 Beyond these findings, we noticed a coappearance of T1-hyperintense lesions and stroke-like lesions. Furthermore, these lesions were not involved in further tumor progression in several index patients. Therefore, we hypothesized that these “double-positive” lesions correspond to regions of focal antitumor activity of BEV. We screened a cohort of 74 BEV-treated patients with recurrent From the Dr. Senckenberg Institute of Neurooncology (O.B., M.W.R., J.R., J.P.S.), Institute of Neurology (Edinger-Institute) (P.N.H., M.M.), Department of Neurosurgery (L.M.W.), and Institute of Neuroradiology (S.-J.Y., E.H.), University Hospital Frankfurt, Goethe University, Frankfurt; and German Cancer Consortium (DKTK) and German Cancer Research Center (DKFZ) (O.B., P.N.H., L.M.W., S.-J.Y., M.M., M.W.R., J.R., J.P.S., E.H.), Heidelberg, Germany. Go to Neurology.org for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article. © 2014 American Academy of Neurology
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
1
Table
Patient characteristics, pretreatment, concomitant therapy, and patterns of progression
All patients (n 5 74)
Patients without doublepositive lesions (n 5 53)
Patients with doublepositive lesions (n 5 21)
Age, y, median (range)
53 (27–76)
54 (27–76)
52 (31–71)
Female, % (n)
37 (27)
36 (19)
38 (8)
KPS, %, median (range)
80 (50–100)
80 (50–100)
80 (50–100)
Primary glioblastoma
88 (65)
85 (45)
95 (20)
Secondary glioblastoma
7 (5)
8 (4)
5 (1)
Characteristics General
Histology, % (n)
Glioblastoma with oligod. component
4 (3)
6 (3)
0 (0)
Gliosarcoma
1 (1)
2 (1)
0 (0)
Not methylated
51 (38)
53 (28)
48 (10)
Methylated
28 (21)
25 (13)
38 (8)
Unknown
20 (15)
23 (12)
14 (3)
No. of resections, median (range)
1 (0–4)
1 (0–4)
1 (0–4)
Patients with 1 resection, % (n)
54 (40)
55 (29)
52 (11)
Patients with biopsy only, % (n)
14 (10)
11 (6)
19 (4)
No. of radiotherapies, median (range)
1 (0–3)
1 (0–3)
1 (1–2)
Patients with 1 radiotherapy, % (n)
66 (49)
70 (37)
57 (12)
Patients with 2 radiotherapies, % (n)
30 (22)
25 (13)
43 (9)
Patients without radiotherapy, % (n)
1 (1)a
2 (1)
0 (0)
MGMT status, % (n)
Surgery
Radiotherapy
Previous chemotherapy No. of chemotherapies, median (range)
2 (0–6)
2 (0–5)
2 (1–6)
Patients without chemotherapy, % (n)
1 (1)b
2 (1)
0 (0)
With 1 chemotherapy, % (n)
18 (13)
15 (8)
24 (5)
With 2 chemotherapies, % (n)
47 (35)
47 (25)
48 (10)
With 3 or more chemotherapies, % (n)
34 (25)
36 (19)
29 (6)
No. of recurrences, median (range)
3 (1–6)
3 (1–6)
3 (1–5)
Patients with 2 or 3 recurrences, % (n)
68 (50)
72 (38)
57 (12)
363 (34–1,703)
357 (34–1,703)
390 (132–1,510)
55 (41)
57 (30)
52 (11)
Recurrences
Time from diagnosis glioblastoma, d, median (range) Concomitant therapy, % (n) BEV monotherapy
c
Irinotecan
26 (19)
23 (12)
33 (7)
Radiotherapy
12 (9)c
15 (8)
5 (1)
Temozolomide
1 (1)
2 (1)
0 (0)
Lomustin
1 (1)
0 (0)
5 (1)
Other
5 (4)
6 (3)
5 (1)
Continued
2
Neurology 83
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Table
Continued
All patients (n 5 74)
Patients without doublepositive lesions (n 5 53)
Patients with doublepositive lesions (n 5 21)
Local
53 (32)
55 (24)
47 (8)
Diffuse
23 (14)
27 (12)
12 (2)
Distant
25 (15)
18 (8)
41 (7)
Characteristics Patterns of progression, % (n)
Abbreviations: BEV 5 bevacizumab; KPS 5 Karnofsky performance status; MGMT 5 O6-methylguanine-DNA methyltransferase; oligod. 5 oligodendroglial. Patient characteristics, pretreatments, concomitant therapy, and MRI pattern at progression at BEV failure for the whole cohort and for patients with and without double-positive lesions are shown. Because 9 patients without and 4 patients with the respective lesions had not yet progressed, the total number of patients for patterns of progression is lower than for the remainder of the table. a One patient without radiotherapy before BEV received radiotherapy as concomitant therapy with BEV. b One patient without chemotherapy before BEV had early progressive disease before the end of initial radiotherapy and received BEV as a rescue therapy. c One patient received radiotherapy and irinotecan.
glioblastoma for these MRI alterations. We evaluated overall survival (OS), histopathology, and patterns of progression. METHODS Study population. We prospectively enrolled patients with the neuropathologic diagnosis of a glioblastoma or gliosarcoma. The radiologic confirmation of recurrence, adequate laboratory values, scheduled initiation of a BEV-based therapy, and a full baseline MRI including DWI and ADC were mandatory. Because this was a noninterventional study, the treating physician defined the treatment plan and any concomitant therapy. To increase the number of evaluable patients, we screened our electronic patient records for patients fulfilling the same criteria. From January 2008 to December 2012, we treated 131 patients with BEV in our neurooncology outpatient unit. Eightyfour patients fulfilled the criteria as defined in our prospective study. Ten did not have sufficient radiologic follow-up. Finally, 74 patients were eligible. Forty-two patients were already prospectively included. Therefore, 32 additional patients were identified and extended our evaluable cohort.
Standard protocol approvals, registrations, and patient consents. Our institutional review board approved the prospective part of this study (University Hospital Frankfurt; ethics committee; reference number 4/09-SNO 01/09). The retrospective add-on approach was also approved (University Hospital Frankfurt; ethics committee; reference number 4/09). All patients gave their written informed consent.
Magnetic resonance imaging. MRI was done on a 3T scanner (Magnetom Trio; Siemens Medical AG, Erlangen, Germany) or a 1.5T scanner (Intera; Philips Medical Systems, Best, the Netherlands). MRI scans before therapy and until further tumor progression were mandatory. In addition, all patients had at least T1-weighted sequences before and after IV administration of gadolinium-containing contrast agent, T2-weighted sequences, and DWI sequences (mostly b 5 0, 500, 1,000 at 3 tesla, and b 5 0, 1,000 at 1.5 tesla) with calculated ADC maps. To determine the response to therapy, we used the updated response assessment criteria for high-grade gliomas (RANO [Revised Assessment in Neuro-Oncology]).18 The precontrast T1-weighted images, DWI, and ADC maps of each time point were jointly analyzed for unequivocal, therapyinduced lesions with a hyperintense signal on T1 and for new lesions with restricted diffusion (high DWI signal and
corresponding low ADC values) by one neuroradiologist (E.H.) and one neurooncologist (O.B.). Regarding T1-hyperintense lesions, MRI scans were rated with 0 for no T1-hyperintense lesions, 1 for punctate or faint T1-hyperintense lesions, and 2 for clear T1-hyperintense lesions. Only patients rated with 2 were regarded as patients with unequivocal T1-hyperintense lesions. For restricted diffusion, MRI scans were rated with 0 for no restricted diffusion and 1 for well-demarcated lesions of restricted diffusion. Lesions that displayed both features were defined as “double-positive.” To evaluate whether these lesions were involved in further tumor progression under BEV, E.H. and O.B. jointly examined follow-up MRIs. Particular attention was given to the positional relationship between double-positive lesions and the area of progressive tumor. To answer the question of whether the occurrence of these lesions was associated with a change in the patterns of progression, the pattern of recurrence at BEV failure was analyzed by an experienced neuroradiologist (S.-J.Y.) who was blinded for the scientific rationale of the analysis. We predefined modified categories as reported by Pope et al.8: 1. Local: Enhancing or nonenhancing tumor at or within 3 cm of the primary site. 2. Diffuse: Recurrent tumor extending more than 3 cm from the primary site with at least 50% of the margin of the recurrent tumor qualitatively assessed as poorly defined. 3. Distant: One or more new noncontiguous lesions (enhancing or not) with at least 1 cm distance to primary site.
Neuropathologic assessment. Tissue of double-positive lesions was available in 4 patients. Two patients had a stereotactic biopsy, and 2 patients donated their brain for scientific research as a voluntary option in this prospective noninterventional study. All samples were analyzed by 2 experienced neuropathologists (P.N.H., M.M.). We used standard hematoxylin & eosin stainings for morphologic analysis and alizarin red staining to detect mineralizations.19 Statistics. OS after start of BEV treatment was determined using the Kaplan-Meier method and the log-rank test (SPSS Statistics version 20.0; IBM Corp., Armonk, NY). The Fisher exact test was used to calculate the correlation between the occurrence of the described double-positive lesions and patterns of progression. RESULTS Patient characteristics and treatment. The table shows the characteristics of the whole cohort including previous treatments. The median number Neurology 83
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
3
of 3 recurrences before start of BEV is the major difference in comparison with the patient characteristics in the BRAIN Study, with approximately 80% of the patients being treated for their first relapse.20 BEV was administered at the standard dosing of 10 mg/kg body weight IV every other week to all of our patients. The table shows the concomitant therapy.
Figure 1
Radiologic findings
Occurrence of T1-hyperintense lesions and lesions with restricted diffusion. Twenty-five of 74 patients (34%)
showed new or significantly increased lesions with a hyperintense signal on T1 in the tumor area. In 35 of 74 patients (47%), new or significantly increased lesions with restricted diffusion (high DWI signal and corresponding low ADC values) became apparent. In 21 of 74 patients (28%), the lesions were double-positive (figure e-1 on the Neurology® Web site at Neurology.org). As described before, these alterations occurred early (2 to 3 months after start of BEV) and were stable with time.13–15,17 Follow-up MRI scans from 5 representative patients with newly emerged lesions under BEV treatment are presented in figure 1. Comparison of patient characteristics. Subsequently, we compared the patient characteristics of the 21 patients who developed double-positive lesions with the other 53 patients who did not. As shown in the table, there were no substantial differences in patient variables. Association with OS. We determined OS for patients
with or without T1-hyperintense lesions, patients with or without lesions with restricted diffusion, and for patients with or without double-positive lesions. As shown in figure 2A, the occurrence of lesions with restricted diffusion was associated with longer OS, confirming earlier reports of this radiographic phenomenon.13–15 The association between T1-hyperintense lesions and outcome was even more pronounced, as shown in figure 2B, and also confirms a previous report.17 Finally, the strongest association with OS was found for lesions that were double-positive (figure 2C). In addition, we analyzed median OS for the following subgroups: patients with no MRI alterations (n 5 35), patients with T1-hyperintense lesions only (n 5 4), patients with restricted diffusion only (n 5 14), and patients with double-positive lesions (n 5 21). Because of the small patient numbers in these groups, the value of these analyses is limited. Nonetheless, we did not find a relevant OS difference among patients with no MRI alterations (median OS 202 days), patients with T1-hyperintense lesions (median OS 235 days), and patients with restricted diffusion only (median OS 195 days). Only the group of patients with double-positive lesions showed a superior median OS of 394 days. Neuropathologic findings. We were able to analyze tissue
Follow-up MRI scans from 5 representative patients with newly emerged lesions under bevacizumab treatment. Lesions with a hyperintense signal on precontrast T1-weighted images are shown in the first column and areas of restricted diffusion are shown in the second and third columns (white arrows). ADC 5 apparent diffusion coefficient; DWI 5 diffusion-weighted imaging. 4
Neurology 83
specimens from double-positive lesions in 4 patients. Two patients underwent a stereotactic biopsy for suspected tumor progression and the stereotactic trajectory included lesions with hyperintense signal on T1-weighted images and restricted diffusion (figure e-2, patients 1 and 2). In
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Figure 2
Survival analysis
faint contrast enhancement. In this new lesion, we found vital tumor with strong Ki67 staining. Most of the double-positive lesion itself did not reveal vital tumor. At the time the biopsy was performed in patient 2 (MGMT status unknown), we still thought that lesions with restricted diffusion rather correspond to progressive tumor, but we only found extensive necrosis with calcifications in the specimens of this patient and mostly absence of vital tumor tissue. In 2 other patients, both with unmethylated MGMT promoter, we analyzed autopsy tissue (figure e-2, patients 3 and 4). As shown in figure 3, all lesions showed extensive necrosis and marked calcifications. Involvement in tumor progression. Next, we investigated the involvement of double-positive lesions in tumor progression at BEV failure. Therefore, we screened the follow-up imaging for the positional relationship between double-positive lesions and the area of progressive tumor. Of the 21 patients who developed double-positive lesions, 4 had not yet progressed. In 3 of the remaining 17 evaluable patients (18%), the lesion was clearly involved in tumor progression. In the other patients, tumor progression was either clearly distant from the described lesion or, in a few cases, close to the lesion but without involvement of the lesion itself. Figure e-3 shows 2 representative cases. Patterns of recurrence. Based on these findings, we asked whether patients with double-positive lesions had more distant recurrences. Therefore, we conducted a blinded analysis of the patterns of progression in the whole cohort. Intriguingly, we found a considerably higher occurrence of distant recurrences in the group of patients with doublepositive lesions (41% vs 18%), while local and diffuse recurrences were less frequently observed (table). However, a statistical comparison of distant vs nondistant (local 1 diffuse) recurrences in the 2 groups reached only borderline significance (2 3 2 contingency table, Fisher exact test, p 5 0.06). Patterns of progression are not associated with progressionfree survival or OS. Because the increased frequency of
(A) OS of patients without (red line) and with lesions with restricted diffusion (black line, DWI/ADC). (B) OS of patients without (red line) and with T1-hyperintense lesions (black line, T1). (C) OS of patients without (red line) and with double-positive lesions (black line). Tick marks indicate the time of patient censoring. ADC 5 apparent diffusion coefficient; DWI 5 diffusion-weighted imaging; OS 5 overall survival.
patient 1 (methylated O6-methylguanine-DNA methyltransferase [MGMT] promoter), the doublepositive lesion was located close to a new lesion with
distant recurrences could simply result from a longer progression-free survival (PFS) or OS in patients with occurrence of double-positive lesions, we analyzed clinical outcome data together with the type of progression. As shown in figure e-4, there was no association between the pattern of progression and PFS (A) or OS (B). This was also true for the subgroup analysis of patients with and without therapy-induced changes on MRI scans (data not shown). DISCUSSION BEV is widely used in the treatment of recurrent glioblastoma although its precise mode Neurology 83
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
5
Figure 3
Neuropathologic findings
H&E (first column) and alizarin red (second column) staining depicting extensive calcified necrotic areas in the tissue specimens of all 4 patients analyzed. The corresponding neuroradiologic areas of the examined tissue specimens are shown in figure e-2. H&E 5 hematoxylin & eosin.
of action and the impact on OS and tumor growth patterns remain uncertain. Our data are compatible with the hypothesis of a sustained focal antitumor activity of BEV in a subpopulation of patients with recurrent glioblastoma. In our cohort of 74 patients treated with BEV for recurrent glioblastoma, 21 (28%) developed doublepositive lesions with hyperintense signals on precontrast T1-weighted images and restricted diffusion (high DWI signal and corresponding low ADC values). This phenomenon was associated with a compelling 6
Neurology 83
survival benefit for these patients (13.0 vs 6.6 months, figure 2). Because patients with T1-hyperintense lesions only and patients with lesions with restricted diffusion only did not show longer survival, this benefit seems to depend on the presence of both MRI alterations. Notably, these patients were comparable regarding other patient variables (table). Histologic evaluation of these double-positive lesions revealed extensive calcified necrosis in 4 of 4 patients (figure 3). Furthermore, these lesions were involved in tumor progression in 3 of 17 recurrent patients only, suggesting that these nonvital
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
double-positive lesions cannot give rise to recurrent tumor. A blinded analysis of the patterns of progression at the time of BEV failure revealed a trend toward an increase in distant recurrences in the group of patients with these BEV-induced lesions (41% vs 18%, p 5 0.06, table). Because we did not find an association between the type of progression and PFS or OS (figure e-4, A and B), it is unlikely that this increase in distant recurrences is simply the result of prolonged PFS and OS in the group of patients with these MRI alterations. Therefore, a difference in tumor biology seems to be more likely responsible for the increase in distant relapses. Two aspects appear obvious: BEV-induced MRI changes could (1) be limited to tumors with a propensity for distant relapses, or (2) be the result of BEVinduced hypoxia driving infiltrative tumor growth and promoting distant relapses.6,7,21 Despite the probable increase in distant recurrences, the impressive survival benefit for patients with double-positive lesions emphasizes the value of BEV in the treatment of recurrent glioblastoma. Moreover, our radiologic and histologic findings suggest that BEV not only reduces vascular permeability, edema, and contrast enhancement but might also have a real and sustained antitumor effect in a relevant subgroup of patients. The tumor tissue in the areas where double-positive lesions develop might be particularly dependent on vascular endothelial growth factor and therefore more vulnerable for a BEV-induced depletion of vascular endothelial growth factor. This could facilitate a true antiangiogenic effect resulting in severe hypoxia and finally in extensive calcified necrosis characterized by the described double-positive lesions on MRI. In line with these considerations, we and others have reported reduced blood perfusion in lesions with restricted diffusion that developed under BEV.13,15 For clinical practice, our findings are important for the evaluation of follow-up MRI of patients with glioblastoma treated with BEV. It should be recognized that double-positive lesions with restricted diffusion and hyperintense signal on T1-weighted images corresponding to calcification on CT scans typically appear early under BEV treatment and are stable with time. They probably correspond to a focal antitumor activity of BEV and should not carelessly be interpreted as progressive tumor, cerebral ischemia, or microhemorrhages. Follow-up MRI after 4 weeks or additional CT can help to distinguish between favorable treatment effects and adverse events of BEV treatment. Thereby, our findings can prevent the unwarranted discontinuation of BEV treatment. Eventually, the question remains whether doublepositive lesions are specific for BEV treatment. A review of the literature did not reveal any study systematically investigating the appearance of T1
hyperintensities and stroke-like DWI lesions under other therapy regimens. From our daily practice, however, we are convinced that these alterations are regularly seen in patients on BEV only. In addition, none of our 74 evaluated, heavily pretreated patients showed double-positive lesions at baseline before treatment with BEV. Our study provides basic imaging markers predictive of a substantial survival benefit for BEV-treated patients with recurrent glioblastoma. The histologic evaluation of these lesions showed extensive calcified necrosis. Furthermore, BEV-induced lesions with restricted diffusion and hyperintense T1 signal are rarely the origin of progressive tumor and might lead to an increase of distant recurrences. We suggest that these lesions are the result of a sustained focal antitumor activity of BEV. AUTHOR CONTRIBUTIONS Oliver Bähr: design and conceptualization, analysis and interpretation, drafting and revising the manuscript. Patrick N. Harter, Lutz M. Weise, Se-Jong You, Michel Mittelbronn, Michael W. Ronellenfitsch, Johannes Rieger, and Joachim Steinbach: analysis and interpretation, drafting and revising the manuscript. Elke Hattingen: design and conceptualization, analysis and interpretation, drafting and revising the manuscript.
ACKNOWLEDGMENT The Dr. Senckenberg Institute of Neurooncology is supported by the Dr. Senckenberg Foundation and the Hertie Foundation. J.P.S. is Hertie Professor of Neurooncology.
STUDY FUNDING No targeted funding reported.
DISCLOSURE O. Bähr has served as a consultant for Roche and has received a travel grant from Roche, the European distributor of bevacizumab (Avastin). P. Harter reports no disclosures relevant to the manuscript. L. Weise has received speaker honoraria from Medtronic GmbH, Meerbusch, Germany. S. You, M. Mittelbronn, and M. Ronellenfitsch report no disclosures relevant to the manuscript. J. Rieger and J. Steinbach have served as consultants for Roche, the European distributor of bevacizumab (Avastin). E. Hattingen reports no disclosures relevant to the manuscript. Go to Neurology.org for full disclosures.
Received October 11, 2013. Accepted in final form April 11, 2014. REFERENCES 1. de Groot JF, Fuller G, Kumar AJ, et al. Tumor invasion after treatment of glioblastoma with bevacizumab: radiographic and pathologic correlation in humans and mice. Neuro Oncol 2010;12:233–242. 2. Iwamoto FM, Abrey LE, Beal K, et al. Patterns of relapse and prognosis after bevacizumab failure in recurrent glioblastoma. Neurology 2009;73:1200–1206. 3. Norden AD, Young GS, Setayesh K, et al. Bevacizumab for recurrent malignant gliomas: efficacy, toxicity, and patterns of recurrence. Neurology 2008;70:779–787. 4. Zuniga RM, Torcuator R, Jain R, et al. Efficacy, safety and patterns of response and recurrence in patients with recurrent high-grade gliomas treated with bevacizumab plus irinotecan. J Neurooncol 2009;91:329–336. Neurology 83
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
7
5. 6.
7.
8.
9.
10.
11.
12.
13.
14.
8
Chamberlain MC. Radiographic patterns of relapse in glioblastoma. J Neurooncol 2011;101:319–323. Ebos JM, Lee CR, Cruz-Munoz W, Bjarnason GA, Christensen JG, Kerbel RS. Accelerated metastasis after short-term treatment with a potent inhibitor of tumor angiogenesis. Cancer Cell 2009;15:232–239. Paez-Ribes M, Allen E, Hudock J, et al. Antiangiogenic therapy elicits malignant progression of tumors to increased local invasion and distant metastasis. Cancer Cell 2009;15:220–231. Pope WB, Xia Q, Paton VE, et al. Patterns of progression in patients with recurrent glioblastoma treated with bevacizumab. Neurology 2011;76:432–437. Wick A, Dorner N, Schafer N, et al. Bevacizumab does not increase the risk of remote relapse in malignant glioma. Ann Neurol 2011;69:586–592. DeLay M, Jahangiri A, Carbonell WS, et al. Microarray analysis verifies two distinct phenotypes of glioblastomas resistant to antiangiogenic therapy. Clin Cancer Res 2012; 18:2930–2942. Hu YL, DeLay M, Jahangiri A, et al. Hypoxia-induced autophagy promotes tumor cell survival and adaptation to antiangiogenic treatment in glioblastoma. Cancer Res 2012;72:1773–1783. Keunen O, Johansson M, Oudin A, et al. Anti-VEGF treatment reduces blood supply and increases tumor cell invasion in glioblastoma. Proc Natl Acad Sci USA 2011; 108:3749–3754. Rieger J, Bahr O, Muller K, Franz K, Steinbach J, Hattingen E. Bevacizumab-induced diffusion-restricted lesions in malignant glioma patients. J Neurooncol 2009;99:49–56. Rieger J, Bahr O, Ronellenfitsch MW, Steinbach J, Hattingen E. Bevacizumab-induced diffusion
Neurology 83
15.
16.
17.
18.
19.
20.
21.
restriction in patients with glioma: tumor progression or surrogate marker of hypoxia? J Clin Oncol 2010;28: e477. Mong S, Ellingson BM, Nghiemphu PL, et al. Persistent diffusion-restricted lesions in bevacizumab-treated malignant gliomas are associated with improved survival compared with matched controls. AJNR Am J Neuroradiol 2012;33:1763–1770. Gerstner ER, Frosch MP, Batchelor TT. Diffusion magnetic resonance imaging detects pathologically confirmed, nonenhancing tumor progression in a patient with recurrent glioblastoma receiving bevacizumab. J Clin Oncol 2010;28: e91–e93. Bähr O, Hattingen E, Rieger J, Steinbach JP. Bevacizumabinduced tumor calcifications as a surrogate marker of outcome in patients with glioblastoma. Neuro Oncol 2011;13: 1020–1029. Wen PY, Macdonald DR, Reardon DA, et al. Updated response assessment criteria for high-grade gliomas: Response Assessment in Neuro-Oncology Working Group. J Clin Oncol 2010;28:1963–1972. Meloan SN, Puchtler H, Valentine LS. Alkaline and acid alizarin red S stains for alkali-soluble and alkaliinsoluble calcium deposits. Arch Pathol 1972;93:190– 197. Friedman HS, Prados MD, Wen PY, et al. Bevacizumab alone and in combination with irinotecan in recurrent glioblastoma. J Clin Oncol 2009;27:4733–4740. Kunkel P, Ulbricht U, Bohlen P, et al. Inhibition of glioma angiogenesis and growth in vivo by systemic treatment with a monoclonal antibody against vascular endothelial growth factor receptor-2. Cancer Res 2001;61: 6624–6628.
July 15, 2014
ª 2014 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Sustained focal antitumor activity of bevacizumab in recurrent glioblastoma Oliver Bähr, Patrick N. Harter, Lutz M. Weise, et al. Neurology published online June 13, 2014 DOI 10.1212/WNL.0000000000000594 This information is current as of June 13, 2014 Updated Information & Services
including high resolution figures, can be found at: http://www.neurology.org/content/early/2014/06/13/WNL.00000 00000000594.full.html
Supplementary Material
Supplementary material can be found at: http://www.neurology.org/content/suppl/2014/06/13/WNL.00000 00000000594.DC1.html
Subspecialty Collections
This article, along with others on similar topics, appears in the following collection(s): All Cerebrovascular disease/Stroke http://www.neurology.org//cgi/collection/all_cerebrovascular_dis ease_stroke Risk factors in epidemiology http://www.neurology.org//cgi/collection/risk_factors_in_epidem iology Stroke prevention http://www.neurology.org//cgi/collection/stroke_prevention
Permissions & Licensing
Information about reproducing this article in parts (figures,tables) or in its entirety can be found online at: http://www.neurology.org/misc/about.xhtml#permissions
Reprints
Information about ordering reprints can be found online: http://www.neurology.org/misc/addir.xhtml#reprintsus
