J Neurooncol (2014) 116:325–331 DOI 10.1007/s11060-013-1298-9
CLINICAL STUDY
Correlation between progression free survival and dynamic susceptibility contrast MRI perfusion in WHO grade III glioma subtypes Rajiv Mangla • Daniel Thomas Ginat • Shervin Kamalian • Michael T. Milano • David N. Korones • Kevin A. Walter • Sven Ekholm
Received: 10 June 2013 / Accepted: 27 October 2013 / Published online: 1 November 2013 Ó Springer Science+Business Media New York 2013
Abstract The purpose of this study was to determine whether dynamic susceptibility contrast MR perfusion relative cerebral blood volume (rCBV) correlates with prognosis of World Health Organization (WHO) grade III glial tumors and their different subtypes. Retrospective evaluation of pretreatment tumor rCBV derived from dynamic susceptibility contrast MR perfusion was performed in 34 patients with histopathologically diagnosed WHO grade III glial tumors (anaplastic astrocytomas (n = 20), oligodendrogliomas (n = 4), and oligoastrocytomas (n = 10)). Progression free survival was correlated with rCBV using Spearman rank analysis. ROC curve analysis was performed to determine the operating point for rCBV in patients with anaplastic astrocytomas dichotomized at the median progression free survival R. Mangla S. Ekholm Department of Imaging Sciences, University of Rochester School of Medicine & Dentistry, Rochester, NY, USA D. T. Ginat (&) Department of Radiology, University of Chicago Medical Center, 5841 S Maryland Avenue, Chicago, IL 60637, USA e-mail: [email protected] S. Kamalian Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA M. T. Milano Department of Radiation Oncology, University of Rochester School of Medicine & Dentistry, Rochester, NY, USA D. N. Korones Department of Pediatric Oncology, University of Rochester School of Medicine & Dentistry, Rochester, NY, USA K. A. Walter Department of Neurosurgery, University of Rochester School of Medicine & Dentistry, Rochester, NY, USA
time. For all grade III tumors (n = 34) the mean rCBV was 2.51 with a progression free survival of 705.5 days. The mean rCBV of anaplastic astrocytomas was 2.47 with progression free survival 495.2 days. In contrast, the mean rCBV for oligodendroglial tumors was 2.56 with a progression free survival of 1005.6 days. Although there was no significant correlation between rCBV and progression free survival among all types of grade III gliomas (P = 0.12), among anaplastic astrocytomas there was a significant correlation between pretreatment rCBV and progression free survival with correlation coefficient of -0.51 (P = 0.02). The operating point for rCBV in patients with anaplastic astrocytomas dichotomized at the median progression free survival time (446.5 days) was 2.86 with 78 % accuracy and there was a significant difference between the survival of patients with anaplastic astrocytomas in the dichotomized groups (P = 0.0009). Pre-treatment rCBV may serve as a prognostic imaging biomarker for anaplastic astrocytomas, but not grade III oligodendroglioma tumors. Keywords MR perfusion Grade III astrocytoma Progression free survival
Introduction The World Health Organization (WHO) classifies gliomas into four grades, of which anaplastic astrocytoma, oligodendroglioma, and oligoastrocytoma constitute the majority of grade III brain tumors [1, 2]. MRI has become an indispensable tool for the diagnosis and treatment planning for brain tumors. Furthermore, MR perfusion can be incorporated into clinical MR examinations to assess tumor vascularity, which correlates with grading and aggressiveness of gliomas [3]. However, there is a paucity of literature
123
326
regarding the role of rCBV as a potential biomarker for progression free survival in grade III gliomas. There are many host-dependent and tumor-dependent factors that determine the prognosis and treatment approach in grade III tumors [4–7]. Histopathology is still considered the gold standard for guiding treatment and predicting survival in patients with gliomas. However, histology has limitations with regard to predicting treatment response, since there can be difficulty obtaining representative tissue samples in histologically heterogeneous tumors. Furthermore, there is considerable interobserver variability among pathologists in grading these tumors, with *30 % discordance for grade III gliomas [8]. It has been suggested that MRI-based imaging parameters, such as perfusion imaging can more accurately assess the treatment response and improve the overall management of gliomas [9–12]. However, a clinical question that remains is whether this information can aid in predicting progression free survival specifically for grade III gliomas. Thus, the aim of this study is to assess the utility of pretreatment rCBV measurements in WHO grade III gliomas for predicting survival time. The hypothesis is that grade III tumors with high rCBV are more likely to have lower progression free survival than those with low rCBV.
Materials and methods Institutional review board approval was obtained and HIPAA guidelines were followed for this retrospective study. Patients Thirty-four patients (mean age 39.6 years; age range 8–73 years; 23 males and 11 females) were recruited in the study. All cases were diagnosed as grade III gliomas on histopathology; by tumor resection (29 cases) or stereotactic biopsy (5 cases). All patients represented newly diagnosed WHO grade III gliomas that had perfusion imaging included in the pre-treatment MRI examination. The median followup was 635 days (range, 67–2188 days). There were 20 cases of AA and 14 with an OG component, including anaplastic oligoastrocytoma (n = 10) and anaplastic oligodendroglioma (n = 4). There were 15 males and 5 females in the AA group (average age 45.6 years) and 12 males and 2 females in the OG group (average age 40.2 years). There were 8 cases (40 %) that demonstrated enhancement at the time of presentation in the AA group and 5 cases (42 %) that demonstrated enhancement at the time of presentation in the OG group. In the AA group, 9 cases received both Temodar and radiation, 2 cases received Bevacizumab only, 1 case received Temodar only, 1 case received radiation, Temodar, and BCNU, 1 case received radiation and BCNU, 1 case
123
J Neurooncol (2014) 116:325–331
received radiation, Temodar, and Bevacizumab, 1 case received radiation, Temodar, Bevacizumab, and irinotecan, and 4 cases received no radiation or chemotherapy. In the OG group, 10 cases received Temodar and radiation, 1 case received only radiation, and 1 case received Temodar, radiation, and Bevacizumab. In the AA group, 3 cases underwent gross total resection, nine underwent subtotal resection, and the remainder underwent biopsy only. In the OG group, five underwent gross total resection, seven underwent subtotal resection, and the remainder underwent biopsy only. Conventional MR imaging Conventional MRI examinations, which included at least axial T1, T2-FSE, T2-FLAIR, as well as post-contrast T1weighted images were been obtained on a 1.5T scanner (GE Healthcare, Milwaukee, Wisconsin). Dynamic susceptibility contrast MR perfusion imaging Twelve axial section levels were chosen for perfusion imaging based on lesion extent as defined on the pre-contrast FLAIR images. The perfusion imaging constituted a T2*weighted echo-planar imaging sequence (TR/TE: 1500/50 ms, NEX: 1, matrix: 128 9 96; section thickness: 6 mm without gap). The first 12 acquisitions were performed before the contrast agent was injected to establish a precontrast baseline. After pre-contrast baseline, 0.15 mmol/kg of body weight gadopentetate Dimeglumine was injected with a power injector at a rate of 5 mL/sec through an 18-or 20-gauge intravenous catheter. This was immediately followed by a bolus injection of saline (20 at 5 mL/sec). The perfusion MR images covered the entire mass in all cases. Post-processing of perfusion images All imaging data from the MR perfusion images were transferred to a separate workstation for post processing and CBV computation. Post-processing was conducted at the workstation by one of the authors, blinded at the time of analysis to clinical outcome and histological data. The analysis was carried out with LUPE (Lund University, Sweden), which is written in IDL (Interactive Data Language; Research Systems, Boulder, Col) and includes leakage correction for T1-effects from blood–brain-barrier leakage, as described in Haselhorst et al. [13]. Grey scale and color CBV maps were available for identification of areas with abnormally increased perfusion. Manuallyselected circular ROI measuring 30–50 mm2 were placed on hot spots in the perfusion map to measure the mean rCBV within these small ROI (Fig. 1, 2) and the ROI with
J Neurooncol (2014) 116:325–331
highest value was chosen to represent the maximum rCBV. T1- and T2-weighted images as well as the raw perfusion images were used to ensure that the ROI did not include hemorrhage or blood vessels. For normalization, manuallyselected circular ROI measuring *30–50 mm2 were drawn in the contralateral normal-appearing white matter. The rCBV ratio was then obtained by dividing the lesion CBV by the values obtained from the contralateral normalappearing white matter. This type of measurement normalization has been shown to have the highest inter- and intra-observer reproducibility [14].
327
conventional MRI. The change in lesion size was measured on axial contrast-enhanced T1-weighted MR images by determining the bidirectional product, as described by Macdonald et al. [15]. An increase in size of the abnormal enhancement by [25 % was considered to represent progression [16]. Alternatively, in the non enhancing tumors, progression was denoted as [25 % increase in lesion size by bidirectional product on FLAIR images. In order to exclude pseudo progression, we considered progression only if the increase in lesion size or new enhancement persisted on at least two consecutive followup scans.
Follow-up All patients were followed with MRI on a regular basis after treatment to monitor for disease progression. Followup imaging was performed using the same protocol for
Statistical analysis
Fig. 1 38-year-old male with an oligodendroglioma. rCBV map at initial presentation with ROI drawn shows elevated perfusion in the tumor (a). There is mild enhancement of the tumor on the corresponding post-contrast MRI (b)
Kaplan–Meier curve analysis was performed to determine progression free survival. Spearman rank correlation analysis was performed for rCBV and progression free survival for all WHO grade III tumors. If a significant correlation was found, patients were categorized as having either survival of greater or less than median survival time for that particular group and receiver operating characteristic (ROC) curve analysis was performed to determine the operating point (maximal sum of sensitivity and specificity) for pretreatment rCBV. Univariate analysis was performed for patient age, tumor enhancement, tumor size, extent of resection, and Bevacizumab administration. Continuous and discrete variables were compared using unpaired t test and Fisher exact or Chi squared tests, respectively. In addition, multivariate analysis for rCBV and the covariate was performed using binary logistic regression. Statistical analyses were performed using MedCalc software (version 11.5.1.0, Mariakerke, Belgium) and significance was defined as a P value of \0.05.
Fig. 2 32-year-male with anaplastic astrocytoma and a progression free disease time of \2 years. rCBV map (a) at time of initial presentation shows elevated perfusion within the tumor compared to normal white matter. The corresponding axial FLAIR (b) and post-
contrast T1-weighted MRI images show a non-enhancing infiltrative mass involving the right basal ganglia and insula. Follow up postcontrast T1-weighted MRI (c) shows marked progression of disease with new enhancement
123
328
J Neurooncol (2014) 116:325–331
Results For WHO grade III tumors overall, the mean rCBV was 2.51 ± 1.37. For the AA and OG subgroups, the mean rCBV measurements were 2.47 ± 1.30 and 2.56 ± 1.52, respectively. The mean progression free survival for the entire cohort of WHO grade III tumors was 705.5 ± 521.6 days. The mean progression free survival for AA subgroup was 495.2 ± 376.8 days was significantly different (P = 0.003) from the mean progression free survival for the OG subgroup was 1005.6 ± 564.3 days. There was no significant correlation between pretreatment rCBV and progression free survival for WHO grade III tumors as a whole (P = 0.12) or the OG group (P = 0.79). However, among patients with anaplastic astrocytomas only, there was a significant inverse correlation between pretreatment rCBV and progression free survival with correlation coefficient of -0.51 (P = 0.02). The distribution of progression free survival with respect to rCBV for AA and OG tumors is shown in Fig. 3. ROC curve analysis revealed that the operating point for rCBV in patients with anaplastic astrocytomas dichotomized at the median progression free survival time (446.5 days) was 2.86 with 78 % accuracy (95 % confidence interval: 56 to 87 %) (Fig. 4). Among patients with AA, those with pretreatment rCBV of \2.86 have significantly (P = 0.0009) longer progression free survival compared to those with rCBV of[2.86, as depicted on the Kaplan–Meier curves (Fig. 5). The results of univariate analysis for the AA group are shown in Table 1. Notably, there was no significant difference in patient age, tumor enhancement, tumor size, extent of resection and Bevacizumab administration among patients with less than median of progression free survival
Fig. 4 ROC curve analysis shows that pretreatment rCBV has a high accuracy (78 %) for differentiating between AA-patients with progression free survival less than versus greater than the median progression free survival time of 446.5 days
Fig. 5 Kaplan-Meier curve analysis shows AA with pretreatment rCBV of \2.86 are associated with significantly (P = 0.0009) longer progression free survival compared to those with rCBV of [2.86
and those with greater than median of progression free survival for the AA and OG groups combined. In addition, multivariate analysis revealed that CBV (P = 0.04) was the strongest predictor of longer than median progression free survival in patients with AA.
Discussion Fig. 3 Scatter plot of progression free survival with respect to rCBV for AA and OG tumors. There is a significant correlation between progression free survival and rCBV for AA tumors (P = 0.02), but not for OG tumors (P = 0.79)
123
We retrospectively analyzed the role of rCBV in predicting the long-term prognosis after treatment in patients with histologically diagnosed WHO grade III gliomas. The salient finding was a significant inverse correlation
J Neurooncol (2014) 116:325–331
329
Table 1 Summary of covariates and univariate analysis for the AA group with progression free survival dichotomized at its median of 446.5 days Covariate
AA patients n = 20
PFS \ 446.5 days n = 10
PFS [ 446.5 days n = 10
P value
Age, mean (SD)
42.2 (17.9)
48.8 (16.6)
35.7 (17.6)
0.10
rCBV, mean (SD)
2.55 (1.27)
3.17 (1.39)
1.93 (0.78)
0.02
12/20 (60 %) 2030.5 (1226.6)
8/10 (80 %) 2449.8 (1468.4)
4/10 (40 %) 1611.2 (793.5)
0.16 0.13
Tumor enhancement, n (%) Tumor size, mean (SD) Extent of resection, n (%)
1.00
Biopsy
6/20 (30 %)
3/6 (50 %)
3/6 (50 %)
Partial
10/20 (50 %)
5/10 (50 %)
5/10 (50 %)
Total
4/20 (20 %)
2/4 (50 %)
2/4 (50 %)
Bevacizumab, n (%)
3/20 (15 %)
1/10 (10 %)
2/10 (20 %)
1.00
Continuous and discrete variables were compared using unpaired t test and Fisher exact/Chi squared tests, respectively
between rCBV and progression free survival for pure astrocytic tumors. Our results parallel the MR perfusion characteristics of low grade gliomas and suggest that pure grade III AA with rCBV values [2.86 have a worse prognosis that is comparable to the median survival of glioblastomas [9, 17]. Grade III gliomas tend to be heterogeneous and may contain regions of different degrees of malignant degeneration within the same tumor [18]. This can be difficult to detect in a biopsy specimen and can result in underestimation of grade on histopathology. However, pure astrocytic tumors may be amenable to grading based on rCBV values. This information may be helpful in predicting treatment response as well as designing individual treatment and imaging follow up paradigms for patients with a grade III astrocytic tumor. Previous studies have assessed the correlation of rCBV and long-term prognosis in patients with glioblastomas. These studies have found that stratification of these tumors, based on area/volume, show a correlation between prognosis and total tumor mass of high rCBV values. There are very few studies that specifically address rCBV and survival in grade III tumors and such studies have grouped the grade III tumor with grade IV tumors [19]. Grade II oligodendroglial tumors are also known to have significantly higher rCBV than grade II pure astrocytic tumors [20]. Nevertheless, the grade II tumors with OG components have an overall better prognosis, despite a higher rCBV [21, 22]. Although grade III tumors with oligodendroglial components generally displayed higher rCBV than pure astrocytic tumors, the difference was not significant. This could be related to the fact that many of the tumors with oligodendroglial components in our study were mixed tumors. On histopathology, neovascularization commonly occurs in low-grade oligodendrogliomas, but unlike astrocytic tumors, endothelial hyperplasia alone is not a strong
independent factor of unfavorable prognosis in OG [23, 24]. The presence of multiple ‘chicken wire’ like vessels in oligodendrogliomas is believed to be responsible for the elevated rCBV in these tumors [20]. These vessels are unlike the typical tortuous leaky vessels of high grade brain tumors, but can cause confounding effects on the grading of these tumors based on rCBV [25] Oligodendroglial tumors have generated much interest over the past decade due to their heightened response to chemotherapy. Combined loss of heterozygosity on chromosomal arms 1p and 19q has emerged as an independent predictive marker of a better response to radiotherapy and chemotherapy as well as longer survival in patients with oligodendroglial tumors [26]. Perfusion analysis of these tumors with 1p and 19q allelic deletions has also shown high rCBV values, which implies that these tumors can have good prognosis despite high rCBV [27]. This phenomenon is not entirely accounted for by the presence of VEGF, since increased VEGF has also been found in anaplastic astrocytomas [28]. Therefore, it has been suggested that angiogenic factors other than VEGF may be present in oligodendroglial tumors [29–31]. More in depth studies with survival analysis and its correlation with degree of oligodendroglial components in mixed gliomas as well as molecular analysis are necessary. A potential limitation of this study was the variability of treatment administered in this patient cohort and may be an independent variable affecting patient outcome. However, the treatment regimens did not significantly differ between the OG and AA groups and multivariate analysis indicated that treatment type was not a significant confounder in our study. There is also no convincing data demonstrating that any particular treatment regimen produces radically better outcomes than are produced by the others. Nevertheless, the treatment regimens were similar in both groups in this study. Although no pre-load of contrast was administered prior to collection of perfusion data, we used post-
123
330
processing software to account for the T1-weighted effects caused by leakage of contrast [32]. In addition, information regarding 1p19q deletion and MGMT methylation status was incomplete in this series and therefore could not be factored into the analysis.
Conclusion The preliminary results in this quantitative analysis suggest that rCBV derived from dynamic susceptibility contrastenhanced perfusion MRI can serve as a useful imaging biomarker for predicting progression free survival in anaplastic astrocytomas, but not may not be predictive for grade III tumors with oligodendroglial components. Conflict of interest None of the authors have conflicts of interests or any relevant disclosures.
References 1. Kleihues P, Cavenee WK (2000) Pathology and genetics of tumours of the nervous system. IARC, Lyon 2. Louis DN, Ohgaki H, Wiestler OD et al (2007) The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114:97–109 3. Aronen HJ, Gazit IE, Louis DN et al (1994) Cerebral blood volume maps of gliomas: comparison with tumor grade and histologic findings. Radiology 191:41–51 4. Leon SP, Folkerth RD, Black P (1996) Microvessel density is a prognostic indicator for patients with astroglial brain tumors. Cancer 77:362–372 5. Von Deimling A, Louis DN, von Ammon K et al (1992) Evidence for a tumor suppressor gene on chromosome 19q associated with human astrocytomas, oligodendrogliomas, and mixed gliomas. Cancer Res 52:4277–4279 6. Curran WJ, Scott CB, Horton J et al (1993) Recursive partitioning analysis of prognostic factors in three radiation therapy oncology group malignant glioma trials. J Natl Cancer Inst 85:704–710 7. Scott SCB, Scarantino C, Urtasun R et al (1998) Validation and predictive power of radiation therapy oncology group (RTOG) recursive partitioning analysis classes for malignant glioma patients: a report using RTOG 90-06. Int J Radiat Oncol Biol Phys 40:51–55 8. Bruner JM, Inouye L, Fuller GN, Langford LA (1997) Diagnostic discrepancies and their clinical impact in a neuropathology referral practice. Cancer 79:796–803 9. Law M, Oh S, Babb JS et al (2006) Low-grade gliomas: dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging—prediction of patient clinical response. Radiology 238:658–667 10. Oh J, Henry RG, Pirzkall A et al (2004) Survival analysis in patients with glioblastoma multiforme: predictive value of choline-to-N-acetylaspartate index, apparent diffusion coefficient, and relative cerebral blood volume. J Magn Reson Imaging 19:546–554 11. Pope WB, Sayre J, Perlina A et al (2005) MR imaging correlates of survival in patients with high-grade gliomas. Am J Neuroradiol 26:2466–2474
123
J Neurooncol (2014) 116:325–331 12. Li Y, Lupo JM, Polley MY et al (2011) Serial analysis of imaging parameters in patients with newly diagnosed glioblastoma multiforme. Neuro Oncol 13:546–557 13. Haselhorst R, Kappos L, Bilecen D et al (2000) Dynamic susceptibility contrast MR imaging of plaque development in multiple sclerosis: application of an extended blood–brain barrier leakage correction. J Magn Reson Imaging 11:495–505 14. Wetzel SG, Cha S, Johnson G et al (2002) Relative cerebral blood volume measurements in intracranial mass lesions: interobserver and intraobserver reproducibility study. Radiology 224:797–803 15. Macdonald DR, Cascino TL, Schold SC Jr, Cairncross JG (1990) Response criteria for phase II studies of supratentorial malignant glioma. J Clin Oncol 8:1277–1280 16. Law M, Young RJ, Babb JS et al (2008) Gliomas: predicting time to progression or survival with cerebral blood volume measurements at dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging. Radiology 247:490–498 17. Stark AM, Nabavi A, Mehdorn HM, Blomer U (2005) Glioblastoma multiforme: report of 267 cases treated at a single institution. Surg Neurol 63:162–169 18. Stupp R, Reni M, Gatta G, Mazza E, Vecht C (2007) Anaplastic astrocytoma in adults. Crit Rev Oncol Hematol 63:72–80 19. Hirai T, Murakami R, Nakamura H et al (2008) Prognostic value of perfusion MR imaging of high-grade astrocytomas: long-term follow-up study. AJNR Am J Neuroradiol 29:1505–1510 20. Cha S, Tihan T, Crawford F et al (2005) Differentiation of lowgrade oligodendrogliomas from low-grade astrocytomas by using quantitative blood-volume measurements derived from dynamic susceptibility contrast-enhanced MR imaging. Am J Neuroradiol 26:266–273 21. Lev MH, Ozsunar Y, Henson JW et al (2004) Glial tumor grading and outcome prediction using dynamic spin-echo MR susceptibility mapping compared with conventional contrast-enhanced MR: confounding effect of elevated rCBV of oligodendroglimoas. AJNR Am J Neuroradiol 25:214–221 22. Bisdas S, Kirkpatrick M, Giglio P et al (2009) Cerebral blood volume measurements by perfusion-weighted MR imaging in gliomas: ready for prime time in predicting short-term outcome and recurrent disease? AJNR Am J Neuroradiol 30:681–688 23. Schiffer D, Dutto A, Cavalla P et al (1997) Prognostic factors in oligodendroglioma. Can J Neurol Sci 24:313–319 24. Mork SJ, Halvorsen TB, Lindegaard KF, Eide GE (1986) Oligodendroglioma: histologic evaluation and prognosis. J Neuropathol Exp Neurol 45:65–78 25. Engelhard HH, Stelea A, Mundt A (2003) Oligodendroglioma and anaplastic oligodendroglioma: clinical features, treatment, and prognosis. Surg Neurol 60:443–456 26. Whitmore RG et al (2007) Prediction of oligodendroglial tumor subtype and grade using perfusion weighted magnetic resonance imaging. J Neurosurg 107:600–609 27. Christov C, Adle-Biassette H, Le Guerinel C et al (1998) Immunohistochemical detection of vascular endothelial growth factor (VEGF) in the vasculature of oligodendrogliomas. Neuropathol Appl Neurobiol 24:29–35 28. Lafuente JV, Ada´n B, Alkiza K, Garibi JM, Rossi M, CruzSa´nchez FF (1999) Expression of vascular endothelial growth factor (VEGF) and platelet-derived growth factor receptor-beta (PDGFR-beta) in human gliomas. J Mol Neurosci 13:177–185 29. Pietsch T, Valter MM, Wolf HK et al (1997) Expression and distribution of vascular endothelial growth factor protein in human brain tumors. Acta Neuropathol 93:109–117 30. Maia AC Jr, Malheiros SM, da Rocha AJ et al (2005) MR cerebral blood volume maps correlated with vascular endothelial growth factor expression and tumor grade in nonenhancing gliomas. AJNR Am J Neuroradiol 26:777–783
J Neurooncol (2014) 116:325–331 31. Netto GC, Bleil CB, Hilbig A et al (2008) Immunohistochemical evaluation of the microvascular density through the expression of TGF-b (CD 105/endoglin) and CD 34 receptors and expression of the vascular endothelial growth factor (VEGF) in oligodendrogliomas. Neuropathology 28:17–23
331 32. Boxerman J, Schmainda KM, Weisskoff RM (2006) Relative cerebral blood volume maps corrected for contrast agent extravasation significantly correlate with glioma tumor grade whereas uncorrected maps do not. AJNR Am J Neuroradiol 27:859–867
123
CLINICAL STUDY
Correlation between progression free survival and dynamic susceptibility contrast MRI perfusion in WHO grade III glioma subtypes Rajiv Mangla • Daniel Thomas Ginat • Shervin Kamalian • Michael T. Milano • David N. Korones • Kevin A. Walter • Sven Ekholm
Received: 10 June 2013 / Accepted: 27 October 2013 / Published online: 1 November 2013 Ó Springer Science+Business Media New York 2013
Abstract The purpose of this study was to determine whether dynamic susceptibility contrast MR perfusion relative cerebral blood volume (rCBV) correlates with prognosis of World Health Organization (WHO) grade III glial tumors and their different subtypes. Retrospective evaluation of pretreatment tumor rCBV derived from dynamic susceptibility contrast MR perfusion was performed in 34 patients with histopathologically diagnosed WHO grade III glial tumors (anaplastic astrocytomas (n = 20), oligodendrogliomas (n = 4), and oligoastrocytomas (n = 10)). Progression free survival was correlated with rCBV using Spearman rank analysis. ROC curve analysis was performed to determine the operating point for rCBV in patients with anaplastic astrocytomas dichotomized at the median progression free survival R. Mangla S. Ekholm Department of Imaging Sciences, University of Rochester School of Medicine & Dentistry, Rochester, NY, USA D. T. Ginat (&) Department of Radiology, University of Chicago Medical Center, 5841 S Maryland Avenue, Chicago, IL 60637, USA e-mail: [email protected] S. Kamalian Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA M. T. Milano Department of Radiation Oncology, University of Rochester School of Medicine & Dentistry, Rochester, NY, USA D. N. Korones Department of Pediatric Oncology, University of Rochester School of Medicine & Dentistry, Rochester, NY, USA K. A. Walter Department of Neurosurgery, University of Rochester School of Medicine & Dentistry, Rochester, NY, USA
time. For all grade III tumors (n = 34) the mean rCBV was 2.51 with a progression free survival of 705.5 days. The mean rCBV of anaplastic astrocytomas was 2.47 with progression free survival 495.2 days. In contrast, the mean rCBV for oligodendroglial tumors was 2.56 with a progression free survival of 1005.6 days. Although there was no significant correlation between rCBV and progression free survival among all types of grade III gliomas (P = 0.12), among anaplastic astrocytomas there was a significant correlation between pretreatment rCBV and progression free survival with correlation coefficient of -0.51 (P = 0.02). The operating point for rCBV in patients with anaplastic astrocytomas dichotomized at the median progression free survival time (446.5 days) was 2.86 with 78 % accuracy and there was a significant difference between the survival of patients with anaplastic astrocytomas in the dichotomized groups (P = 0.0009). Pre-treatment rCBV may serve as a prognostic imaging biomarker for anaplastic astrocytomas, but not grade III oligodendroglioma tumors. Keywords MR perfusion Grade III astrocytoma Progression free survival
Introduction The World Health Organization (WHO) classifies gliomas into four grades, of which anaplastic astrocytoma, oligodendroglioma, and oligoastrocytoma constitute the majority of grade III brain tumors [1, 2]. MRI has become an indispensable tool for the diagnosis and treatment planning for brain tumors. Furthermore, MR perfusion can be incorporated into clinical MR examinations to assess tumor vascularity, which correlates with grading and aggressiveness of gliomas [3]. However, there is a paucity of literature
123
326
regarding the role of rCBV as a potential biomarker for progression free survival in grade III gliomas. There are many host-dependent and tumor-dependent factors that determine the prognosis and treatment approach in grade III tumors [4–7]. Histopathology is still considered the gold standard for guiding treatment and predicting survival in patients with gliomas. However, histology has limitations with regard to predicting treatment response, since there can be difficulty obtaining representative tissue samples in histologically heterogeneous tumors. Furthermore, there is considerable interobserver variability among pathologists in grading these tumors, with *30 % discordance for grade III gliomas [8]. It has been suggested that MRI-based imaging parameters, such as perfusion imaging can more accurately assess the treatment response and improve the overall management of gliomas [9–12]. However, a clinical question that remains is whether this information can aid in predicting progression free survival specifically for grade III gliomas. Thus, the aim of this study is to assess the utility of pretreatment rCBV measurements in WHO grade III gliomas for predicting survival time. The hypothesis is that grade III tumors with high rCBV are more likely to have lower progression free survival than those with low rCBV.
Materials and methods Institutional review board approval was obtained and HIPAA guidelines were followed for this retrospective study. Patients Thirty-four patients (mean age 39.6 years; age range 8–73 years; 23 males and 11 females) were recruited in the study. All cases were diagnosed as grade III gliomas on histopathology; by tumor resection (29 cases) or stereotactic biopsy (5 cases). All patients represented newly diagnosed WHO grade III gliomas that had perfusion imaging included in the pre-treatment MRI examination. The median followup was 635 days (range, 67–2188 days). There were 20 cases of AA and 14 with an OG component, including anaplastic oligoastrocytoma (n = 10) and anaplastic oligodendroglioma (n = 4). There were 15 males and 5 females in the AA group (average age 45.6 years) and 12 males and 2 females in the OG group (average age 40.2 years). There were 8 cases (40 %) that demonstrated enhancement at the time of presentation in the AA group and 5 cases (42 %) that demonstrated enhancement at the time of presentation in the OG group. In the AA group, 9 cases received both Temodar and radiation, 2 cases received Bevacizumab only, 1 case received Temodar only, 1 case received radiation, Temodar, and BCNU, 1 case received radiation and BCNU, 1 case
123
J Neurooncol (2014) 116:325–331
received radiation, Temodar, and Bevacizumab, 1 case received radiation, Temodar, Bevacizumab, and irinotecan, and 4 cases received no radiation or chemotherapy. In the OG group, 10 cases received Temodar and radiation, 1 case received only radiation, and 1 case received Temodar, radiation, and Bevacizumab. In the AA group, 3 cases underwent gross total resection, nine underwent subtotal resection, and the remainder underwent biopsy only. In the OG group, five underwent gross total resection, seven underwent subtotal resection, and the remainder underwent biopsy only. Conventional MR imaging Conventional MRI examinations, which included at least axial T1, T2-FSE, T2-FLAIR, as well as post-contrast T1weighted images were been obtained on a 1.5T scanner (GE Healthcare, Milwaukee, Wisconsin). Dynamic susceptibility contrast MR perfusion imaging Twelve axial section levels were chosen for perfusion imaging based on lesion extent as defined on the pre-contrast FLAIR images. The perfusion imaging constituted a T2*weighted echo-planar imaging sequence (TR/TE: 1500/50 ms, NEX: 1, matrix: 128 9 96; section thickness: 6 mm without gap). The first 12 acquisitions were performed before the contrast agent was injected to establish a precontrast baseline. After pre-contrast baseline, 0.15 mmol/kg of body weight gadopentetate Dimeglumine was injected with a power injector at a rate of 5 mL/sec through an 18-or 20-gauge intravenous catheter. This was immediately followed by a bolus injection of saline (20 at 5 mL/sec). The perfusion MR images covered the entire mass in all cases. Post-processing of perfusion images All imaging data from the MR perfusion images were transferred to a separate workstation for post processing and CBV computation. Post-processing was conducted at the workstation by one of the authors, blinded at the time of analysis to clinical outcome and histological data. The analysis was carried out with LUPE (Lund University, Sweden), which is written in IDL (Interactive Data Language; Research Systems, Boulder, Col) and includes leakage correction for T1-effects from blood–brain-barrier leakage, as described in Haselhorst et al. [13]. Grey scale and color CBV maps were available for identification of areas with abnormally increased perfusion. Manuallyselected circular ROI measuring 30–50 mm2 were placed on hot spots in the perfusion map to measure the mean rCBV within these small ROI (Fig. 1, 2) and the ROI with
J Neurooncol (2014) 116:325–331
highest value was chosen to represent the maximum rCBV. T1- and T2-weighted images as well as the raw perfusion images were used to ensure that the ROI did not include hemorrhage or blood vessels. For normalization, manuallyselected circular ROI measuring *30–50 mm2 were drawn in the contralateral normal-appearing white matter. The rCBV ratio was then obtained by dividing the lesion CBV by the values obtained from the contralateral normalappearing white matter. This type of measurement normalization has been shown to have the highest inter- and intra-observer reproducibility [14].
327
conventional MRI. The change in lesion size was measured on axial contrast-enhanced T1-weighted MR images by determining the bidirectional product, as described by Macdonald et al. [15]. An increase in size of the abnormal enhancement by [25 % was considered to represent progression [16]. Alternatively, in the non enhancing tumors, progression was denoted as [25 % increase in lesion size by bidirectional product on FLAIR images. In order to exclude pseudo progression, we considered progression only if the increase in lesion size or new enhancement persisted on at least two consecutive followup scans.
Follow-up All patients were followed with MRI on a regular basis after treatment to monitor for disease progression. Followup imaging was performed using the same protocol for
Statistical analysis
Fig. 1 38-year-old male with an oligodendroglioma. rCBV map at initial presentation with ROI drawn shows elevated perfusion in the tumor (a). There is mild enhancement of the tumor on the corresponding post-contrast MRI (b)
Kaplan–Meier curve analysis was performed to determine progression free survival. Spearman rank correlation analysis was performed for rCBV and progression free survival for all WHO grade III tumors. If a significant correlation was found, patients were categorized as having either survival of greater or less than median survival time for that particular group and receiver operating characteristic (ROC) curve analysis was performed to determine the operating point (maximal sum of sensitivity and specificity) for pretreatment rCBV. Univariate analysis was performed for patient age, tumor enhancement, tumor size, extent of resection, and Bevacizumab administration. Continuous and discrete variables were compared using unpaired t test and Fisher exact or Chi squared tests, respectively. In addition, multivariate analysis for rCBV and the covariate was performed using binary logistic regression. Statistical analyses were performed using MedCalc software (version 11.5.1.0, Mariakerke, Belgium) and significance was defined as a P value of \0.05.
Fig. 2 32-year-male with anaplastic astrocytoma and a progression free disease time of \2 years. rCBV map (a) at time of initial presentation shows elevated perfusion within the tumor compared to normal white matter. The corresponding axial FLAIR (b) and post-
contrast T1-weighted MRI images show a non-enhancing infiltrative mass involving the right basal ganglia and insula. Follow up postcontrast T1-weighted MRI (c) shows marked progression of disease with new enhancement
123
328
J Neurooncol (2014) 116:325–331
Results For WHO grade III tumors overall, the mean rCBV was 2.51 ± 1.37. For the AA and OG subgroups, the mean rCBV measurements were 2.47 ± 1.30 and 2.56 ± 1.52, respectively. The mean progression free survival for the entire cohort of WHO grade III tumors was 705.5 ± 521.6 days. The mean progression free survival for AA subgroup was 495.2 ± 376.8 days was significantly different (P = 0.003) from the mean progression free survival for the OG subgroup was 1005.6 ± 564.3 days. There was no significant correlation between pretreatment rCBV and progression free survival for WHO grade III tumors as a whole (P = 0.12) or the OG group (P = 0.79). However, among patients with anaplastic astrocytomas only, there was a significant inverse correlation between pretreatment rCBV and progression free survival with correlation coefficient of -0.51 (P = 0.02). The distribution of progression free survival with respect to rCBV for AA and OG tumors is shown in Fig. 3. ROC curve analysis revealed that the operating point for rCBV in patients with anaplastic astrocytomas dichotomized at the median progression free survival time (446.5 days) was 2.86 with 78 % accuracy (95 % confidence interval: 56 to 87 %) (Fig. 4). Among patients with AA, those with pretreatment rCBV of \2.86 have significantly (P = 0.0009) longer progression free survival compared to those with rCBV of[2.86, as depicted on the Kaplan–Meier curves (Fig. 5). The results of univariate analysis for the AA group are shown in Table 1. Notably, there was no significant difference in patient age, tumor enhancement, tumor size, extent of resection and Bevacizumab administration among patients with less than median of progression free survival
Fig. 4 ROC curve analysis shows that pretreatment rCBV has a high accuracy (78 %) for differentiating between AA-patients with progression free survival less than versus greater than the median progression free survival time of 446.5 days
Fig. 5 Kaplan-Meier curve analysis shows AA with pretreatment rCBV of \2.86 are associated with significantly (P = 0.0009) longer progression free survival compared to those with rCBV of [2.86
and those with greater than median of progression free survival for the AA and OG groups combined. In addition, multivariate analysis revealed that CBV (P = 0.04) was the strongest predictor of longer than median progression free survival in patients with AA.
Discussion Fig. 3 Scatter plot of progression free survival with respect to rCBV for AA and OG tumors. There is a significant correlation between progression free survival and rCBV for AA tumors (P = 0.02), but not for OG tumors (P = 0.79)
123
We retrospectively analyzed the role of rCBV in predicting the long-term prognosis after treatment in patients with histologically diagnosed WHO grade III gliomas. The salient finding was a significant inverse correlation
J Neurooncol (2014) 116:325–331
329
Table 1 Summary of covariates and univariate analysis for the AA group with progression free survival dichotomized at its median of 446.5 days Covariate
AA patients n = 20
PFS \ 446.5 days n = 10
PFS [ 446.5 days n = 10
P value
Age, mean (SD)
42.2 (17.9)
48.8 (16.6)
35.7 (17.6)
0.10
rCBV, mean (SD)
2.55 (1.27)
3.17 (1.39)
1.93 (0.78)
0.02
12/20 (60 %) 2030.5 (1226.6)
8/10 (80 %) 2449.8 (1468.4)
4/10 (40 %) 1611.2 (793.5)
0.16 0.13
Tumor enhancement, n (%) Tumor size, mean (SD) Extent of resection, n (%)
1.00
Biopsy
6/20 (30 %)
3/6 (50 %)
3/6 (50 %)
Partial
10/20 (50 %)
5/10 (50 %)
5/10 (50 %)
Total
4/20 (20 %)
2/4 (50 %)
2/4 (50 %)
Bevacizumab, n (%)
3/20 (15 %)
1/10 (10 %)
2/10 (20 %)
1.00
Continuous and discrete variables were compared using unpaired t test and Fisher exact/Chi squared tests, respectively
between rCBV and progression free survival for pure astrocytic tumors. Our results parallel the MR perfusion characteristics of low grade gliomas and suggest that pure grade III AA with rCBV values [2.86 have a worse prognosis that is comparable to the median survival of glioblastomas [9, 17]. Grade III gliomas tend to be heterogeneous and may contain regions of different degrees of malignant degeneration within the same tumor [18]. This can be difficult to detect in a biopsy specimen and can result in underestimation of grade on histopathology. However, pure astrocytic tumors may be amenable to grading based on rCBV values. This information may be helpful in predicting treatment response as well as designing individual treatment and imaging follow up paradigms for patients with a grade III astrocytic tumor. Previous studies have assessed the correlation of rCBV and long-term prognosis in patients with glioblastomas. These studies have found that stratification of these tumors, based on area/volume, show a correlation between prognosis and total tumor mass of high rCBV values. There are very few studies that specifically address rCBV and survival in grade III tumors and such studies have grouped the grade III tumor with grade IV tumors [19]. Grade II oligodendroglial tumors are also known to have significantly higher rCBV than grade II pure astrocytic tumors [20]. Nevertheless, the grade II tumors with OG components have an overall better prognosis, despite a higher rCBV [21, 22]. Although grade III tumors with oligodendroglial components generally displayed higher rCBV than pure astrocytic tumors, the difference was not significant. This could be related to the fact that many of the tumors with oligodendroglial components in our study were mixed tumors. On histopathology, neovascularization commonly occurs in low-grade oligodendrogliomas, but unlike astrocytic tumors, endothelial hyperplasia alone is not a strong
independent factor of unfavorable prognosis in OG [23, 24]. The presence of multiple ‘chicken wire’ like vessels in oligodendrogliomas is believed to be responsible for the elevated rCBV in these tumors [20]. These vessels are unlike the typical tortuous leaky vessels of high grade brain tumors, but can cause confounding effects on the grading of these tumors based on rCBV [25] Oligodendroglial tumors have generated much interest over the past decade due to their heightened response to chemotherapy. Combined loss of heterozygosity on chromosomal arms 1p and 19q has emerged as an independent predictive marker of a better response to radiotherapy and chemotherapy as well as longer survival in patients with oligodendroglial tumors [26]. Perfusion analysis of these tumors with 1p and 19q allelic deletions has also shown high rCBV values, which implies that these tumors can have good prognosis despite high rCBV [27]. This phenomenon is not entirely accounted for by the presence of VEGF, since increased VEGF has also been found in anaplastic astrocytomas [28]. Therefore, it has been suggested that angiogenic factors other than VEGF may be present in oligodendroglial tumors [29–31]. More in depth studies with survival analysis and its correlation with degree of oligodendroglial components in mixed gliomas as well as molecular analysis are necessary. A potential limitation of this study was the variability of treatment administered in this patient cohort and may be an independent variable affecting patient outcome. However, the treatment regimens did not significantly differ between the OG and AA groups and multivariate analysis indicated that treatment type was not a significant confounder in our study. There is also no convincing data demonstrating that any particular treatment regimen produces radically better outcomes than are produced by the others. Nevertheless, the treatment regimens were similar in both groups in this study. Although no pre-load of contrast was administered prior to collection of perfusion data, we used post-
123
330
processing software to account for the T1-weighted effects caused by leakage of contrast [32]. In addition, information regarding 1p19q deletion and MGMT methylation status was incomplete in this series and therefore could not be factored into the analysis.
Conclusion The preliminary results in this quantitative analysis suggest that rCBV derived from dynamic susceptibility contrastenhanced perfusion MRI can serve as a useful imaging biomarker for predicting progression free survival in anaplastic astrocytomas, but not may not be predictive for grade III tumors with oligodendroglial components. Conflict of interest None of the authors have conflicts of interests or any relevant disclosures.
References 1. Kleihues P, Cavenee WK (2000) Pathology and genetics of tumours of the nervous system. IARC, Lyon 2. Louis DN, Ohgaki H, Wiestler OD et al (2007) The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114:97–109 3. Aronen HJ, Gazit IE, Louis DN et al (1994) Cerebral blood volume maps of gliomas: comparison with tumor grade and histologic findings. Radiology 191:41–51 4. Leon SP, Folkerth RD, Black P (1996) Microvessel density is a prognostic indicator for patients with astroglial brain tumors. Cancer 77:362–372 5. Von Deimling A, Louis DN, von Ammon K et al (1992) Evidence for a tumor suppressor gene on chromosome 19q associated with human astrocytomas, oligodendrogliomas, and mixed gliomas. Cancer Res 52:4277–4279 6. Curran WJ, Scott CB, Horton J et al (1993) Recursive partitioning analysis of prognostic factors in three radiation therapy oncology group malignant glioma trials. J Natl Cancer Inst 85:704–710 7. Scott SCB, Scarantino C, Urtasun R et al (1998) Validation and predictive power of radiation therapy oncology group (RTOG) recursive partitioning analysis classes for malignant glioma patients: a report using RTOG 90-06. Int J Radiat Oncol Biol Phys 40:51–55 8. Bruner JM, Inouye L, Fuller GN, Langford LA (1997) Diagnostic discrepancies and their clinical impact in a neuropathology referral practice. Cancer 79:796–803 9. Law M, Oh S, Babb JS et al (2006) Low-grade gliomas: dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging—prediction of patient clinical response. Radiology 238:658–667 10. Oh J, Henry RG, Pirzkall A et al (2004) Survival analysis in patients with glioblastoma multiforme: predictive value of choline-to-N-acetylaspartate index, apparent diffusion coefficient, and relative cerebral blood volume. J Magn Reson Imaging 19:546–554 11. Pope WB, Sayre J, Perlina A et al (2005) MR imaging correlates of survival in patients with high-grade gliomas. Am J Neuroradiol 26:2466–2474
123
J Neurooncol (2014) 116:325–331 12. Li Y, Lupo JM, Polley MY et al (2011) Serial analysis of imaging parameters in patients with newly diagnosed glioblastoma multiforme. Neuro Oncol 13:546–557 13. Haselhorst R, Kappos L, Bilecen D et al (2000) Dynamic susceptibility contrast MR imaging of plaque development in multiple sclerosis: application of an extended blood–brain barrier leakage correction. J Magn Reson Imaging 11:495–505 14. Wetzel SG, Cha S, Johnson G et al (2002) Relative cerebral blood volume measurements in intracranial mass lesions: interobserver and intraobserver reproducibility study. Radiology 224:797–803 15. Macdonald DR, Cascino TL, Schold SC Jr, Cairncross JG (1990) Response criteria for phase II studies of supratentorial malignant glioma. J Clin Oncol 8:1277–1280 16. Law M, Young RJ, Babb JS et al (2008) Gliomas: predicting time to progression or survival with cerebral blood volume measurements at dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging. Radiology 247:490–498 17. Stark AM, Nabavi A, Mehdorn HM, Blomer U (2005) Glioblastoma multiforme: report of 267 cases treated at a single institution. Surg Neurol 63:162–169 18. Stupp R, Reni M, Gatta G, Mazza E, Vecht C (2007) Anaplastic astrocytoma in adults. Crit Rev Oncol Hematol 63:72–80 19. Hirai T, Murakami R, Nakamura H et al (2008) Prognostic value of perfusion MR imaging of high-grade astrocytomas: long-term follow-up study. AJNR Am J Neuroradiol 29:1505–1510 20. Cha S, Tihan T, Crawford F et al (2005) Differentiation of lowgrade oligodendrogliomas from low-grade astrocytomas by using quantitative blood-volume measurements derived from dynamic susceptibility contrast-enhanced MR imaging. Am J Neuroradiol 26:266–273 21. Lev MH, Ozsunar Y, Henson JW et al (2004) Glial tumor grading and outcome prediction using dynamic spin-echo MR susceptibility mapping compared with conventional contrast-enhanced MR: confounding effect of elevated rCBV of oligodendroglimoas. AJNR Am J Neuroradiol 25:214–221 22. Bisdas S, Kirkpatrick M, Giglio P et al (2009) Cerebral blood volume measurements by perfusion-weighted MR imaging in gliomas: ready for prime time in predicting short-term outcome and recurrent disease? AJNR Am J Neuroradiol 30:681–688 23. Schiffer D, Dutto A, Cavalla P et al (1997) Prognostic factors in oligodendroglioma. Can J Neurol Sci 24:313–319 24. Mork SJ, Halvorsen TB, Lindegaard KF, Eide GE (1986) Oligodendroglioma: histologic evaluation and prognosis. J Neuropathol Exp Neurol 45:65–78 25. Engelhard HH, Stelea A, Mundt A (2003) Oligodendroglioma and anaplastic oligodendroglioma: clinical features, treatment, and prognosis. Surg Neurol 60:443–456 26. Whitmore RG et al (2007) Prediction of oligodendroglial tumor subtype and grade using perfusion weighted magnetic resonance imaging. J Neurosurg 107:600–609 27. Christov C, Adle-Biassette H, Le Guerinel C et al (1998) Immunohistochemical detection of vascular endothelial growth factor (VEGF) in the vasculature of oligodendrogliomas. Neuropathol Appl Neurobiol 24:29–35 28. Lafuente JV, Ada´n B, Alkiza K, Garibi JM, Rossi M, CruzSa´nchez FF (1999) Expression of vascular endothelial growth factor (VEGF) and platelet-derived growth factor receptor-beta (PDGFR-beta) in human gliomas. J Mol Neurosci 13:177–185 29. Pietsch T, Valter MM, Wolf HK et al (1997) Expression and distribution of vascular endothelial growth factor protein in human brain tumors. Acta Neuropathol 93:109–117 30. Maia AC Jr, Malheiros SM, da Rocha AJ et al (2005) MR cerebral blood volume maps correlated with vascular endothelial growth factor expression and tumor grade in nonenhancing gliomas. AJNR Am J Neuroradiol 26:777–783
J Neurooncol (2014) 116:325–331 31. Netto GC, Bleil CB, Hilbig A et al (2008) Immunohistochemical evaluation of the microvascular density through the expression of TGF-b (CD 105/endoglin) and CD 34 receptors and expression of the vascular endothelial growth factor (VEGF) in oligodendrogliomas. Neuropathology 28:17–23
331 32. Boxerman J, Schmainda KM, Weisskoff RM (2006) Relative cerebral blood volume maps corrected for contrast agent extravasation significantly correlate with glioma tumor grade whereas uncorrected maps do not. AJNR Am J Neuroradiol 27:859–867
123
