- Email: [email protected]
Perfusion Computed Tomography Parameters Are Useful for Differentiating Glioblastoma, Lymphoma, and Metastasis
Original Article
Perfusion Computed Tomography Parameters Are Useful for Differentiating Glioblastoma, Lymphoma, and Metastasis Shumpei Onishi1,5, Yoshinori Kajiwara6, Takeshi Takayasu1, Manish Kolakshyapati1, Minoru Ishifuro3, Vishwa Jeet Amatya2, Yukio Takeshima2, Kazuhiko Sugiyama4, Kaoru Kurisu1, Fumiyuki Yamasaki1
-
BACKGROUND: Perfusion computed tomography (PCT) reflects blood flow and capillary condition, which is valuable in assessing brain tumors. We evaluated PCT parameters at the tumor (t) and peritumoral (p) region to differentiate malignant brain tumors.
The combination of rCBVt and rPSt could differentiate GBM from other tumors with sensitivity and specificity of 81.8% and 94.1%. The combination of rCBFp and rMTTp could differentiate METS from other tumors with sensitivity and specificity of 90.9% and 82.1%.
METHODS: We performed PCT in 39 patients with supratentorial malignant brain tumors (22 glioblastomas, 6 lymphomas, 11 metastases). Regions of interests were placed manually at tumor, peritumoral region, and contralateral normal-appearing white matter. Cerebral blood volume (CBV), cerebral blood flow (CBF), mean transit time (MTT), and permeability surface (PS) were measured. These parameters were divided by those of contralateral normal-appearing white matter to normalize at tumor (relative [r]CBVt, rCBFt, rMTTt, rPSt) and peritumoral regions (rCBVp, rCBFp, rMTTp, rPSp). The parameters were evaluated with ManneWhitney U test and receiver operating characteristics analyses. Stepwise analyses also were performed to select useful PCT parameters for differentiating these tumors.
-
-
RESULTS: The rCBFt and rCBVt of glioblastoma (GBM) were greater than those of primary central nervous system lymphoma (PCNSL) (P [ 0.0005, 0.0002) and brain metastasis (METS) (P [ 0.0044, 0.0028). The rMTTp of METS was greater than that of GBM and PCNSL (P [ 0.0001, 0.0007).
-
Key words Brain metastasis - Glioblastoma - Llymphoma - Perfusion CT - Primary central nervous system -
Abbreviations and Acronyms CBF: Cerebral blood flow CBV: Cerebral blood volume CT: Computed tomography GBM: Glioblastoma METS: Brain metastasis MTT: Mean transit time p: Peritumoral region PCNSL: Primary central nervous system lymphoma PCT: Perfusion computed tomography PS: Permeability surface
WORLD NEUROSURGERY -: e1-e8, - 2018
CONCLUSIONS: Our study introduces and supports the usefulness of PCT parameters for differentiation among GBM, PCNSL, and METS. rCBVt and rPSt may be the best predictors of GBM. rCBFp and rMTTp may be the best predictors of METS.
INTRODUCTION
P
reoperative differentiation among glioblastoma (GBM), primary central nervous system lymphoma (PCNSL), and brain metastasis (METS) can be challenging. These tumors often present as contrast-enhancing on conventional computed tomography (CT), which is a sometimes-indiscernible radiographic characteristic. Preoperative differentiation among these tumors is of high clinical relevance, as the treatment of these tumors is entirely different. GBM requires maximum resection followed by intensive chemo- and radiotherapy,1 PCNSL only requires biopsy followed by chemotherapy and sometimes with radiotherapy,2 and METS requires stereotactic irradiation or systemic treatment.3
r: Relative ROC: Receiver operating characteristic t: Tumor From the Departments of 1Neurosurgery and 2Pathology, Graduate School of Biomedical and Health Sciences, Hiroshima University, Hiroshima; Departments of 3Diagnostic Imaging and 4 Clinical Oncology and Neuro-oncology Program, Hiroshima University Hospital, Hiroshima; 5 Department of Neurosurgery, Higashihiroshima Medical Center, Hiroshima; and 6 Department of Neurosurgery, Matsuyama Red Cross Hospital, Ehime, Japan To whom correspondence should be addressed: Fumiyuki Yamasaki, M.D., Ph.D. [E-mail: [email protected]] Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.07.291 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
www.WORLDNEUROSURGERY.org
e1
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Figure 1. Box plots demonstrate perfusion computed tomography parameters of glioblastoma (GBM), primary central nervous system lymphoma (PCNSL), and brain metastasis (METS) at tumor and peritumoral regions. rCBFt, relative cerebral blood volume, tumor; rCBVt, relative cerebral blood volume, tumor; rMTTt, relative mean transit time,
tumor; rPSt, relative permeability surface, tumor; rCBFp, relative cerebral blood volume, peritumoral region; rCBVp, relative cerebral blood volume, peritumoral region; rMTTp, relative mean transit time, peritumoral region; rPSp, relative permeability surface, peritumoral region.
Perfusion imaging by CT and magnetic resonance imaging has been used to visualize the ischemic changes in brain tissue and the hemodynamics and vascularity of brain tumors.4-6 The usefulness of perfusion computed tomography (PCT) has been reported in grading gliomas,4,7 differentiating tumor progression from treatment-induced effects,8 and predicting prognosis after treatment.9 However, the integrated interpretation of PCT parameters for differentiating malignant tumors is unclear. We evaluated PCT parameters at tumor and peritumoral regions to differentiate GBM, METS, and PCNSL.
MATERIALS AND METHODS Patients This retrospective study was approved by the institutional review board of Hiroshima University Hospital. We included 39 patients with histologically proven supratentorial malignant brain tumors treated at Hiroshima University Hospital from August 2004 to August 2009. The patients included 19 men and 20 women, ranging between 25 and 83 years (mean age 60.6 years). For all 39 patients, the histologic diagnosis was confirmed by a neuropathologist. This study group consisted of 22 patients (11 men and 11 women) with GBM aged 25e83 years (mean age 60.6 years), 6 patients (4 men and 2 women) with PCNSL aged 43e76 years (mean age 65.3 years), and 11 patients (4 men and 7 women) with METS aged 34e78 years (mean age 61.8 years). All patients underwent PCT before any neurosurgical treatment.
e2
www.SCIENCEDIRECT.com
Figure 2. Receiver operating characteristic analysis for differentiating glioblastoma from other tumors. The analysis showed performances with relative cerebral blood volume, tumor (rCBFt; area under the curve [AUC] ¼ 0.8476), relative cerebral blood volume, peritumoral region (rCBFp; AUC ¼ 0.7955), relative cerebral blood volume, tumor (rCBVt; AUC ¼ 0.8636), relative cerebral blood volume, peritumoral region (rCBVp; AUC ¼ 0.7513), relative mean transit time, tumor (rMTTt; AUC ¼ 0.6511), relative mean transit time, peritumoral region (rMTTp; AUC ¼ 0.7299), relative permeability surface, tumor (rPSt; AUC ¼ 0. 6257), relative permeability surface, peritumoral region (rPSp; AUC ¼ 0.488).
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.291
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Table 1. Comparison of PCT Parameters to Differentiate GBM From Other Tumors Parameters
AUC
95% CI
Cut-Off Value
Sensitivity
Specificity
P Value
rCBFt
0.8476
(0.6645e0.9398)
2.31
0.9545
0.6471
0.0163*
rCBFp
0.7955
(0.6645e0.9398)
0.59
0.8182
0.8235
0.0316*
rCBVt
0.8636
(0.6699e0.9518)
3.65
0.8182
0.8824
0.0067*
rCBVp
0.7513
(0.6699e0.9518)
0.7
0.7727
0.7647
0.0352*
rMTTt
0.6511
(0.4643e0.8007)
0.9
0.5
0.8235
0.1692
rMTTp
0.7299
(0.4634e0.8007)
1.44
0.8636
0.5882
0.109
rPSt
0.6257
(0.4359e0.7833)
11.75
0.5909
0.7059
0.6665
rPSp
0.488
(0.4359e0.7833)
1.76
0.3182
0.8235
0.5363
PCT, perfusion computed tomography; GBM, glioblastoma; AUC, area under the curve; 95% CI, 95% confidence interval; rCBFt, relative cerebral blood volume, tumor; rCBFp, relative cerebral blood volume, peritumoral region; rCBVt, relative cerebral blood volume, tumor; rCBVp, relative cerebral blood volume, peritumoral region; rMTTt, relative mean transit time, tumor; rMTTp, relative mean transit time, peritumoral region; rPSt, relative permeability surface, tumor; rPSp, relative permeability surface, peritumoral region. *Indicates statistical significance.
Image Acquisition Perfusion studies were performed with a 16-slice helical CT scanner (LightSpeed Ultra; General Electric Medical Systems, Chicago, Illinois, USA). A noncontrast CT scan was obtained to localize the region of interest before we obtained a perfusion scan. For the perfusion studies, 40 mL of nonionic contrast medium (350 mg/mL) was injected at a rate of 5 mL/s through a 20-gauge intravenous line. At 5 seconds into the injection, a helical scan was initiated with the following technique: 80 kVp, 80 mA, and 1 second per rotation. After the initial 5-second cine scan, 8 more
Figure 3. Receiver operating characteristic analysis for differentiating glioblastoma (GBM) from other tumors with relative cerebral blood volume, tumor and relative permeability surface, tumor. These parameters could differentiate GBM with sensitivity and specificity of 81.8% and 94.1%.
WORLD NEUROSURGERY -: e1-e8, - 2018
axial images were acquired, 1 image every 2 seconds for an additional 52 seconds, thereby taking a total acquisition time of 57 seconds. Perfusion maps of cerebral blood volume (CBV), cerebral blood flow (CBF), mean transit time (MTT), and permeability surface (PS) were generated with an Advantage Windows Workstation (Ver. 4.2; GE Medical Systems) using CT Perfusion 3
Figure 4. Receiver operating characteristic analysis for differentiating brain metastasis from other tumors. The analysis showed the best diagnostic performance with relative cerebral blood volume, tumor (rCBFt; area under the curve [AUC] ¼ 0.7159), relative cerebral blood volume, peritumoral region (rCBFp; AUC ¼ 0.8571), relative cerebral blood volume, tumor (rCBVt; AUC ¼ 0.7338), relative cerebral blood volume, peritumoral region (rCBVp; AUC ¼ 0.789), relative mean transit time, tumor (rMTTt; AUC ¼ 0.6623), relative mean transit time, peritumoral region (rMTTp; AUC ¼ 0.901), relative permeability surface, tumor (rPSt; AUC ¼ 0.6688), relative permeability surface, peritumoral region (rPSp; AUC ¼ 0.5422).
www.WORLDNEUROSURGERY.org
e3
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Table 2. Comparison of PCT Parameters to Differentiate MET From Other Tumors Parameters rCBFt
AUC
95% CI
Cutoff Value
Sensitivity
Specificity
P Value
0.7159
(0.4912e0.8680)
3.1
0.8182
0.6429
0.1796
rCBFp
0.8571
(0.6974e0.9398)
0.58
1
0.75
0.0083*
rCBVt
0.7338
(0.4962e0.8852)
2.83
0.8182
0.6786
0.1405
rCBVp
0.789
(0.6061e0.9008)
0.69
0.8182
0.6786
0.0177*
rMTTt
0.6623
(0.4548e0.8218)
1.63
0.3636
92.86
0.1054
rMTTp
0.901
(0.7377e0.9671)
1.46
0.9091
0.8929
0.0241*
rPSt
0.6688
(0.4617e0.8263)
13.03
0.8182
0.5714
0.3415
rPSp
0.5422
(0.3521e0.7208)
1.66
1
0.3571
0.2528
PCT, perfusion computed tomography; MET, brain metastasis; AUC, area under the curve; 95% CI, 95% confidence interval; rCBFt, relative cerebral blood volume, tumor; rCBFp, relative cerebral blood volume, peritumoral region; rCBVt, relative cerebral blood volume, tumor; rCBVp, relative cerebral blood volume, peritumoral region; rMTTt, relative mean transit time, tumor; rMTTp, relative mean transit time, peritumoral region; rPSt, relative permeability surface, tumor; rPSp, relative permeability surface, peritumoral region. *Indicates statistical significance.
software (GE Medical System) in all patients. Regions of interest were placed manually at the enhanced tumor, peritumoral region, and contralateral normal-appearing white matter confirmed on contrast-enhanced and noncontrast CT images. Peritumoral region were defined as low density around enhanced tumor on CT image. The PCT parameters were divided by those of contralateral normal-appearing white matter to normalize the parameters at tumor (rCBVt, rCBFt, rMTTt, rPSt) and peritumoral region (rCBVp, rCBFp, rMTTp, rPSp), where r indicates relative; t indicates tumor, and p indicates peritumoral region.
Histologic Study Tumor specimens after surgical resection or biopsy were fixed in 10% phosphate-buffered formalin and embedded in paraffin blocks. Representative slides were then stained with hematoxylineeosin reagent for standard histologic diagnosis. Statistical Analysis Statistical analyses were performed with GraphPad Prism, version7.00 for Mac (GraphPad Software, San Diego California, USA) and JMP Pro, version 13.0 (SAS institute, Cary, North Carolina, USA). GBM, METS, and PCNSL were compared at tumor (t) and peritumoral (p) region respectively by using normalized PCT parameters (rCBF, rCBV, rMTT, rPS). These parameters were evaluated with the ManneWhitney U test and receiver operating characteristics (ROC) analyses; statistical significance was assigned when P < 0.05. Stepwise analyses were also performed to select useful PCT parameters for differentiating these tumors. RESULTS
Figure 5. Receiver operating characteristic analysis for differentiating brain metastasis from other tumors with relative cerebral blood volume, peritumoral region and relative mean transit time, peritumoral region. These parameters could differentiate glioblastoma with sensitivity and specificity of 90.9% and 82.1%, respectively.
e4
www.SCIENCEDIRECT.com
In total, 39 patients were included in this study (22 GBM, 6 PCNSL, 11 METS). There was no significant difference in the age distribution among GBM (mean standard deviation; 60.6 14.8 years), PCNSL (65.3 11.1 years), and METS (61.8 12.9 years) (P ¼ 0.5364, KruskaleWallis test). rCBFt and rCBVt of GBM were statistically greater than those of PCNSL (P ¼ 0.0005, 0.0002, respectively) and METS (P ¼ 0.0044, 0.0028, respectively; ManneWhitney U test). rCBFp and rCBVp of GBM were statistically greater than those of METS (P ¼ 0.0002, 0.0017, respectively; ManneWhitney U test). rMTTp of METS was statistically greater than that of GBM and PCNSL (P ¼ 0.0001, P ¼ 0.0007, respectively; ManneWhitney U test) (Figure 1). ROC analysis for differentiating GBM from other tumors showed the best diagnostic performance with rCBVt (area under the curve ¼ 0.8636, 95% confidence interval 0.6699e0.9518). The cut-off value was 3.65 with a sensitivity and specificity of 81.8% and 88.2%, respectively. There were statistical differences in rCBFt, rCBFp, rCBVt, and rCBVp (P < 0.05, c2 test) (Figure 2,
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.291
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
Figure 6. A representative case of glioblastoma. (A) Contrast-enhanced computed tomography showed left frontal enhanced tumor. Cerebral blood flow (CBF) map (B) and cerebral blood volume (CBV) map (C) showed
Table 1). A stepwise regression analysis showed that greater rCBVt and lower rPSt correlated with GBM. Combination of rCBVt and rPSt could differentiate GBM with a sensitivity and specificity of 81.8% and 94.1%, respectively (Figure 3). ROC analysis for differentiating MET from other tumors showed the best diagnostic performance with rMTTp (area under the curve ¼ 0.901, 95% confidence interval 0.7377e0.9671). The cut-off value was 1.46 with a sensitivity and specificity of 90.9% and 89.2%, respectively. There were statistical differences in rCBFp, rCBVp, and rMTTp (P < 0.05, c2 test) (Figure 4, Table 2). A stepwise regression analysis showed that greater rMTTp and lower rCBFp correlated with METS. The combination of rMTTp and rCBFp could differentiate METS with a sensitivity and specificity of 90.9% and 82.1%, respectively (Figure 5). Representative images are shown in Figures 6e8.
DISCUSSION This study investigated PCT parameters of the tumor and peritumoral regions for differentiating GBM, PCNSL, and METS. We conclude that integrated application of several PCT parameters was valuable to differentiate these entities with high sensitivity and specificity. PCT can demonstrate the histologic features of brain tissue. CBV and CBF correlate with microvascular density, and PS corresponds with microvascular cellular proliferation.10 Microvascular
WORLD NEUROSURGERY -: e1-e8, - 2018
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
elevated CBF and CBV at tumor. (D) Mean transit time (MTT) map did not show significant change in MTT at tumor. (E) Permeability surface (PS) map showed elevated PS at tumor.
density and microvascular cellular proliferation are histologic angiogenetic markers, implying that PCT can predict histologic vascularity and angiogenetic activity. The clinical efficacy of PCT in preoperative histologic grading of glioma already has been reported.4,7 Reflecting the neovascularization and increased angiogenetic activity, high-grade glioma demonstrated greater CBV, CBF, and PS compared with low-grade glioma.4,5 PCT is also applicable to differentiate treatment-induced necrosis from recurrent/progressive tumor.8 In our study, we applied PCT parameters for differentiation among GBM, METS, and PCNSL. CBV and CBF of high-grade glioma obtained by PCT are reported to be significantly greater than those of PCNSL.11-13 However, a statistical difference in CBV and CBF between high grade glioma and METS has not been established.12,13 Magnetic resonanceebased perfusion study revealed that CBV of GBM was greatest among GBM, PCNSL, and METS, although it was not significantly different between GBM and METS.14 Bauer et al.15 also reported greater CBV in GBM than in METS but without statistical significance. Furthermore, Sunwoo et al.16 reported that CBF of GBM is significantly greater than that of METS.16 Our results of elevated rCBVt and rCBFt of GBM than those of PCNSL are consistent with previous studies. We also showed that rCBVt and rCBFt of GBM were statistically greater than those of METS, wherein parameters of METS might differ depending on the primary lesion. Tumor vascularization is one of the characteristics of GBM,17 which would result in elevated CBV and CBF in these studies.
www.WORLDNEUROSURGERY.org
e5
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Figure 7. A representative case of primary central nervous system lymphoma. (A) Contrast-enhanced computed tomography showed left parietal enhanced tumor. Cerebral blood flow (CBF) map (B) and cerebral blood volume (CBV) (C) map showed mildly enhanced
PS alone did not show statistical differences for discriminating brain tumors in our study. PS indicates flow of the molecule through capillary membranes in a certain volume of tissue18 and increased permeability due to bloodebrain barrier dysfunction.19 The other previous PCT study reported increased PS values in GBM and METS than in contralateral normal-appearing brain tissue, but no statistical differences were found.20 Another perfusion study also demonstrated increased PS in GBM and PCNSL than that in contralateral normal-appearing brain tissue, but statistical difference was not established.11 Stecco et al.21 reported that postoperative peritumoral PS parameters are informative to detect the infiltration of GBM. In addition, they described that the postoperative peritumoral PS of GBM is greater than the preoperative PS of MET. GBM, PCNSL, and METS alter the bloodebrain barrier with varying mechanisms,22-24 resulting in diffusion of blood or contrast molecules into extravascular space. Consistent with previous studies, our results also suggest increased PS in these malignant brain tumors; however, statistical differences could not be established. Future, larger studies are necessary to confirm the role of PS for differentiation among these tumors. Diagnostic accuracy of a single PCT parameter is limited. Lee et al.25 reported improved accuracy with the combined analysis of CBV and PS to differentiate intracranial masses. PS assumed an
e6
www.SCIENCEDIRECT.com
CBF and CBV within tumor. (D) Mean transit time (MTT) map did not show significant change in MTT at tumor. (E) Permeability surface (PS) map showed elevated PS within tumor.
important role in differentiating GBM, PCNSL, and MES from low-grade tumor and radiation necrosis. CBV was determined as an important factor for differentiation among malignant brain tumors. Our study is congruent with previously published study and stepwise analysis established elevated CBV and lower PS as useful PCT parameters for differentiating GBM from these malignant tumors. Tumoral behavior at the adjacent brain parenchyma is different and peritumoral assessment provided useful information, especially for differentiating METS from other tumors in our study. The authors of a previous PCT study reported elevated CBV at peritumoral region of GBM than those of METS and PCNSL.13 Other magnetic resonanceebased perfusion studies also demonstrated elevated CBV at peritumoral region of GBM than that of METS.15,26 Because GBM is one of the most aggressive brain tumors and notorious for rapid invasion of adjacent brain structure,27 it increases cellularity and promotes neoangiogenesis. Therefore, the elevated peritumoral CBV and CBF reflect the invasive biological characteristics of GBM. PCNSL also invades the normal white matter, and histopathologic study shows diffusely spreading malignant lymphocyte into normal brain parenchyma.28 On the contrary, most METS are well demarcated from the brain parenchyma29 and dominantly cause vasogenic edema of the peritumoral region.15 In summary, GBM and
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.291
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Figure 8. A representative case of brain metastasis. (A) Contrast-enhanced computed tomography showed left parietal enhanced tumor. Cerebral blood flow (CBF) map (B) and Cerebral blood volume (CBV) (C) map showed
PCNSL infiltrate the adjacent brain parenchyma, whereas most METS compress adjacent normal brain tissue and interrupt peritumoral circulatory network. MTT stands for the time interval between arterial inflow and venous outflow of contrast medium.18 In our study, peritumoral MTT of METS was more elevated compared with that of GBM and PCNSL. Interrupted peritumoral circulatory network in METS resulted in prolonged peritumoral MTT compared with that of GBM and PCNSL. Stepwise analysis also showed that prolonged rMTTp and lower rCBFp discerned METS among these malignant tumors. These results are concordant with the biological characteristics of these malignant tumors. Based on this radiologic information, clinicians will be able to predict the histologic diagnosis and select appropriate surgical strategy. Patients with GBM will undergo extensive tumor resection, and a biopsy will be performed in only some of the patients because of the anatomical or functional location. Patients with PCNSL will undergo a biopsy only. Subsequently, they will undergo chemotherapy and/or radiotherapy. Patients with MET will undergo multidisciplinary treatment for the primary
REFERENCES 1. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005;352:987-996.
mildly enhanced CBF and CBV at tumor. (D) Mean transit time (MTT) map did not show significant change in MTT at tumor. (E) Permeability surface (PS) map showed elevated PS at tumor.
tumor and metastatic brain tumor based on the patients’ condition. In addition to the retrospective design, this study has a few limitations. First, we did not have a sufficient number of the cases in the study, particularly PCNSL. Second, this study might have sampling bias because the specimen might not have corresponded to the intended area of PCT map. Larger study and immunehistological diagnosis are necessary to confirm the usefulness of PCT parameters.
CONCLUSIONS PCT parameters provide valuable information for the preoperative diagnosis of GBM, PCNSL, and METS. We noted that comprehensive assessment of PCT parameters increased the diagnostic accuracy of these malignant brain tumors. This study revealed that rCBVt and rPSt best predicted GBM and rCBFp and rMTTp best predicted METS.
2. Gerstner ER, Batchelor TT. Primary central nervous system lymphoma. Arch Neurol. 2010;67:291-297. 3. Soon YY, Tham IWK, Lim KH, Koh WY, Lu JJ. Surgery or radiosurgery plus whole brain radiotherapy versus surgery or radiosurgery alone
WORLD NEUROSURGERY -: e1-e8, - 2018
for brain metastases. Cochrane Database Syst Rev. 2014:CD009454. https://doi.org/10.1002/14651858. CD009454.pub2. 4. Shankar JJS, Woulfe J, Silva VD, Nguyen TB. Evaluation of perfusion CT in grading and
www.WORLDNEUROSURGERY.org
e7
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
prognostication of high-grade gliomas at diagnosis: a pilot study. AJR Am J Roentgenol. 2013;200: W504-W509.
normalized cerebral blood volume in enhancing and perienhancing lesions. AJNR Am J Neuroradiol. 2010;31:1699-1706.
5. Yeung TPC, Bauman G, Yartsev S, Fainardi E, Macdonald D, Lee T-Y. Dynamic perfusion CT in brain tumors. Eur J Radiol. 2015;84:2386-2392.
14. Calli C, Kitis O, Yunten N, Yurtseven T, Islekel S, Akalin T. Perfusion and diffusion MR imaging in enhancing malignant cerebral tumors. Eur J Radiol. 2006;58:394-403.
6. Lui YW, Tang ER, Allmendinger AM, Spektor V. Evaluation of CT perfusion in the setting of cerebral ischemia: patterns and pitfalls. AJNR Am J Neuroradiol. 2010;31:1552-1563. 7. Beppu T, Sasaki M, Kudo K, Kurose A, Takeda M, Kashimura H, et al. Prediction of malignancy grading using computed tomography perfusion imaging in nonenhancing supratentorial gliomas. J Neurooncol. 2011;103:619-627. 8. Jain R, Narang J, Schultz L, Scarpace L, Saksena S, Brown S, et al. Permeability estimates in histopathology-proved treatment-induced necrosis using perfusion CT: can these add to other perfusion parameters in differentiating from recurrent/progressive tumors? Am J Neuroradiol. 2011;32:658-663. 9. Spampinato MV, Schiarelli C, Cianfoni A, Giglio P, Welsh CT, Bisdas S, et al. Correlation between cerebral blood volume measurements by perfusion-weighted magnetic resonance imaging and two-year progression-free survival in gliomas. Neuroradiol J. 2013;26:385-395. 10. Jain RK, di Tomaso E, Duda DG, Loeffler JS, Sorensen AG, Batchelor TT. Angiogenesis in brain tumours. Nat Rev Neurosci. 2007;8:610-622. 11. Schramm P, Xyda A, Klotz E, Tronnier V, Knauth M, Hartmann M. Dynamic CT perfusion imaging of intra-axial brain tumours: differentiation of high-grade gliomas from primary CNS lymphomas. Eur Radiol. 2010;20:2482-2490. 12. Xyda A, Haberland U, Klotz E, Jung K, Bock HC, Schramm R, et al. Diagnostic performance of whole brain volume perfusion CT in intra-axial brain tumors: preoperative classification accuracy and histopathologic correlation. Eur J Radiol. 2012; 81:4105-4111. 13. Ma JH, Kim HS, Rim NJ, Kim SH, Cho KG. Differentiation among glioblastoma multiforme, solitary metastatic tumor, and lymphoma using whole-tumor histogram analysis of the
e8
www.SCIENCEDIRECT.com
15. Bauer AH, Erly W, Moser FG, Maya M, Nael K. Differentiation of solitary brain metastasis from glioblastoma multiforme: a predictive multiparametric approach using combined MR diffusion and perfusion. Neuroradiology. 2015;57: 697-703. 16. Sunwoo L, Yun T-J, You S-H, Yoo RE, Kang KM, Choi SH, et al. Differentiation of glioblastoma from brain metastasis: qualitative and quantitative analysis using arterial spin labeling MR imaging. PLoS One. 2016;11:e0166662. 17. Das S, Marsden PA. Angiogenesis in glioblastoma. N Engl J Med. 2013;369:1561-1563. 18. Cuenod CA, Balvay D. Perfusion and vascular permeability: basic concepts and measurement in DCE-CT and DCE-MRI. Diagn Interv Imaging. 2013; 94:1187-1204. 19. Avsenik J, Bisdas S, Popovic KS. Blood-brain barrier permeability imaging using perfusion computed tomography. Radiol Oncol. 2015;49: 107-114. 20. Fainardi E, Di Biase F, Borrelli M, Saletti A, Cavallo M, Sarubbo S, et al. Potential role of CT perfusion parameters in the identification of solitary intra-axial brain tumor grading. Acta Neurochir Suppl. 2010;106:283-287.
barrier by invading glioma cells. Nat Commun. 2014;5:4196. 24. Molnár PP, O’Neill BP, Scheithauer BW, Groothuis DR. The blood-brain barrier in primary CNS lymphomas: ultrastructural evidence of endothelial cell death. Neuro Oncol. 1999;1:89-100. 25. Lee Y-J, Ahn K-J, Kim B-S, Yoo W-J. Role of perfusion CT in differentiating between various cerebral masses using normalized permeability surface area product and cerebral blood volume. Clin Imaging. 2012;36:680-687. 26. Lehmann P, Saliou G, de Marco G, Monet P, Souraya SE, Bruniau A, et al. Cerebral peritumoral oedema study: does a single dynamic MR sequence assessing perfusion and permeability can help to differentiate glioblastoma from metastasis? Eur J Radiol. 2012;81:522-527. 27. Burger PC, Heinz ER, Shibata T, Kleihues P. Topographic anatomy and CT correlations in the untreated glioblastoma multiforme. J Neurosurg. 1988;68:698-704. 28. Aho R, Ekfors T, Haltia M, Kalimo H. Pathogenesis of primary central nervous system lymphoma: invasion of malignant lymphoid cells into and within the brain parenchyme. Acta Neuropathol. 1993;86:71-76. 29. Berghoff AS, Rajky O, Winkler F, Bartsch R, Furtner J, Hainfellner JA, et al. Invasion patterns in brain metastases of solid cancers. Neuro Oncol. 2013;15:1664-1672.
21. Stecco A, Fabbiano F, Amatuzzo P, Cignini E, Brambilla M, Masini L, et al. Perfusion computed tomography assessments of peri-enhancing brain tissue in high-grade gliomas. Radiol Med. 2013;118: 431-443.
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
22. Wilhelm I, Molnár J, Fazakas C, Haskó J, Krizbai IA. Role of the blood-brain barrier in the formation of brain metastases. Int J Mol Sci. 2013; 14:1383-1411.
Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.07.291
Received 2 March 2018; accepted 31 July 2018
Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com
23. Watkins S, Robel S, Kimbrough IF, Robert SM, Ellis-Davies G, Sontheimer H. Disruption of astrocyte-vascular coupling and the blood-brain
1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.291
Perfusion Computed Tomography Parameters Are Useful for Differentiating Glioblastoma, Lymphoma, and Metastasis Shumpei Onishi1,5, Yoshinori Kajiwara6, Takeshi Takayasu1, Manish Kolakshyapati1, Minoru Ishifuro3, Vishwa Jeet Amatya2, Yukio Takeshima2, Kazuhiko Sugiyama4, Kaoru Kurisu1, Fumiyuki Yamasaki1
-
BACKGROUND: Perfusion computed tomography (PCT) reflects blood flow and capillary condition, which is valuable in assessing brain tumors. We evaluated PCT parameters at the tumor (t) and peritumoral (p) region to differentiate malignant brain tumors.
The combination of rCBVt and rPSt could differentiate GBM from other tumors with sensitivity and specificity of 81.8% and 94.1%. The combination of rCBFp and rMTTp could differentiate METS from other tumors with sensitivity and specificity of 90.9% and 82.1%.
METHODS: We performed PCT in 39 patients with supratentorial malignant brain tumors (22 glioblastomas, 6 lymphomas, 11 metastases). Regions of interests were placed manually at tumor, peritumoral region, and contralateral normal-appearing white matter. Cerebral blood volume (CBV), cerebral blood flow (CBF), mean transit time (MTT), and permeability surface (PS) were measured. These parameters were divided by those of contralateral normal-appearing white matter to normalize at tumor (relative [r]CBVt, rCBFt, rMTTt, rPSt) and peritumoral regions (rCBVp, rCBFp, rMTTp, rPSp). The parameters were evaluated with ManneWhitney U test and receiver operating characteristics analyses. Stepwise analyses also were performed to select useful PCT parameters for differentiating these tumors.
-
-
RESULTS: The rCBFt and rCBVt of glioblastoma (GBM) were greater than those of primary central nervous system lymphoma (PCNSL) (P [ 0.0005, 0.0002) and brain metastasis (METS) (P [ 0.0044, 0.0028). The rMTTp of METS was greater than that of GBM and PCNSL (P [ 0.0001, 0.0007).
-
Key words Brain metastasis - Glioblastoma - Llymphoma - Perfusion CT - Primary central nervous system -
Abbreviations and Acronyms CBF: Cerebral blood flow CBV: Cerebral blood volume CT: Computed tomography GBM: Glioblastoma METS: Brain metastasis MTT: Mean transit time p: Peritumoral region PCNSL: Primary central nervous system lymphoma PCT: Perfusion computed tomography PS: Permeability surface
WORLD NEUROSURGERY -: e1-e8, - 2018
CONCLUSIONS: Our study introduces and supports the usefulness of PCT parameters for differentiation among GBM, PCNSL, and METS. rCBVt and rPSt may be the best predictors of GBM. rCBFp and rMTTp may be the best predictors of METS.
INTRODUCTION
P
reoperative differentiation among glioblastoma (GBM), primary central nervous system lymphoma (PCNSL), and brain metastasis (METS) can be challenging. These tumors often present as contrast-enhancing on conventional computed tomography (CT), which is a sometimes-indiscernible radiographic characteristic. Preoperative differentiation among these tumors is of high clinical relevance, as the treatment of these tumors is entirely different. GBM requires maximum resection followed by intensive chemo- and radiotherapy,1 PCNSL only requires biopsy followed by chemotherapy and sometimes with radiotherapy,2 and METS requires stereotactic irradiation or systemic treatment.3
r: Relative ROC: Receiver operating characteristic t: Tumor From the Departments of 1Neurosurgery and 2Pathology, Graduate School of Biomedical and Health Sciences, Hiroshima University, Hiroshima; Departments of 3Diagnostic Imaging and 4 Clinical Oncology and Neuro-oncology Program, Hiroshima University Hospital, Hiroshima; 5 Department of Neurosurgery, Higashihiroshima Medical Center, Hiroshima; and 6 Department of Neurosurgery, Matsuyama Red Cross Hospital, Ehime, Japan To whom correspondence should be addressed: Fumiyuki Yamasaki, M.D., Ph.D. [E-mail: [email protected]] Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.07.291 Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
www.WORLDNEUROSURGERY.org
e1
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Figure 1. Box plots demonstrate perfusion computed tomography parameters of glioblastoma (GBM), primary central nervous system lymphoma (PCNSL), and brain metastasis (METS) at tumor and peritumoral regions. rCBFt, relative cerebral blood volume, tumor; rCBVt, relative cerebral blood volume, tumor; rMTTt, relative mean transit time,
tumor; rPSt, relative permeability surface, tumor; rCBFp, relative cerebral blood volume, peritumoral region; rCBVp, relative cerebral blood volume, peritumoral region; rMTTp, relative mean transit time, peritumoral region; rPSp, relative permeability surface, peritumoral region.
Perfusion imaging by CT and magnetic resonance imaging has been used to visualize the ischemic changes in brain tissue and the hemodynamics and vascularity of brain tumors.4-6 The usefulness of perfusion computed tomography (PCT) has been reported in grading gliomas,4,7 differentiating tumor progression from treatment-induced effects,8 and predicting prognosis after treatment.9 However, the integrated interpretation of PCT parameters for differentiating malignant tumors is unclear. We evaluated PCT parameters at tumor and peritumoral regions to differentiate GBM, METS, and PCNSL.
MATERIALS AND METHODS Patients This retrospective study was approved by the institutional review board of Hiroshima University Hospital. We included 39 patients with histologically proven supratentorial malignant brain tumors treated at Hiroshima University Hospital from August 2004 to August 2009. The patients included 19 men and 20 women, ranging between 25 and 83 years (mean age 60.6 years). For all 39 patients, the histologic diagnosis was confirmed by a neuropathologist. This study group consisted of 22 patients (11 men and 11 women) with GBM aged 25e83 years (mean age 60.6 years), 6 patients (4 men and 2 women) with PCNSL aged 43e76 years (mean age 65.3 years), and 11 patients (4 men and 7 women) with METS aged 34e78 years (mean age 61.8 years). All patients underwent PCT before any neurosurgical treatment.
e2
www.SCIENCEDIRECT.com
Figure 2. Receiver operating characteristic analysis for differentiating glioblastoma from other tumors. The analysis showed performances with relative cerebral blood volume, tumor (rCBFt; area under the curve [AUC] ¼ 0.8476), relative cerebral blood volume, peritumoral region (rCBFp; AUC ¼ 0.7955), relative cerebral blood volume, tumor (rCBVt; AUC ¼ 0.8636), relative cerebral blood volume, peritumoral region (rCBVp; AUC ¼ 0.7513), relative mean transit time, tumor (rMTTt; AUC ¼ 0.6511), relative mean transit time, peritumoral region (rMTTp; AUC ¼ 0.7299), relative permeability surface, tumor (rPSt; AUC ¼ 0. 6257), relative permeability surface, peritumoral region (rPSp; AUC ¼ 0.488).
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.291
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Table 1. Comparison of PCT Parameters to Differentiate GBM From Other Tumors Parameters
AUC
95% CI
Cut-Off Value
Sensitivity
Specificity
P Value
rCBFt
0.8476
(0.6645e0.9398)
2.31
0.9545
0.6471
0.0163*
rCBFp
0.7955
(0.6645e0.9398)
0.59
0.8182
0.8235
0.0316*
rCBVt
0.8636
(0.6699e0.9518)
3.65
0.8182
0.8824
0.0067*
rCBVp
0.7513
(0.6699e0.9518)
0.7
0.7727
0.7647
0.0352*
rMTTt
0.6511
(0.4643e0.8007)
0.9
0.5
0.8235
0.1692
rMTTp
0.7299
(0.4634e0.8007)
1.44
0.8636
0.5882
0.109
rPSt
0.6257
(0.4359e0.7833)
11.75
0.5909
0.7059
0.6665
rPSp
0.488
(0.4359e0.7833)
1.76
0.3182
0.8235
0.5363
PCT, perfusion computed tomography; GBM, glioblastoma; AUC, area under the curve; 95% CI, 95% confidence interval; rCBFt, relative cerebral blood volume, tumor; rCBFp, relative cerebral blood volume, peritumoral region; rCBVt, relative cerebral blood volume, tumor; rCBVp, relative cerebral blood volume, peritumoral region; rMTTt, relative mean transit time, tumor; rMTTp, relative mean transit time, peritumoral region; rPSt, relative permeability surface, tumor; rPSp, relative permeability surface, peritumoral region. *Indicates statistical significance.
Image Acquisition Perfusion studies were performed with a 16-slice helical CT scanner (LightSpeed Ultra; General Electric Medical Systems, Chicago, Illinois, USA). A noncontrast CT scan was obtained to localize the region of interest before we obtained a perfusion scan. For the perfusion studies, 40 mL of nonionic contrast medium (350 mg/mL) was injected at a rate of 5 mL/s through a 20-gauge intravenous line. At 5 seconds into the injection, a helical scan was initiated with the following technique: 80 kVp, 80 mA, and 1 second per rotation. After the initial 5-second cine scan, 8 more
Figure 3. Receiver operating characteristic analysis for differentiating glioblastoma (GBM) from other tumors with relative cerebral blood volume, tumor and relative permeability surface, tumor. These parameters could differentiate GBM with sensitivity and specificity of 81.8% and 94.1%.
WORLD NEUROSURGERY -: e1-e8, - 2018
axial images were acquired, 1 image every 2 seconds for an additional 52 seconds, thereby taking a total acquisition time of 57 seconds. Perfusion maps of cerebral blood volume (CBV), cerebral blood flow (CBF), mean transit time (MTT), and permeability surface (PS) were generated with an Advantage Windows Workstation (Ver. 4.2; GE Medical Systems) using CT Perfusion 3
Figure 4. Receiver operating characteristic analysis for differentiating brain metastasis from other tumors. The analysis showed the best diagnostic performance with relative cerebral blood volume, tumor (rCBFt; area under the curve [AUC] ¼ 0.7159), relative cerebral blood volume, peritumoral region (rCBFp; AUC ¼ 0.8571), relative cerebral blood volume, tumor (rCBVt; AUC ¼ 0.7338), relative cerebral blood volume, peritumoral region (rCBVp; AUC ¼ 0.789), relative mean transit time, tumor (rMTTt; AUC ¼ 0.6623), relative mean transit time, peritumoral region (rMTTp; AUC ¼ 0.901), relative permeability surface, tumor (rPSt; AUC ¼ 0.6688), relative permeability surface, peritumoral region (rPSp; AUC ¼ 0.5422).
www.WORLDNEUROSURGERY.org
e3
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Table 2. Comparison of PCT Parameters to Differentiate MET From Other Tumors Parameters rCBFt
AUC
95% CI
Cutoff Value
Sensitivity
Specificity
P Value
0.7159
(0.4912e0.8680)
3.1
0.8182
0.6429
0.1796
rCBFp
0.8571
(0.6974e0.9398)
0.58
1
0.75
0.0083*
rCBVt
0.7338
(0.4962e0.8852)
2.83
0.8182
0.6786
0.1405
rCBVp
0.789
(0.6061e0.9008)
0.69
0.8182
0.6786
0.0177*
rMTTt
0.6623
(0.4548e0.8218)
1.63
0.3636
92.86
0.1054
rMTTp
0.901
(0.7377e0.9671)
1.46
0.9091
0.8929
0.0241*
rPSt
0.6688
(0.4617e0.8263)
13.03
0.8182
0.5714
0.3415
rPSp
0.5422
(0.3521e0.7208)
1.66
1
0.3571
0.2528
PCT, perfusion computed tomography; MET, brain metastasis; AUC, area under the curve; 95% CI, 95% confidence interval; rCBFt, relative cerebral blood volume, tumor; rCBFp, relative cerebral blood volume, peritumoral region; rCBVt, relative cerebral blood volume, tumor; rCBVp, relative cerebral blood volume, peritumoral region; rMTTt, relative mean transit time, tumor; rMTTp, relative mean transit time, peritumoral region; rPSt, relative permeability surface, tumor; rPSp, relative permeability surface, peritumoral region. *Indicates statistical significance.
software (GE Medical System) in all patients. Regions of interest were placed manually at the enhanced tumor, peritumoral region, and contralateral normal-appearing white matter confirmed on contrast-enhanced and noncontrast CT images. Peritumoral region were defined as low density around enhanced tumor on CT image. The PCT parameters were divided by those of contralateral normal-appearing white matter to normalize the parameters at tumor (rCBVt, rCBFt, rMTTt, rPSt) and peritumoral region (rCBVp, rCBFp, rMTTp, rPSp), where r indicates relative; t indicates tumor, and p indicates peritumoral region.
Histologic Study Tumor specimens after surgical resection or biopsy were fixed in 10% phosphate-buffered formalin and embedded in paraffin blocks. Representative slides were then stained with hematoxylineeosin reagent for standard histologic diagnosis. Statistical Analysis Statistical analyses were performed with GraphPad Prism, version7.00 for Mac (GraphPad Software, San Diego California, USA) and JMP Pro, version 13.0 (SAS institute, Cary, North Carolina, USA). GBM, METS, and PCNSL were compared at tumor (t) and peritumoral (p) region respectively by using normalized PCT parameters (rCBF, rCBV, rMTT, rPS). These parameters were evaluated with the ManneWhitney U test and receiver operating characteristics (ROC) analyses; statistical significance was assigned when P < 0.05. Stepwise analyses were also performed to select useful PCT parameters for differentiating these tumors. RESULTS
Figure 5. Receiver operating characteristic analysis for differentiating brain metastasis from other tumors with relative cerebral blood volume, peritumoral region and relative mean transit time, peritumoral region. These parameters could differentiate glioblastoma with sensitivity and specificity of 90.9% and 82.1%, respectively.
e4
www.SCIENCEDIRECT.com
In total, 39 patients were included in this study (22 GBM, 6 PCNSL, 11 METS). There was no significant difference in the age distribution among GBM (mean standard deviation; 60.6 14.8 years), PCNSL (65.3 11.1 years), and METS (61.8 12.9 years) (P ¼ 0.5364, KruskaleWallis test). rCBFt and rCBVt of GBM were statistically greater than those of PCNSL (P ¼ 0.0005, 0.0002, respectively) and METS (P ¼ 0.0044, 0.0028, respectively; ManneWhitney U test). rCBFp and rCBVp of GBM were statistically greater than those of METS (P ¼ 0.0002, 0.0017, respectively; ManneWhitney U test). rMTTp of METS was statistically greater than that of GBM and PCNSL (P ¼ 0.0001, P ¼ 0.0007, respectively; ManneWhitney U test) (Figure 1). ROC analysis for differentiating GBM from other tumors showed the best diagnostic performance with rCBVt (area under the curve ¼ 0.8636, 95% confidence interval 0.6699e0.9518). The cut-off value was 3.65 with a sensitivity and specificity of 81.8% and 88.2%, respectively. There were statistical differences in rCBFt, rCBFp, rCBVt, and rCBVp (P < 0.05, c2 test) (Figure 2,
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.291
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
Figure 6. A representative case of glioblastoma. (A) Contrast-enhanced computed tomography showed left frontal enhanced tumor. Cerebral blood flow (CBF) map (B) and cerebral blood volume (CBV) map (C) showed
Table 1). A stepwise regression analysis showed that greater rCBVt and lower rPSt correlated with GBM. Combination of rCBVt and rPSt could differentiate GBM with a sensitivity and specificity of 81.8% and 94.1%, respectively (Figure 3). ROC analysis for differentiating MET from other tumors showed the best diagnostic performance with rMTTp (area under the curve ¼ 0.901, 95% confidence interval 0.7377e0.9671). The cut-off value was 1.46 with a sensitivity and specificity of 90.9% and 89.2%, respectively. There were statistical differences in rCBFp, rCBVp, and rMTTp (P < 0.05, c2 test) (Figure 4, Table 2). A stepwise regression analysis showed that greater rMTTp and lower rCBFp correlated with METS. The combination of rMTTp and rCBFp could differentiate METS with a sensitivity and specificity of 90.9% and 82.1%, respectively (Figure 5). Representative images are shown in Figures 6e8.
DISCUSSION This study investigated PCT parameters of the tumor and peritumoral regions for differentiating GBM, PCNSL, and METS. We conclude that integrated application of several PCT parameters was valuable to differentiate these entities with high sensitivity and specificity. PCT can demonstrate the histologic features of brain tissue. CBV and CBF correlate with microvascular density, and PS corresponds with microvascular cellular proliferation.10 Microvascular
WORLD NEUROSURGERY -: e1-e8, - 2018
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
elevated CBF and CBV at tumor. (D) Mean transit time (MTT) map did not show significant change in MTT at tumor. (E) Permeability surface (PS) map showed elevated PS at tumor.
density and microvascular cellular proliferation are histologic angiogenetic markers, implying that PCT can predict histologic vascularity and angiogenetic activity. The clinical efficacy of PCT in preoperative histologic grading of glioma already has been reported.4,7 Reflecting the neovascularization and increased angiogenetic activity, high-grade glioma demonstrated greater CBV, CBF, and PS compared with low-grade glioma.4,5 PCT is also applicable to differentiate treatment-induced necrosis from recurrent/progressive tumor.8 In our study, we applied PCT parameters for differentiation among GBM, METS, and PCNSL. CBV and CBF of high-grade glioma obtained by PCT are reported to be significantly greater than those of PCNSL.11-13 However, a statistical difference in CBV and CBF between high grade glioma and METS has not been established.12,13 Magnetic resonanceebased perfusion study revealed that CBV of GBM was greatest among GBM, PCNSL, and METS, although it was not significantly different between GBM and METS.14 Bauer et al.15 also reported greater CBV in GBM than in METS but without statistical significance. Furthermore, Sunwoo et al.16 reported that CBF of GBM is significantly greater than that of METS.16 Our results of elevated rCBVt and rCBFt of GBM than those of PCNSL are consistent with previous studies. We also showed that rCBVt and rCBFt of GBM were statistically greater than those of METS, wherein parameters of METS might differ depending on the primary lesion. Tumor vascularization is one of the characteristics of GBM,17 which would result in elevated CBV and CBF in these studies.
www.WORLDNEUROSURGERY.org
e5
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Figure 7. A representative case of primary central nervous system lymphoma. (A) Contrast-enhanced computed tomography showed left parietal enhanced tumor. Cerebral blood flow (CBF) map (B) and cerebral blood volume (CBV) (C) map showed mildly enhanced
PS alone did not show statistical differences for discriminating brain tumors in our study. PS indicates flow of the molecule through capillary membranes in a certain volume of tissue18 and increased permeability due to bloodebrain barrier dysfunction.19 The other previous PCT study reported increased PS values in GBM and METS than in contralateral normal-appearing brain tissue, but no statistical differences were found.20 Another perfusion study also demonstrated increased PS in GBM and PCNSL than that in contralateral normal-appearing brain tissue, but statistical difference was not established.11 Stecco et al.21 reported that postoperative peritumoral PS parameters are informative to detect the infiltration of GBM. In addition, they described that the postoperative peritumoral PS of GBM is greater than the preoperative PS of MET. GBM, PCNSL, and METS alter the bloodebrain barrier with varying mechanisms,22-24 resulting in diffusion of blood or contrast molecules into extravascular space. Consistent with previous studies, our results also suggest increased PS in these malignant brain tumors; however, statistical differences could not be established. Future, larger studies are necessary to confirm the role of PS for differentiation among these tumors. Diagnostic accuracy of a single PCT parameter is limited. Lee et al.25 reported improved accuracy with the combined analysis of CBV and PS to differentiate intracranial masses. PS assumed an
e6
www.SCIENCEDIRECT.com
CBF and CBV within tumor. (D) Mean transit time (MTT) map did not show significant change in MTT at tumor. (E) Permeability surface (PS) map showed elevated PS within tumor.
important role in differentiating GBM, PCNSL, and MES from low-grade tumor and radiation necrosis. CBV was determined as an important factor for differentiation among malignant brain tumors. Our study is congruent with previously published study and stepwise analysis established elevated CBV and lower PS as useful PCT parameters for differentiating GBM from these malignant tumors. Tumoral behavior at the adjacent brain parenchyma is different and peritumoral assessment provided useful information, especially for differentiating METS from other tumors in our study. The authors of a previous PCT study reported elevated CBV at peritumoral region of GBM than those of METS and PCNSL.13 Other magnetic resonanceebased perfusion studies also demonstrated elevated CBV at peritumoral region of GBM than that of METS.15,26 Because GBM is one of the most aggressive brain tumors and notorious for rapid invasion of adjacent brain structure,27 it increases cellularity and promotes neoangiogenesis. Therefore, the elevated peritumoral CBV and CBF reflect the invasive biological characteristics of GBM. PCNSL also invades the normal white matter, and histopathologic study shows diffusely spreading malignant lymphocyte into normal brain parenchyma.28 On the contrary, most METS are well demarcated from the brain parenchyma29 and dominantly cause vasogenic edema of the peritumoral region.15 In summary, GBM and
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.291
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
Figure 8. A representative case of brain metastasis. (A) Contrast-enhanced computed tomography showed left parietal enhanced tumor. Cerebral blood flow (CBF) map (B) and Cerebral blood volume (CBV) (C) map showed
PCNSL infiltrate the adjacent brain parenchyma, whereas most METS compress adjacent normal brain tissue and interrupt peritumoral circulatory network. MTT stands for the time interval between arterial inflow and venous outflow of contrast medium.18 In our study, peritumoral MTT of METS was more elevated compared with that of GBM and PCNSL. Interrupted peritumoral circulatory network in METS resulted in prolonged peritumoral MTT compared with that of GBM and PCNSL. Stepwise analysis also showed that prolonged rMTTp and lower rCBFp discerned METS among these malignant tumors. These results are concordant with the biological characteristics of these malignant tumors. Based on this radiologic information, clinicians will be able to predict the histologic diagnosis and select appropriate surgical strategy. Patients with GBM will undergo extensive tumor resection, and a biopsy will be performed in only some of the patients because of the anatomical or functional location. Patients with PCNSL will undergo a biopsy only. Subsequently, they will undergo chemotherapy and/or radiotherapy. Patients with MET will undergo multidisciplinary treatment for the primary
REFERENCES 1. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005;352:987-996.
mildly enhanced CBF and CBV at tumor. (D) Mean transit time (MTT) map did not show significant change in MTT at tumor. (E) Permeability surface (PS) map showed elevated PS at tumor.
tumor and metastatic brain tumor based on the patients’ condition. In addition to the retrospective design, this study has a few limitations. First, we did not have a sufficient number of the cases in the study, particularly PCNSL. Second, this study might have sampling bias because the specimen might not have corresponded to the intended area of PCT map. Larger study and immunehistological diagnosis are necessary to confirm the usefulness of PCT parameters.
CONCLUSIONS PCT parameters provide valuable information for the preoperative diagnosis of GBM, PCNSL, and METS. We noted that comprehensive assessment of PCT parameters increased the diagnostic accuracy of these malignant brain tumors. This study revealed that rCBVt and rPSt best predicted GBM and rCBFp and rMTTp best predicted METS.
2. Gerstner ER, Batchelor TT. Primary central nervous system lymphoma. Arch Neurol. 2010;67:291-297. 3. Soon YY, Tham IWK, Lim KH, Koh WY, Lu JJ. Surgery or radiosurgery plus whole brain radiotherapy versus surgery or radiosurgery alone
WORLD NEUROSURGERY -: e1-e8, - 2018
for brain metastases. Cochrane Database Syst Rev. 2014:CD009454. https://doi.org/10.1002/14651858. CD009454.pub2. 4. Shankar JJS, Woulfe J, Silva VD, Nguyen TB. Evaluation of perfusion CT in grading and
www.WORLDNEUROSURGERY.org
e7
ORIGINAL ARTICLE SHUMPEI ONISHI ET AL.
PCT PARAMETERS TO DIFFERENTIATE GBM, PCNSL, METS
prognostication of high-grade gliomas at diagnosis: a pilot study. AJR Am J Roentgenol. 2013;200: W504-W509.
normalized cerebral blood volume in enhancing and perienhancing lesions. AJNR Am J Neuroradiol. 2010;31:1699-1706.
5. Yeung TPC, Bauman G, Yartsev S, Fainardi E, Macdonald D, Lee T-Y. Dynamic perfusion CT in brain tumors. Eur J Radiol. 2015;84:2386-2392.
14. Calli C, Kitis O, Yunten N, Yurtseven T, Islekel S, Akalin T. Perfusion and diffusion MR imaging in enhancing malignant cerebral tumors. Eur J Radiol. 2006;58:394-403.
6. Lui YW, Tang ER, Allmendinger AM, Spektor V. Evaluation of CT perfusion in the setting of cerebral ischemia: patterns and pitfalls. AJNR Am J Neuroradiol. 2010;31:1552-1563. 7. Beppu T, Sasaki M, Kudo K, Kurose A, Takeda M, Kashimura H, et al. Prediction of malignancy grading using computed tomography perfusion imaging in nonenhancing supratentorial gliomas. J Neurooncol. 2011;103:619-627. 8. Jain R, Narang J, Schultz L, Scarpace L, Saksena S, Brown S, et al. Permeability estimates in histopathology-proved treatment-induced necrosis using perfusion CT: can these add to other perfusion parameters in differentiating from recurrent/progressive tumors? Am J Neuroradiol. 2011;32:658-663. 9. Spampinato MV, Schiarelli C, Cianfoni A, Giglio P, Welsh CT, Bisdas S, et al. Correlation between cerebral blood volume measurements by perfusion-weighted magnetic resonance imaging and two-year progression-free survival in gliomas. Neuroradiol J. 2013;26:385-395. 10. Jain RK, di Tomaso E, Duda DG, Loeffler JS, Sorensen AG, Batchelor TT. Angiogenesis in brain tumours. Nat Rev Neurosci. 2007;8:610-622. 11. Schramm P, Xyda A, Klotz E, Tronnier V, Knauth M, Hartmann M. Dynamic CT perfusion imaging of intra-axial brain tumours: differentiation of high-grade gliomas from primary CNS lymphomas. Eur Radiol. 2010;20:2482-2490. 12. Xyda A, Haberland U, Klotz E, Jung K, Bock HC, Schramm R, et al. Diagnostic performance of whole brain volume perfusion CT in intra-axial brain tumors: preoperative classification accuracy and histopathologic correlation. Eur J Radiol. 2012; 81:4105-4111. 13. Ma JH, Kim HS, Rim NJ, Kim SH, Cho KG. Differentiation among glioblastoma multiforme, solitary metastatic tumor, and lymphoma using whole-tumor histogram analysis of the
e8
www.SCIENCEDIRECT.com
15. Bauer AH, Erly W, Moser FG, Maya M, Nael K. Differentiation of solitary brain metastasis from glioblastoma multiforme: a predictive multiparametric approach using combined MR diffusion and perfusion. Neuroradiology. 2015;57: 697-703. 16. Sunwoo L, Yun T-J, You S-H, Yoo RE, Kang KM, Choi SH, et al. Differentiation of glioblastoma from brain metastasis: qualitative and quantitative analysis using arterial spin labeling MR imaging. PLoS One. 2016;11:e0166662. 17. Das S, Marsden PA. Angiogenesis in glioblastoma. N Engl J Med. 2013;369:1561-1563. 18. Cuenod CA, Balvay D. Perfusion and vascular permeability: basic concepts and measurement in DCE-CT and DCE-MRI. Diagn Interv Imaging. 2013; 94:1187-1204. 19. Avsenik J, Bisdas S, Popovic KS. Blood-brain barrier permeability imaging using perfusion computed tomography. Radiol Oncol. 2015;49: 107-114. 20. Fainardi E, Di Biase F, Borrelli M, Saletti A, Cavallo M, Sarubbo S, et al. Potential role of CT perfusion parameters in the identification of solitary intra-axial brain tumor grading. Acta Neurochir Suppl. 2010;106:283-287.
barrier by invading glioma cells. Nat Commun. 2014;5:4196. 24. Molnár PP, O’Neill BP, Scheithauer BW, Groothuis DR. The blood-brain barrier in primary CNS lymphomas: ultrastructural evidence of endothelial cell death. Neuro Oncol. 1999;1:89-100. 25. Lee Y-J, Ahn K-J, Kim B-S, Yoo W-J. Role of perfusion CT in differentiating between various cerebral masses using normalized permeability surface area product and cerebral blood volume. Clin Imaging. 2012;36:680-687. 26. Lehmann P, Saliou G, de Marco G, Monet P, Souraya SE, Bruniau A, et al. Cerebral peritumoral oedema study: does a single dynamic MR sequence assessing perfusion and permeability can help to differentiate glioblastoma from metastasis? Eur J Radiol. 2012;81:522-527. 27. Burger PC, Heinz ER, Shibata T, Kleihues P. Topographic anatomy and CT correlations in the untreated glioblastoma multiforme. J Neurosurg. 1988;68:698-704. 28. Aho R, Ekfors T, Haltia M, Kalimo H. Pathogenesis of primary central nervous system lymphoma: invasion of malignant lymphoid cells into and within the brain parenchyme. Acta Neuropathol. 1993;86:71-76. 29. Berghoff AS, Rajky O, Winkler F, Bartsch R, Furtner J, Hainfellner JA, et al. Invasion patterns in brain metastases of solid cancers. Neuro Oncol. 2013;15:1664-1672.
21. Stecco A, Fabbiano F, Amatuzzo P, Cignini E, Brambilla M, Masini L, et al. Perfusion computed tomography assessments of peri-enhancing brain tissue in high-grade gliomas. Radiol Med. 2013;118: 431-443.
Conflict of interest statement: The authors declare that the article content was composed in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
22. Wilhelm I, Molnár J, Fazakas C, Haskó J, Krizbai IA. Role of the blood-brain barrier in the formation of brain metastases. Int J Mol Sci. 2013; 14:1383-1411.
Citation: World Neurosurg. (2018). https://doi.org/10.1016/j.wneu.2018.07.291
Received 2 March 2018; accepted 31 July 2018
Journal homepage: www.WORLDNEUROSURGERY.org Available online: www.sciencedirect.com
23. Watkins S, Robel S, Kimbrough IF, Robert SM, Ellis-Davies G, Sontheimer H. Disruption of astrocyte-vascular coupling and the blood-brain
1878-8750/$ - see front matter ª 2018 Elsevier Inc. All rights reserved.
WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2018.07.291