Comparison of diffusion tensor imaging and 11C-methionine positron emission tomography for reliable prediction of tumor cell density in gliomas

Full access

OBJECTIVE

Diffusion MRI is attracting increasing interest for tissue characterization of gliomas, especially after the introduction of antiangiogenic therapy to treat malignant gliomas. The goal of the current study is to elucidate the actual magnitude of the correlation between diffusion MRI and cell density within the tissue. The obtained results were further extended and compared with metabolic imaging with 11C-methionine (MET) PET.

METHODS

Ninety-eight tissue samples from 37 patients were stereotactically obtained via an intraoperative neuronavigation system. Diffusion tensor imaging (DTI) and MET PET were performed as routine presurgical imaging studies for these patients. DTI was converted into fractional anisotropy (FA) and apparent diffusion coefficient (ADC) maps, and MET PET images were registered to Gd-administered T1-weighted images that were used for navigation. Metrics of FA, ADC, and tumor-to-normal tissue ratio of MET PET along with relative values of FA (rFA) and ADC (rADC) compared with normal-appearing white matter were correlated with cell density of the stereotactically obtained tissues.

RESULTS

rADC was significantly lower in lesions obtained from Gd-enhancing lesions than from nonenhancing lesions. Although rADC showed a moderate but statistically significant negative correlation with cell density (p = 0.010), MET PET showed a superb positive correlation with cell density (p < 0.0001). On the other hand, rFA showed little correlation with cell density.

CONCLUSIONS

The presented data validated the use of rADC for estimating the treatment response of gliomas but also caution against overestimating its limited accuracy compared with MET PET.

ABBREVIATIONSADC = apparent diffusion coefficient; DTI = diffusion tensor imaging; FA = fractional anisotropy; MET = 11C-methionine; rADC = relative value of ADC; rFA = relative value of FA; T/Nr = tumor-to-normal tissue ratio; VOI = voxel of interest.

OBJECTIVE

Diffusion MRI is attracting increasing interest for tissue characterization of gliomas, especially after the introduction of antiangiogenic therapy to treat malignant gliomas. The goal of the current study is to elucidate the actual magnitude of the correlation between diffusion MRI and cell density within the tissue. The obtained results were further extended and compared with metabolic imaging with 11C-methionine (MET) PET.

METHODS

Ninety-eight tissue samples from 37 patients were stereotactically obtained via an intraoperative neuronavigation system. Diffusion tensor imaging (DTI) and MET PET were performed as routine presurgical imaging studies for these patients. DTI was converted into fractional anisotropy (FA) and apparent diffusion coefficient (ADC) maps, and MET PET images were registered to Gd-administered T1-weighted images that were used for navigation. Metrics of FA, ADC, and tumor-to-normal tissue ratio of MET PET along with relative values of FA (rFA) and ADC (rADC) compared with normal-appearing white matter were correlated with cell density of the stereotactically obtained tissues.

RESULTS

rADC was significantly lower in lesions obtained from Gd-enhancing lesions than from nonenhancing lesions. Although rADC showed a moderate but statistically significant negative correlation with cell density (p = 0.010), MET PET showed a superb positive correlation with cell density (p < 0.0001). On the other hand, rFA showed little correlation with cell density.

CONCLUSIONS

The presented data validated the use of rADC for estimating the treatment response of gliomas but also caution against overestimating its limited accuracy compared with MET PET.

Malignant gliomas are primary brain tumors associated with a seriously short life expectancy for patients.21 They consist of several different pathological diseases, such as glioblastoma, anaplastic astrocytoma, and oligodendroglioma. When these lesions are imaged using MRI, they usually present as gadolinium-enhanced lesions on T1-weighted images surrounded by hyperintense lesions on T2-weighted images that are thought to represent tumor-infiltrative or vasogenic brain edema around the tumor core.12–14,18 Detailed and precise radiological determination of the extent of tumor cell aggregation or infiltration within the brain tissue is not only a crucial step for treatment planning but also for accurate evaluation of therapeutic response of the disease. Introduction of antiangiogenic therapy to treat malignant gliomas has made the latter objective more challenging than ever, as those agents strongly modify key radiological features of the tumor, including the magnitude of contrast enhancement of the tumor and the extent of tumor-associated brain edema.10

Diffusion MR imaging techniques combined with mathematical and statistical analyses were proposed to solve this challenge with the hope that the reconstructed images could be used to visualize the treatment response in situ and reveal the regions that are unresponsive to treatment to minimize delays in intervention.5,8,9 The theoretical background supporting this technique is founded on the hypothesis that cell density within a tissue and the apparent diffusion coefficient (ADC) are negatively correlated. Higher cell density within a region hypothetically should more strongly restrict water diffusion within that region, leading to a decrease in ADC. Although this technique is appealing, the fact that ADC may be affected by necrosis16 and conflicting reports regarding the correlation between cell density and ADC11,15,22 complicate the above hypothesis. As more and more attempts are being made to use ADC as an imaging biomarker to evaluate the treatment response, understanding the actual magnitude of the correlation between ADC and cell density within the tissue is crucial.

This research was conducted to address this question by comparing glioma tissue samples that were stereotactically obtained during image guidance with presurgical ADC and fractional anisotropy (FA) maps calculated following diffusion tensor imaging (DTI). The investigation was further extended to compare DTI and metabolic imaging with 11C-methionine (MET) PET regarding the magnitude of the correlation between imaging and cell density to identify the most reliable imaging modality to evaluate glioma tissue characteristics.

Methods

Patient Selection

This study was approved by the local ethics committee and was found to conform to generally accepted scientific principles and ethical standards. We collected data from all glioma patients who underwent DTI and/or MET PET for presurgical examination, as well as intraoperative navigation-guided stereotactic tissue sampling between 2007 and 2011. There were 5 patients with low-grade glioma (1 pleomorphic xanthoastrocytoma, 2 diffuse astrocytomas, and 2 oligoastrocytomas) and 32 patients with high-grade glioma (8 anaplastic astrocytomas, 3 anaplastic oligoastrocytomas, and 21 glioblastomas). In total, 98 lesions were obtained from this patient population.

MRI

All patients were studied using a 3.0-T MRI scanner (Signa, GE Medical Systems) prior to surgery. T1-weighted images with and without gadolinium enhancement, T2-weighted images, and FLAIR images were acquired in all cases for delineation of tumors. DTI was performed in 27 of the 37 patients included for analysis. As DTI was performed primarily for surgical planning focusing on white matter fiber tracking, DTI was not performed in 10 patients in whom DTI was considered unnecessary as presurgical imaging. Images were acquired using a single-shot echo-planar imaging technique, with TE = 80 msec and TR = 10,000 msec. Diffusion gradient encoding in 25 directions, with b = 2000 sec/mm2 and an additional measurement without the diffusion gradient (b = 0 sec/mm2) were performed.26 A parallel imaging technique was used to record data with a spatial resolution of 128 × 128 and a field of view of 260 × 260 mm. A total of 50 sections were obtained, with a section thickness of 3 mm and no intersection gap. ADC and FA maps were processed using Diffusion Toolkit (Martinos Center for Biomedical Imaging, Massachusetts General Hospital; http://www.trackvis.org/dtk/).

PET

PET studies were performed using the Eminence instrument (Shimazu). MET (111–222 MBq; 3–6 mCi), synthesized according to the method of Berger et al.,3 was injected intravenously. Tracer accumulation was recorded over 12 minutes in 99 transaxial slices of the entire brain. Total activity from 20 to 32 minutes after tracer injection was used for image reconstruction. A 3-minute transmission scan in which 137Cs was used as the line source preceded data acquisition. Images were stored in 256 × 256 × 99 anisotropic voxels, with 1 × 1 × 2.6 mm for each voxel. MET PET was performed in all 37 patients included for analysis.

Image Fusion and Registration

After DTI and PET images were obtained, images were registered onto contrast-enhanced T1- or T2-weighted standard anatomical images using normalized mutual information with the Vinci image-analyzing software from the Max-Planck Institute for Neurological Research Cologne (http://www.nf.mpg.de/vinci/). Registration of the images was confirmed visually. The reported registration error for normalized mutual information is less than 1 mm.24 After image registration was completed, all image sets, including the standard anatomical MR images, ADC map, FA map, and MET PET, were converted to 256 × 256 × 256 isotropic, 1 × 1 × 1–mm images, to enable further voxel-wise analysis (Fig. 1).

FIG. 1.
FIG. 1.

Schematic illustration of the study design. Glioma tissues were stereotactically sampled by image guidance, and multimodal images such as FA, ADC maps, and MET PET were registered to anatomical images used for navigation. An image-histology comparison was performed preserving the location information of the sampled tissue. Figure is available in color online only.

Stereotactic Tissue Sampling

Thin-slice contrast-enhanced T1-weighted images were transferred to the neuronavigation system, and biopsy targets were planned for histopathological examination. Either the Stryker surgical navigation system (Stryker) or the VectorVision neuronavigation system (Brainlab) was used for image guidance. Tissue samples were obtained randomly from the tumor core for histological diagnosis or from the tumor periphery considered safe for resection for clinical purposes to evaluate viable tumor cells at the resection margin. To minimize the effects of brain shift during resection, biopsies were performed at the earliest stages of surgery. The accuracy of the navigation system was verified by visual confirmation of anatomical landmarks, such as cortical veins and the sulcus. To minimize the influence of brain shift, a Nelaton catheter was inserted under navigation guidance, aiming at the planned biopsy site to “anchor” the target of interest. Subsequently, the targeted area was biopsied by accurately tracing the catheter. Real-time navigation at each biopsy site in the tumor was performed to confirm the biopsy position.

Histopathological Analysis

As previous exploratory investigation showed that cell density, rather than MIB-1 activity, showed stronger correlation with MET PET,19 cell density was chosen as the reference for evaluating tissue characterization capability of the images. Formalin-fixed specimens were embedded in paraffin for histopathological analysis. Hematoxylin and eosin–stained specimens were evaluated to calculate cell density. Cell counting was performed at ×400 magnification under light microscopy (Nikon), and all cells were counted, except those that were apparently different from tumor cells, such as endothelial cells or lymphocytes. The area for the tumor cell count was 0.0497 mm2, and data for cell density were recorded as the mean from 3 different locations within each specimen.

Data Processing and Selection of Sampling Target-Specific Voxels of Interest

Four data sets (standard anatomical images, ADC map, FA map, and MET PET) were exported to in-house software written in MATLAB R2015 (MathWorks) for further analysis. A target voxel of interest (VOI) was set at the location recorded as the site of tissue sampling for image-tissue comparison. The target was obtained either by screenshot images or by stored coordinates of the navigation system. An average of 3 × 3 × 3 voxels was reported as the obtained value of the target site. A subset of data of the MET PET stereotactic tissue sampling data was previously published.1,12,19 The reported values were ADC, rADC (ADC value of the VOI divided by the ADC value contralateral to the VOI), FA, rFA (FA value of the VOI divided by the FA value contralateral to the VOI), and tumor-to-normal tissue ratio (T/Nr) of MET PET. For T/Nr of MET PET, the standardized uptake value of the contralateral tumor-unaffected gray matter at the axial plane of the thalamus was averaged, and the derived value was used to normalize the standardized uptake value in a voxel-wise manner, enabling calculation of T/Nr of the VOI.

Statistical Analysis

Statistical analysis was performed using JMP (version 8, SAS Institute). The Student t-test was used for 2-group comparisons. A linear regression model using the method for least squares was used for modeling 2 or more independent variables. A p value < 0.05 was considered statistically significant for both analyses.

Results

Sampling Tissue Characteristics

Because equal sampling of tissues was performed for contrast-enhanced and nonenhanced lesions, characteristics of the obtained tissue were evaluated. Among the 98 obtained samples, 41 samples were from contrast-enhanced and 57 were from nonenhanced lesions. The obtained cohort showed no difference in cell density of the tissue between contrast-enhanced and nonenhanced lesions (Fig. 2A).

FIG. 2.
FIG. 2.

Comparison of cell density, ADC, and rADC between gadolinium-enhanced (+) and nonenhanced (−) samples. A comparison of cell density (A), ADC (B), and rADC (C) between tissue samples obtained from gadolinium-enhanced and nonenhanced samples is presented. rADC was the only metric that exhibited a statistically significant difference between the 2 groups (p = 0.018, C). Figure is available in color online only.

DTI Metrics and MET PET for Contrast-Enhanced Versus Nonenhanced Lesions

ADC (30 enhanced, 45 nonenhanced), rADC (30 enhanced, 45 nonenhanced), FA (30 enhanced, 45 nonenhanced), rFA (30 enhanced, 45 nonenhanced), and T/Nr of MET PET (41 enhanced, 57 nonenhanced) were compared between contrast-enhanced and nonenhanced lesions (Figs. 2B and C and 3A–C). rADC was the only metric that exhibited a statistically significant difference between the 2 groups (p = 0.018, Fig. 2C). This was also true even when analyzing only high-grade gliomas (Supplementary Fig. A–F).

FIG. 3.
FIG. 3.

Comparison of FA, rFA, and T/Nr of MET PET between gadolinium-enhanced and nonenhanced samples. Comparison of FA (A), rFA (B), and T/Nr of MET PET (C) between tissue samples obtained from gadolinium-enhanced and nonenhanced samples is presented. No difference was observed in any of these comparisons. Figure is available in color online only.

ADC Metrics and Cell Density in Glioma

ADC, rADC, FA, rFA, and T/Nr of MET PET were correlated with cell density within the stereotactically obtained tissues. For ADC and rADC, only rADC showed a moderate but statistically significant negative correlation with cell density (p = 0.010, Fig. 4B). The linear regression fitting was as follows: rADC = −3.4 × 10−5 × [cell density(cells/mm2)] + 1.2.

When the analysis was restricted to high-grade glioma, both metrics showed statistically significant negative correlation with cell density with rADC showing stronger correlation (p = 0.020 for ADC [Fig. 4C] and p = 0.010 for rADC [Fig. 4D]). FA and rFA both showed little correlation with cell density suitable enough for cell density estimation by using those metrics (Fig. 5 and Supplementary Fig. G and H). Finally, for MET PET, the T/Nr of MET PET showed a superb correlation with cell density (p < 0.0001, Fig. 6), which was reproducible when analysis was restricted to high-grade glioma (p < 0.0001, Fig. 6 right).

FIG. 4.
FIG. 4.

Correlation between ADC metrics and cell density. Correlation between ADC (A) or rADC (B) and cell density of the exact location where these metrics were obtained using all samples is presented. rADC but not ADC showed a moderate but statistically significant negative correlation with cell density of the obtained tissue (p = 0.010, B). Analysis restricted by using only high-grade glioma is also presented (C and D). Both rADC and ADC showed a moderate but statistically significant negative correlation with cell density of the obtained tissue (p = 0.020, C; and p = 0.010, D). Figure is available in color online only.

FIG. 5.
FIG. 5.

Correlation between FA metrics and cell density. Correlation between FA (left) or rFA (right) and cell density of the exact location where these metrics were obtained is presented.

FIG. 6.
FIG. 6.

Correlation between T/Nr of MET PET and cell density. Correlation between T/Nr of MET PET and cell density of the exact location where these metrics were obtained using all samples is presented (left). T/Nr of MET PET showed a superb positive correlation with cell density (p < 0.0001). This correlation was confirmed when analyzing only high-grade glioma (right, p < 0.0001). Figure is available in color online only.

The linear regression fitting was as follows: MET PET T/Nr = 2.0 × 10−4 × [cell density(cells/mm2)] + 1.0, and as follows for high-grade glioma only-analysis: MET PET T/Nr = 2.1 × 10−4 × [cell density(cells/mm2)] + 1.0. This was also shown with multiple linear regression analysis of rADC and T/Nr of MET PET for the estimation of cell density. Although T/Nr of MET PET was statistically significant for modeling (p < 0.001), rADC was not (p = 0.06).

Discussion

Radiological assessment of brain tumors is one of the most crucial steps in the management of malignant gliomas. Without accurate information about the tumor, one can neither plan surgical and nonsurgical treatment nor perform optimal follow-up of the patient. To make matters more complicated, the introduction of antiangiogenic therapy such as bevacizumab to malignant glioma treatment has altered the way gliomas should be radiologically followed.10 As bevacizumab pharmacologically diminishes contrast enhancement of the tumors, which used to be considered one of the hallmarks of tumor recurrence, an alternative and objective imaging biomarker is required for follow-up of patients treated with bevacizumab. Response assessment criteria for high-grade gliomas published by the Response Assessment in Neuro-Oncology Working Group are the most reliable and widely used parameters for radiological response assessment during glioma treatment in real clinical settings.25 However, the fact that these criteria rely heavily on information regarding the performance status of the patient and the dosage of steroids prescribed to the patient brings to light the challenges of “pure radiological images” for revealing the treatment response in situ.

The theoretical hypothesis or assumption that increased high cell density restricts water diffusion led to the idea of using ADC to monitor tumor treatment.20 A decrease or increase in cell density of the lesion following either successful or unsuccessful treatment could increase or decrease ADC. The concept of a functional diffusion map or a parametric response map was introduced according to this theoretical assumption.5,8,9 Based on these theories, it is crucial that imaging parameters such as ADC guarantee precise estimation of cell density within the tissues. However, little histological confirmation has been performed.11,15,22 A study that compared autopsy specimens and diffusion images suggested that a decrease in ADC could be attributed to both high cell density and the presence of necrosis.16 The fact that contrast-enhancing lesions showed a lower rADC than nonenhancing lesions seems to support the above finding (Fig. 2C), as contrast-enhancing lesions are well known to be accompanied by necrotic tissue.

Next, when ADC and rADC were plotted as a function of cell density, a small but statistically significant negative correlation was observed for rADC but not for ADC itself (Fig. 4). This finding does support the use of ADC for analysis such as functional diffusion mapping5,8,9 or the CIMPLE method,6,7 which claims to independently visualize the proliferation and invasion capability of gliomas. Although these analytical methods use absolute ADC for calculation, the mathematical calculations are shifted toward examining the “relative” changes in ADC, which we have shown to be a valid metric as mentioned above.

Another observation was that FA showed little correlation with cell density (Fig. 5 and Supplementary Fig. G and H). Although a slight positive correlation seems to exist when cell density is greater than 5000 cells/mm2, it is unlikely that this weak correlation can be clinically useful. This result can be drawn from the fact that FA can be influenced by multiple factors. Not only cell density but also the magnitude of white matter disruption by tumor cell invasion can interfere with FA. FA would initially decrease in relation to disrupted white matter that is invaded by tumor cells,23 and at a certain point where the white matter is completely destroyed, FA would then be influenced only by the magnitude of cell density, which results in a gradual increase in FA (Fig. 5 and Supplementary Fig. G and H).2,15,17 Finally, the present study showed that MET PET is far more accurate and reliable for estimating cell density of glioma tissue than diffusion-based MR imaging (Figs. 5 and 6). The current study could be criticized for its limitation regarding the accuracy of biopsy sample locations and their image coregistration. Clear differences, however, in the reliability for cell density estimation between diffusion MR and MET PET, as shown in Figs. 4 and 6, elucidates the superiority of MET PET over diffusion MRI as an imaging modality for tissue characterization in gliomas. Positive correlation of MET PET and cell density was previously demonstrated both at the tumor core19 and tumor periphery.1,12 The current investigation is unique in that direct comparison between MET PET and diffusion MRI was performed regarding cell density estimation within glioma tissues. Furthermore, we have previously demonstrated that MET PET could be used as a surrogate imaging modality for glioma treatment response for immunotherapy.4 Although diffusion MR imaging is attracting more and more interest for tissue characterization of gliomas, our current data demonstrate that amino acid PET is still the de facto standard and that the limitation of diffusion MR imaging in gliomas requires caution. As MET PET is still not widely clinically available in many countries or institutions, it should be mentioned that more efforts should be made to prevail amino acid PET for glioma imaging in daily clinical settings.

Conclusions

Our current study elucidated that rADC showed a mild negative correlation with cell density within glioma tissue, and MET PET demonstrated a more robust and reliable estimation of cell density. We also revealed that FA showed a “J-shaped” curve when correlated with cell density. The presented data suggest that although diffusion MR imaging metrics such as ADC are still valid for estimating the treatment response of glioma tissues in terms of cell density, caution is needed regarding its limited accuracy compared with MET PET.

Acknowledgments

This investigation was supported by the Aichi Cancer Research Foundation, the SENSIN Medical Research Foundation, the Life Science Foundation of Japan, the Japanese Foundation for Multidisciplinary Treatment of Cancer, and JSPS KAKENHI (25462256 and 26670642).

References

  • 1

    Arita HKinoshita MKagawa NFujimoto YKishima HHashimoto N: 11C-methionine uptake and intraoperative 5-aminolevulinic acid-induced fluorescence as separate index markers of cell density in glioma: a stereotactic image-histological analysis. Cancer 118:161916272012

  • 2

    Beppu TInoue TShibata YYamada NKurose AOgasawara K: Fractional anisotropy value by diffusion tensor magnetic resonance imaging as a predictor of cell density and proliferation activity of glioblastomas. Surg Neurol 63:56612005

  • 3

    Berger GMaziere MKnipper RPrenant CComar D: Automated synthesis of 11C-labelled radiopharmaceuticals: imipramine, chlorpromazine, nicotine and methionine. Int J Appl Radiat Isot 30:3933991979

  • 4

    Chiba YKinoshita MOkita YTsuboi AIsohashi KKagawa N: Use of 11C-methionine PET parametric response map for monitoring WT1 immunotherapy response in recurrent malignant glioma. J Neurosurg 116:8358422012

  • 5

    Ellingson BMCloughesy TFLai AMischel PSNghiemphu PLLalezari S: Graded functional diffusion map-defined characteristics of apparent diffusion coefficients predict overall survival in recurrent glioblastoma treated with bevacizumab. Neuro Oncol 13:115111612011

  • 6

    Ellingson BMCloughesy TFLai ANghiemphu PLPope WB: Cell invasion, motility, and proliferation level estimate (CIMPLE) maps derived from serial diffusion MR images in recurrent glioblastoma treated with bevacizumab. J Neurooncol 105:911012011

  • 7

    Ellingson BMLaViolette PSRand SDMalkin MGConnelly JMMueller WM: Spatially quantifying microscopic tumor invasion and proliferation using a voxel-wise solution to a glioma growth model and serial diffusion MRI. Magn Reson Med 65:113111432011

  • 8

    Hamstra DAChenevert TLMoffat BAJohnson TDMeyer CRMukherji SK: Evaluation of the functional diffusion map as an early biomarker of time-to-progression and overall survival in high-grade glioma. Proc Natl Acad Sci U S A 102:16759167642005

  • 9

    Hamstra DAGalbán CJMeyer CRJohnson TDSundgren PCTsien C: Functional diffusion map as an early imaging biomarker for high-grade glioma: correlation with conventional radiologic response and overall survival. J Clin Oncol 26:338733942008

  • 10

    Hutterer MHattingen EPalm CProescholdt MAHau P: Current standards and new concepts in MRI and PET response assessment of antiangiogenic therapies in high-grade glioma patients. Neuro Oncol 17:7848002015

  • 11

    Iima MReynaud OTsurugizawa TCiobanu LLi JRGeffroy F: Characterization of glioma microcirculation and tissue features using intravoxel incoherent motion magnetic resonance imaging in a rat brain model. Invest Radiol 49:4854902014

  • 12

    Kinoshita MArita HGoto TOkita YIsohashi KWatabe T: A novel PET index, 18F-FDG–11C-methionine uptake decoupling score, reflects glioma cell infiltration. J Nucl Med 53:170117082012

  • 13

    Kinoshita MGoto TArita HOkita YIsohashi KKagawa N: Imaging 18F-fluorodeoxy glucose/11C-methionine uptake decoupling for identification of tumor cell infiltration in peritumoral brain edema. J Neurooncol 106:4174252012

  • 14

    Kinoshita MGoto TOkita YKagawa NKishima HHashimoto N: Diffusion tensor-based tumor infiltration index cannot discriminate vasogenic edema from tumor-infiltrated edema. J Neurooncol 96:4094152010

  • 15

    Kinoshita MHashimoto NGoto TKagawa NKishima HIzumoto S: Fractional anisotropy and tumor cell density of the tumor core show positive correlation in diffusion tensor magnetic resonance imaging of malignant brain tumors. Neuroimage 43:29352008

  • 16

    LaViolette PSMickevicius NJCochran EJRand SDConnelly JBovi JA: Precise ex vivo histological validation of heightened cellularity and diffusion-restricted necrosis in regions of dark apparent diffusion coefficient in 7 cases of high-grade glioma. Neuro Oncol 16:159916062014

  • 17

    Liu XTian WKolar BYeaney GAQiu XJohnson MD: MR diffusion tensor and perfusion-weighted imaging in preoperative grading of supratentorial nonenhancing gliomas. Neuro Oncol 13:4474552011

  • 18

    Lu SAhn DJohnson GLaw MZagzag DGrossman RI: Diffusion-tensor MR imaging of intracranial neoplasia and associated peritumoral edema: introduction of the tumor infiltration index. Radiology 232:2212282004

  • 19

    Okita YKinoshita MGoto TKagawa NKishima HShimosegawa E: 11C-methionine uptake correlates with tumor cell density rather than with microvessel density in glioma: a stereotactic image-histology comparison. Neuroimage 49:297729822010

  • 20

    Padhani ARLiu GKoh DMChenevert TLThoeny HCTakahara T: Diffusion-weighted magnetic resonance imaging as a cancer biomarker: consensus and recommendations. Neoplasia 11:1021252009

  • 21

    Report of Brain Tumor Registry of Japan 1984–2000. Neurol Med Chir (Tokyo) 49:SupplPS1PS962009

  • 22

    Stadlbauer AGanslandt OBuslei RHammen TGruber SMoser E: Gliomas: histopathologic evaluation of changes in directionality and magnitude of water diffusion at diffusion-tensor MR imaging. Radiology 240:8038102006

  • 23

    Stadlbauer ANimsky CBuslei RSalomonowitz EHammen TBuchfelder M: Diffusion tensor imaging and optimized fiber tracking in glioma patients: histopathologic evaluation of tumor-invaded white matter structures. Neuroimage 34:9499562007

  • 24

    Veninga THuisman Hvan der Maazen RWMHuizenga H: Clinical validation of the normalized mutual information method for registration of CT and MR images in radiotherapy of brain tumors. J Appl Clin Med Phys 5:66792004

  • 25

    Wen PYMacdonald DRReardon DACloughesy TFSorensen AGGalanis E: Updated response assessment criteria for high-grade gliomas: response assessment in neurooncology working group. J Clin Oncol 28:196319722010

  • 26

    Yoshiura TMihara FTanaka AOgomori KOhyagi YTaniwaki T: High b value diffusion-weighted imaging is more sensitive to white matter degeneration in Alzheimer's disease. Neuroimage 20:4134192003

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Kinoshita. Acquisition of data: all authors. Analysis and interpretation of data: all authors. Drafting the article: Kinoshita. Critically revising the article: Kinoshita, Hashimoto, Tanaka, Watanabe, Shimosegawa, Hatazawa. Approved the final version of the manuscript on behalf of all authors: Kinoshita. Statistical analysis: Kinoshita. Administrative/technical/material support: Kinoshita, Hashimoto, Yoshimine. Study supervision: Kinoshita, Yoshimine.

Supplemental Information

Online-Only Content

Supplemental material is available with the online version of the article.

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

INCLUDE WHEN CITING Published online February 26, 2016; DOI: 10.3171/2015.11.JNS151848.

Correspondence Manabu Kinoshita, Department of Neurosurgery, Osaka Medical Center for Cancer and Cardiovascular Diseases, 1-3-3 Nakamichi, Higashinari-ku, Osaka 537-8511, Japan. email: mail@manabukinoshita.com.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Schematic illustration of the study design. Glioma tissues were stereotactically sampled by image guidance, and multimodal images such as FA, ADC maps, and MET PET were registered to anatomical images used for navigation. An image-histology comparison was performed preserving the location information of the sampled tissue. Figure is available in color online only.

  • View in gallery

    Comparison of cell density, ADC, and rADC between gadolinium-enhanced (+) and nonenhanced (−) samples. A comparison of cell density (A), ADC (B), and rADC (C) between tissue samples obtained from gadolinium-enhanced and nonenhanced samples is presented. rADC was the only metric that exhibited a statistically significant difference between the 2 groups (p = 0.018, C). Figure is available in color online only.

  • View in gallery

    Comparison of FA, rFA, and T/Nr of MET PET between gadolinium-enhanced and nonenhanced samples. Comparison of FA (A), rFA (B), and T/Nr of MET PET (C) between tissue samples obtained from gadolinium-enhanced and nonenhanced samples is presented. No difference was observed in any of these comparisons. Figure is available in color online only.

  • View in gallery

    Correlation between ADC metrics and cell density. Correlation between ADC (A) or rADC (B) and cell density of the exact location where these metrics were obtained using all samples is presented. rADC but not ADC showed a moderate but statistically significant negative correlation with cell density of the obtained tissue (p = 0.010, B). Analysis restricted by using only high-grade glioma is also presented (C and D). Both rADC and ADC showed a moderate but statistically significant negative correlation with cell density of the obtained tissue (p = 0.020, C; and p = 0.010, D). Figure is available in color online only.

  • View in gallery

    Correlation between FA metrics and cell density. Correlation between FA (left) or rFA (right) and cell density of the exact location where these metrics were obtained is presented.

  • View in gallery

    Correlation between T/Nr of MET PET and cell density. Correlation between T/Nr of MET PET and cell density of the exact location where these metrics were obtained using all samples is presented (left). T/Nr of MET PET showed a superb positive correlation with cell density (p < 0.0001). This correlation was confirmed when analyzing only high-grade glioma (right, p < 0.0001). Figure is available in color online only.

References

  • 1

    Arita HKinoshita MKagawa NFujimoto YKishima HHashimoto N: 11C-methionine uptake and intraoperative 5-aminolevulinic acid-induced fluorescence as separate index markers of cell density in glioma: a stereotactic image-histological analysis. Cancer 118:161916272012

  • 2

    Beppu TInoue TShibata YYamada NKurose AOgasawara K: Fractional anisotropy value by diffusion tensor magnetic resonance imaging as a predictor of cell density and proliferation activity of glioblastomas. Surg Neurol 63:56612005

  • 3

    Berger GMaziere MKnipper RPrenant CComar D: Automated synthesis of 11C-labelled radiopharmaceuticals: imipramine, chlorpromazine, nicotine and methionine. Int J Appl Radiat Isot 30:3933991979

  • 4

    Chiba YKinoshita MOkita YTsuboi AIsohashi KKagawa N: Use of 11C-methionine PET parametric response map for monitoring WT1 immunotherapy response in recurrent malignant glioma. J Neurosurg 116:8358422012

  • 5

    Ellingson BMCloughesy TFLai AMischel PSNghiemphu PLLalezari S: Graded functional diffusion map-defined characteristics of apparent diffusion coefficients predict overall survival in recurrent glioblastoma treated with bevacizumab. Neuro Oncol 13:115111612011

  • 6

    Ellingson BMCloughesy TFLai ANghiemphu PLPope WB: Cell invasion, motility, and proliferation level estimate (CIMPLE) maps derived from serial diffusion MR images in recurrent glioblastoma treated with bevacizumab. J Neurooncol 105:911012011

  • 7

    Ellingson BMLaViolette PSRand SDMalkin MGConnelly JMMueller WM: Spatially quantifying microscopic tumor invasion and proliferation using a voxel-wise solution to a glioma growth model and serial diffusion MRI. Magn Reson Med 65:113111432011

  • 8

    Hamstra DAChenevert TLMoffat BAJohnson TDMeyer CRMukherji SK: Evaluation of the functional diffusion map as an early biomarker of time-to-progression and overall survival in high-grade glioma. Proc Natl Acad Sci U S A 102:16759167642005

  • 9

    Hamstra DAGalbán CJMeyer CRJohnson TDSundgren PCTsien C: Functional diffusion map as an early imaging biomarker for high-grade glioma: correlation with conventional radiologic response and overall survival. J Clin Oncol 26:338733942008

  • 10

    Hutterer MHattingen EPalm CProescholdt MAHau P: Current standards and new concepts in MRI and PET response assessment of antiangiogenic therapies in high-grade glioma patients. Neuro Oncol 17:7848002015

  • 11

    Iima MReynaud OTsurugizawa TCiobanu LLi JRGeffroy F: Characterization of glioma microcirculation and tissue features using intravoxel incoherent motion magnetic resonance imaging in a rat brain model. Invest Radiol 49:4854902014

  • 12

    Kinoshita MArita HGoto TOkita YIsohashi KWatabe T: A novel PET index, 18F-FDG–11C-methionine uptake decoupling score, reflects glioma cell infiltration. J Nucl Med 53:170117082012

  • 13

    Kinoshita MGoto TArita HOkita YIsohashi KKagawa N: Imaging 18F-fluorodeoxy glucose/11C-methionine uptake decoupling for identification of tumor cell infiltration in peritumoral brain edema. J Neurooncol 106:4174252012

  • 14

    Kinoshita MGoto TOkita YKagawa NKishima HHashimoto N: Diffusion tensor-based tumor infiltration index cannot discriminate vasogenic edema from tumor-infiltrated edema. J Neurooncol 96:4094152010

  • 15

    Kinoshita MHashimoto NGoto TKagawa NKishima HIzumoto S: Fractional anisotropy and tumor cell density of the tumor core show positive correlation in diffusion tensor magnetic resonance imaging of malignant brain tumors. Neuroimage 43:29352008

  • 16

    LaViolette PSMickevicius NJCochran EJRand SDConnelly JBovi JA: Precise ex vivo histological validation of heightened cellularity and diffusion-restricted necrosis in regions of dark apparent diffusion coefficient in 7 cases of high-grade glioma. Neuro Oncol 16:159916062014

  • 17

    Liu XTian WKolar BYeaney GAQiu XJohnson MD: MR diffusion tensor and perfusion-weighted imaging in preoperative grading of supratentorial nonenhancing gliomas. Neuro Oncol 13:4474552011

  • 18

    Lu SAhn DJohnson GLaw MZagzag DGrossman RI: Diffusion-tensor MR imaging of intracranial neoplasia and associated peritumoral edema: introduction of the tumor infiltration index. Radiology 232:2212282004

  • 19

    Okita YKinoshita MGoto TKagawa NKishima HShimosegawa E: 11C-methionine uptake correlates with tumor cell density rather than with microvessel density in glioma: a stereotactic image-histology comparison. Neuroimage 49:297729822010

  • 20

    Padhani ARLiu GKoh DMChenevert TLThoeny HCTakahara T: Diffusion-weighted magnetic resonance imaging as a cancer biomarker: consensus and recommendations. Neoplasia 11:1021252009

  • 21

    Report of Brain Tumor Registry of Japan 1984–2000. Neurol Med Chir (Tokyo) 49:SupplPS1PS962009

  • 22

    Stadlbauer AGanslandt OBuslei RHammen TGruber SMoser E: Gliomas: histopathologic evaluation of changes in directionality and magnitude of water diffusion at diffusion-tensor MR imaging. Radiology 240:8038102006

  • 23

    Stadlbauer ANimsky CBuslei RSalomonowitz EHammen TBuchfelder M: Diffusion tensor imaging and optimized fiber tracking in glioma patients: histopathologic evaluation of tumor-invaded white matter structures. Neuroimage 34:9499562007

  • 24

    Veninga THuisman Hvan der Maazen RWMHuizenga H: Clinical validation of the normalized mutual information method for registration of CT and MR images in radiotherapy of brain tumors. J Appl Clin Med Phys 5:66792004

  • 25

    Wen PYMacdonald DRReardon DACloughesy TFSorensen AGGalanis E: Updated response assessment criteria for high-grade gliomas: response assessment in neurooncology working group. J Clin Oncol 28:196319722010

  • 26

    Yoshiura TMihara FTanaka AOgomori KOhyagi YTaniwaki T: High b value diffusion-weighted imaging is more sensitive to white matter degeneration in Alzheimer's disease. Neuroimage 20:4134192003

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 191 191 9
PDF Downloads 185 185 6
EPUB Downloads 0 0 0

PubMed

Google Scholar