Perfusion MR imaging of enhancing brain tumors: Comparison of arterial spin labeling technique with dynamic susceptibility contrast technique
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/neuroindia.NI_871_16
Source of Support: None, Conflict of Interest: None
Objective: Arterial spin labeling (ASL) magnetic resonance (MR) perfusion is a noninvasive and repeatable method for quantitatively measuring cerebral blood flow (CBF). This study aims to compare measurements of ASL-derived CBF with dynamic susceptibility contrast (DSC) MRI in the assessment of enhancing brain tumors (primary and metastatic), with an aim to use ASL as an alternative to DSC.
Keywords: Brain tumor imaging, dynamic susceptibility contrast, magnetic resonance imaging, normalized tumor blood flow, normalized tumor blood volume, pseudo-continuous arterial spin labeling
The two most common methods of magnetic resonance imaging (MRI) used for measuring brain perfusion are dynamic susceptibility contrast (DSC) and arterial spin labeling (ASL), with more extensive clinical experience existing for DSC perfusion MRI. Perfusion MRI estimates tumor neoangiogenesis, which helps in tumor grading, guiding stereotactic biopsy, surgical planning, differentiating radiation necrosis from recurrent tumor, and assessing therapeutic response in clinical trials of new anti-angiogenic agents.,,,, Both DSC and ASL perfusion MRI have been compared in various studies of brain tumors.,,,,,,,
ASL technique was first reported by Deter et al., who provided absolute quantification of CBF using magnetically-labeled blood-water as an endogenous tracer, which makes ASL a promising technique for studying perfusion in patients with renal failure and in those who require repetitive follow-ups. Recently developed three-dimensional pseudo-continuous arterial spin labeling (3D PCASL) sequence, incorporating high-field, parallel imaging with background suppression, provides increased sensitivity. These features may move ASL from the research stage towards clinical usage. DSC-MRI based on T2-weighted imaging measures perfusion using an exogenous contrast agent. Comparative studies between ASL and DSC-MRI have found a close linear correlation between these two perfusion techniques.,,,,,, The present study was conducted to compare the role of perfusion parameters obtained from ASL and DSC perfusion MRI in enhancing brain tumors (primary and metastatic) detected on the conventional MRI in our set of population, and to assess the feasibility of using ASL as an alternative to DSC.
Thirty patients (age range = 19–71 years; 16 male and 14 female patients) with newly diagnosed enhancing brain tumors (primary and metastatic) were enrolled in this institutional ethics committee (IEC)-approved prospective study (July 2013 to September 2014). Informed consent was obtained from each patient prior to their undergoing the perfusion MRI. Patients with nonenhancing lesions, previous surgical interventions, and imaging with susceptibility artifacts that affects evaluation, were excluded from the study. The final diagnosis of the patients was obtained based on the clinical background, histopathology (after acquiring the perfusion MRI), and follow-up.
Magnetic resonance imaging acquisition
MRI was performed on a 3T MR scanner (Signa HDxt, General Electric, Milwaukee, USA) with a 16-channel head-neck-spine coil. The lesions were evaluated with conventional imaging using routine axial T2/T1-weighted, fluid-attenuated inversion recovery (FLAIR), diffusion-weighted imaging (DWI) sequences including 3D PCASL (combining pseudo continuous arterial spin labeling [PCASL], and three dimensional [3D] fast spin echo encoding and spiral trajectory acquisition techniques), and DSC perfusion MRI followed by postcontrast axial T1-weighted and 3D BRAVO (brain volume) sequences with same planning of DSC and ASL, respectively, to overlay CBF, and relative cerebral blood volume (rCBV) perfusion maps. Acquisition parameters for the 3D PCASL are described in our previous article by Soni et al. DSC-MRI, a gradient echo EPI (echo planar imaging) sequence, was performed with the following parameters: TR = 2s/TE = 20.2 ms/slices = 20/slice thickness = 6 mm/spacing = 0/matrix = 128 × 96/average = 1/FOV = 24 cm × 24 cm/bandwidth = 62.5 kHz/scan time = 1 min 20s/axial plane. A series of 42 such scans were obtained at 1 s/image during the first pass of dynamic intravenous administration of gadolinium-DTPA (0.1 mmol/kg) at a rate of 3–5 ml/s followed by 20 ml saline flush at the same rate.
Image processing and data analysis
Post processing of ASL and DSC data was performed on ADW4.4 GE workstation using Functool 3D ASL and Brain stat Software to obtain color-coded CBF, rCBF, and rCBV maps, respectively, from each patient [Figure 1]. At least 3 regions of interest (ROIs) of 4–7 mm 2 were drawn over the tumor region showing the highest perfusion value in the CBF, rCBV, and rCBF maps, avoiding the regions of vessels, calcification, hemorrhage, cyst, and necrosis [Figure 2].
For normalization, to avoid age and patient dependent CBF variations,
ROIs were placed in the the contralateral normal appearing frontal GM and periventricular WM [Figure 2]. The 3 ROI values were averaged to obtain the mean. The mean value of tumor ROI was divided by the mean value of contralateral normal GM and WM ROI to estimate normalized nTBF and nTBV values based on the CBF and rCBV maps.
ASL nTBF was compared with DSC nTBF values using paired t-test. Linear regression and Pearson's correlation were employed to examine the correlation between DSC and ASL-derived CBF values in tumor as well as normal GM and WM region. These statistical analyses were performed using the SPSS, version 16.0 software package (SPSS Inc., Chicago, IL, USA).
All the patients recruited had a newly detected tumor and had not undergone any prior treatment. Tissue for histological analysis had been obtained at biopsy or during surgical resection of the tumor, and revealed 16 meningiomas, 6 gliomas, 3 metastases, 2 cerebellopontine angle schwannomas, and 1 central neurocytoma. Two patients did not undergo a histopathological examination; however, they had the characteristic MR features of a low-grade glioma and are on regular follow up.
Quantitative normalized perfusion value measurements
Quantitative perfusion values normalized to GM and WM (i.e., ASL nTBF, DSC nTBF, and DSC nTBV) as reference region were obtained in all 30 tumors [Table 1] and [Table 2]. Distribution of ASL nTBF, DSC nTBF, and DSC nTBV showed the highest values for meningiomas and the lowest values for gliomas among all the tumors [Table 3] and [Figure 3].
Pearson's correlation coefficient between ASL and DSC normalized perfusion values
A highly strong correlation coefficient between ASL nTBF and DSC nTBF was found in all 30 tumors when normalized to GM [Figure 4] rather than to WM as the reference region [Table 4], with the highest correlation being seen in gliomas. Similarly, correlation coefficient for DSC nTBF and DSC nTBV was also calculated and found to be high (r = 0.83 with R2= 0.701) when normalized to GM. Our study results support that PCASL could be used as an alternative to DSC-MRI in non-invasively estimating the in vivo intracranial tumor blood flow.
The mean ASL absolute tumor blood flow (TBF) values were also calculated among all tumors and subtypes [Table 5]. Similarly, the mean of normal absolute ASL CBF values were calculated for normal GM and WM separately in all 30 patients [Table 5].
The purpose of this study was to quantitatively compare perfusion values from ASL with DSC MRI in brain tumors, with an aim to evaluate the usage of ASL as an alternative to DSC in routine clinical and research studies. In routine clinical practice, brain tumor perfusion is mainly calculated from rCBF and rCBV perfusion maps using DSC MRI. ASL is a relatively new, evolving technique for clinical and research studies. The types of contrast agents (exogenous gadolinium-based contrast agent in DSC MRI and endogenous tracer in the ASL technique) and the post-processing algorithm used in these two perfusion techniques provide different perfusion values.,,,, DSC-MRI is based on T2-weighted imaging and requires a faster dynamic echo-planar imaging scan with parallel imaging to cover the whole brain., ASL provides absolute CBF values, which is difficult with DSC MRI because of a careful selection of arterial input function (AIF) and advanced post-processing deconvolutional algorithms. ASL was, therefore, quantitatively compared with DSC MRI by means of normalized perfusion values in this study. Our study demonstrated evidence that supported a close linear correlation between normalized perfusion values derived from ASL and DSC [Figure 4], and the results are consistent with previous studies [Table 6].,,,
The correlation coefficient value in our study is in close approximation with the study done by Van Westen et al., where GM was used as the reference region for normalization. The slight difference which is seen when we compared our results with that of other studies may be due to reference region choice and tumor composition. Our ASL nTBF and DSC nTBF values normalized to WM showed a weak correlation [Figure 4]. Underestimation of WM CBF by ASL due to the long transit times and higher water content of the normal contralateral white matter in brain tumors make WM a questionable reference region.,,
Most previous MRI studies performed DSC nTBV for the evaluation of brain tumor perfusion and showed a diagnostic comparison with CBF.,,,, As seen in the previous studies, a strong correlation was seen between DSC nTBF and DSC nTBV (r = 0.837) [Figure 4], supporting the fact that DSC nTBF may be as good as DSC nTBV for the assessment of brain tumor perfusion. Ata et al., also reported a strong correlation (r = 0.91, P> 0.001) in the comparison of relative regional DSC rCBF and rCBV values.
Correlation coefficient in gliomas was reported to be stronger than meningiomas [0.98 vs 0.814]. Although the sample size of gliomas in our study was small (n = 8), nTBF ASL and nTBF DSC values lie in the range that were similar to the findings of the previous studies [Table 7]. The slight difference from the other studies could be mainly due to a higher number of patients with a high-grade gliomas recruited by them. In our study, among all the tumors, meningiomas have shown the highest ASL nTBF (3.7 ± 1.63) and absolute ASL TBF (177 ± 89.2 ml/min/100 g) values, which is in agreement with a previously published study [Table 3] and [Table 5]. Thus, our study results indicate that ASL can be used as an alternative to DSC.
Similar good correlation has been reported for overall brain tumors in a few recent studies.,,,,,,, Close linear correlation (r = 0.83; P<.005) was found by Warmuth et al., between ASL and DSC in histologically proved brain tumors. Weber et al., suggested that ASL and DSC MRI techniques determining rCBF in brain metastasis after stereotactic radiosurgery allow prediction of the treatment outcome with equal confidence. Jarnum et al., also found a good correlation between the normalized ASL and DSC nTBF values (r = 0.82) in 28 enhancing brain tumors. Similarly, Lehmann et al., analyzed patients of gliomas, metastases, and meningiomas with ASL and DSC MRI sequences, and concluded that ASL is a good alternative to DSC MRI. Other recent studies, have also found a good correlation between ASL and DSC techniques, and suggested that ASL perfusion can be used in clinical practice.,,, More recently, Ata et al., also reported a strong correlation (r = 0.86) with a high sensitivity and specificity between DSC and multiphase ASL in 27 brain tumor patients. In general, most of the studies support the opinion that ASL could be a noninvasive alternative to DSC.
ASL offers several advantages over DSC due to its noninvasive nature, simple post-processing, no exogenous contrast use, being safe in patients with impaired renal function, and having the ability to provide quantitative tissue perfusion. This renders ASL an ideal technique for routine clinical and research studies. Certain pathological conditions severely impair the blood–brain barrier and result in extravascular contrast agent leakage, which results in miscalculation of the DSC perfusion mapping. There are a few limitations of ASL, such as generating only the CBF perfusion map, which are of relatively lower resolution and takes a longer acquisition time than the DSC perfusion map.,
Although data from our current study is encouraging, the study included a small number of patients, especially those with a high or a low-grade glioma, and patients with metastases.
This study has shown a very strong correlation between 3D PCASL and DSC MRI perfusion values and support the possibility of ASL being used as an alternative to DSC MRI for the evaluation of brain tumors. Nowadays, with the more extensive use of 3T scanners, improved ASL perfusion sequences with whole brain coverage, low cost, safety in impaired renal function, easy availability, and simple quantification are possible to execute. The ASL method can be introduced in everyday clinical practice and has a wide range of applications including the monitoring and follow-up of brain tumors while avoiding the requirement of repeated contrast injections.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7]