The Value of CBF Combined With Temporal Information in Grading High-Grade Astrocytomas: A Multi-Inversion-Time Arterial-Spin-Labeling Magnetic Resonance Study
Shuang Yang1, Tianyi Qian2, Jianwei Xiang1, Yingchao Liu3, Fei Gao1, Peng Zhao3, Josef Pfeuffer4, Guangbin Wang1, and Bin Zhao1

1Shandong Medical Imaging Research Institute, Shandong University, Jinan, China, People's Republic of, 2MR Collaborations NE Asia, Siemens Healthcare, Beijing, China, People's Republic of, 3Shandong provincial Hospital, Shandong University, Jinan, China, People's Republic of, 4Application Development, Siemens Healthcare, Erlangen, Germany


This study aimed to demonstrate the feasibility of multi-inversion-time arterial spin labeling (mTI-ASL) for differentiating between WHO III and WHO IV grade astrocytomas, as well as the added value of bolus arrival time (BAT) information in evaluating tumor perfusion. In the first part of this study, we evaluated the reproducibility of mTI-ASL in healthy subjects, and then mTI-ASL was used to evaluate 45 astrocytoma patients. There was no major variation between two consecutive mTI-ASL measurements in healthy volunteers. Furthermore, mTI-ASL provided valuable information for the classification of astrocytomas, while BAT added relevant information for grading by estimating the temporal dynamics of local tumor-mass perfusion.


Systematic and accurate glioma grading has very high diagnostic relevance, especially for high-grade astrocytomas, because patient outcome largely relies on a timely and appropriate surgery and therapy. In previous studies, CBF obtained by using standard 3D arterial spin labeling (ASL) MRI with a single inversion time (TI) could only differentiate between low- and high-grade gliomas, but was unable to provide sufficient information to differentiate between WHO III and WHO IV grades. The multi-parametric multi-TI ASL (mTI-ASL) technique allows for the quantification of CBF and provides accurate bolus arrival time (BAT) estimations to better characterize brain perfusion changes. The purpose of this study was to demonstrate the feasibility of mTI-ASL in differentiating high-grade astrocytomas and the value of bolus arrival time (BAT) in evaluating tumor perfusion.


Data were collected on a MAGNETOM Skyra 3T MR scanner (Siemens Healthcare, Erlangen, Germany) with a 32-channel head coil. In the first part of this study, ten healthy volunteers (7 males; mean 50 years; range, 21–62 years) underwent two consecutive mTI-ASL MRI scans. Then, we performed a volume-based paired t-test analysis (VBA) by SPM8. 1 The test-retest variability (TRV) of each voxel was calculated using this equation: 200 %* (test 1 - test 2)/ (test 1 + test 2). The second part, forty-five patients (24 males; mean 52 years; range, 15–73 years (13 glioblastomas , World Health Organization [WHO] IV; 15 anaplastic astrocytomas, WHO III; 17 diffuse astrocytomas, WHO II). The mTI-ASL images parameters: TR/TE = 4600/22 ms, FOV = 220 × 220 mm2, GRAPPA (PE) 2, slice thickness = 4 mm, voxel size = 3.4 × 3.4 × 4.0 mm3, 20 slices, PASL PICORE Q2TIPS bolus length = 700 ms, 16 equidistant TIs between 500 to 4100 ms, and total acquisition time = 5:09 min including an M0 scan. The normalized values (nCBF) were obtained by dividing the absolute value of tumor area by that of the contralateral normal-appearing white-matter area for each patient, and were marked as nCBF-mTI (derived from mTI-ASL), nBAT, and nCBF-sTI (derived from single TI-ASL). Raw coupled control and labeled images (TI = 1920 ms) were used to evaluate the efficacy of the CBF derived from sTI-ASL (vs. mTI-ASL) for grading tumors.


Figure 1 shows the average of the TRV of all healthy volunteers. There was no significant difference between the two consecutive measurements, and the TRV was around 10% in most cortex areas. The nCBF-sTI (P = 0.017), nCBF-mTI (P < 0.001) and nBAT (P = 0.043) could all independently differentiate LGGs from HGGs grades. However, a significant difference was observed only in nCBF-mTI between the WHO III and IV groups (P = 0.039). The nBAT can differentiate the WHO II and III (P = 0.046), but there is no difference between the other group pairs. The ROC analysis of nCBF-mTI demonstrated an AUC = 0.728, sensitivity = 69.2%, and specificity = 80.0% at the best cut-off point in grading WHO III and WHO IV groups. The diagnostic accuracies for simultaneous distinction between WHO II, III, and IV grades were calculated. The nCBF-mTI had the best performance, with an overall accuracy of 60.0% compared with the nCBF-sTI (33.3%) and nBAT (53.3%). When combing the nCBF-mTI with nBAT, the diagnostic accuracy effectively improved from 60% to 68.9% compared to the use of nCBF-mTI alone.


In previous reports, CBF obtained by using sTI-ASL could only help differentiate between low- and high-grade gliomas. 2 Here, even though we obtained similar results with CBF-sTI, CBF obtained by mTI-ASL demonstrated to be of great value for discriminating between tumor grades. Moreover, nCBF-mTI allowed for differentiation between pairs of WHO II, III, and IV tumor grades. The CBF-mTI value was significantly higher than the CBF-sTI value, especially in high-grade astrocytomas, suggesting that the mTI-ASL is more sensitive, while the CBF obtained by sTI-ASL was likely underestimated, especially in high-grade gliomas with BATs shorter than in normal tissue. Our results suggest that the BAT can be instrumental for increasing the accuracy of tumor grading. To accurately estimate CBF in tumor when blinded to their histological grade, several repetitive sTI-ASL scans would be needed to capture the heterogeneous BAT value distribution within the tumor area. Our results show that mTI-ASL effectively provide this information within a single measurement.


mTI-ASL demonstrated high test-retest reliability in healthy subjects. Furthermore, this technique could effectively support the differentiation of high grade astrocytomas, by combining CBF with the dynamic perfusion information provided by BAT, increasing diagnostic accuracy for astrocytomas.


No acknowledgement found.


1. Mezue M, Segerdahl AR, Okell TW, et al. Optimization and reliability of multiple post labeling delay pseudo-continuous arterial spin labeling during rest and stimulus-induced functional task activation. Journal of cerebral blood flow and metabolism: official journal of the International Society of Cerebral Blood Flow and Metabolism. 2014; 34: 1919-1927

2. Wolf RL, Wang J, Wang S, et al. Grading of cns neoplasms using continuous arterial spin labeled perfusion mr imaging at 3 tesla. Journal of magnetic resonance imaging : JMRI. 2005;22:475-482


Figure 1: Voxel-based results showing high test-retest reproducibility of mTI-ASL (average TRV for ten subjects). The test-retest stability index is mainly around 10% for both BAT and CBF.

Figure 2: Examples of high grade astrocytomas. (a) Images from a 50-year-old female with an anaplastic astrocytoma (WHO III) in the left temporal lobe. (b) A 56-year-old female patient with glioblastoma (WHO IV) in the left parietal lobe. Hyperperfusion is demonstrated in the CBF-mTI map, and the BAT is shortened.

Figure 3: The diagnostic accuracies (as percentages) of nCBF-mTI, nBAT, and nCBF-sTI in distinguishing between WHO II, III and IV grades of astrocytoma as calculated. All the diagnostic accuracy values are labeled on top of the columns.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)