PURPOSE:

Gadolinium based contrast agents (GBCA’s) decrease T₁ times in pathologic regions of the brain where there is disruption of the blood-brain barrier; hence these regions appear brighter on T₁-weighted scans. This improves image contrast and lesion visualization with increased sensitivity and specificity [1]. Around 30–40% of clinical MR applications have applied contrast agent, including intravenous injection of GBCA’s for brain tumors [2]. 3 dimensional fast gradient echo sequences such as MPRAGE offer thinner slices, near seamless reformatting options, and lower specific absorption rates than 2 dimensional spin-echo based technique. In clinical practice, contrast-enhanced (CE) magnetic resonance imaging (MRI), is performed with either spin-echo or gradient-echo T₁-weighted sequences, and there has been controversy regarding which sequence is better in detecting brain tumor [3-6]. In our opinions, the lack of using optimized sequences in the comparison for brain tumor detection caused this controversy. The purpose of our study is to optimize MPRAGE sequence using computer simulation to improve brain tumor enhancement and detection. we compare its performance with that of a Siemens default MPRAGE sequence after Gadavist administration at 3.0T.

METHOD

Simulation: We simulated contrast efficiencies between CE tumor tissue and normal white matter (WM) (CE tumor-WM), using Bloch’s equation, based on the values of T₁, T₂, and proton density of the WM, GM and Gadavist enhanced tumor tissue at 3.0 T: 1400/850/250 ms, 100/90/180 ms, and 0.75/0.65/1.0, respectively [7, 8].

In vivo experiment: Four subjects with brain tumor were scanned on a 3T Siemens Skyra scanner that was equipped with a 32-channel head coil. All subjects were scanned pre and post contrast administration of 0.1 mmol/kg Gadavist. Brain images were obtained using MPRAGE sequence with FOV 256 x 232 mm², matrix 256 x 232, number of slices 160, slice thickness 1 mm. Our optimal imaging parameters: TI 450 ms, Flip angle (FA) 19^o, TR 1350 ms, total scan time (TA) 2min and 40s. Siemens default imaging parameters: TI 900, FA 8^o, TR 2300 ms, and 6min and 9s.

Evaluation: (1) The performance of the optimization was then evaluated using contrast efficiency of CE tumor-WM in post-contrast images, which were defined as contrast per square root of TA (second), CN_eff = Contrast / (TA^0.5) ; and (2) The percentage of pre-contrast (pre) and post-contrast (post) signal intensity (SI) alteration in the region of tumor lesion due to enhancement, which was denoted as β ratio; β=100%x(SI(post)-SI(pre))/ SI(pre).

RESULT

Fig. 1 show in vivo brain images acquired with MPRAGE sequence after the administration of Gadavist. Visually, the size of enhanced lesion looked more or bigger in the images acquiring using our optimized MPRAGE sequence (Fig.1b) than those acquired by Siemens default MPRAGE sequence (Fig. 1a). Further quantitative analysis in Table 1 indicated that (1) The size of detected enhanced tumor using our optimal protocol is 4% more than that using Siemens default’s protocol. This implies that our optimized protocol may increase the detection sensitivity for tumor lesion; (2) β ratio produced by our protocol was 326% which was comparable to that of 328% produced by Siemens default protocol. However, the scan time of our protocol was just about 40% of that of Siemens default protocol. It implied that our optimization increased scan efficiency without sacrificing the detection accuracy; (3) Our protocol improved CE tumor-WM contrast efficiency by 64%. It showed a great potential in the use of low dose contrast agent for detecting tumor lesion. The computer simulation in Fig. 2 indicated that the optimal FA and TI are 19^o and 450 ms, respectively. The simulation also suggested that our optimal protocol would improve 80% CE tumor-WM contrast efficiency, compared with Siemens default protocol. The simulation is in agreement with in vivo experiments.

DISCUSSION AND CONCLUSION

Potential reasons why there have been inconsistent in the evaluation of MRI protocols for brain tumor detection may include: (1) lack of sequence optimization; (2) use of improper MRI properties (T₁, T₂ and proton density) for different brain tumor types; (3) ignorance of the effect of contrast agent lifetime/transmit time ; (4) inconsistent evaluation metrics. In this study, we optimized CE MPRAGE sequence to improve the detection of brain tumor. Our optimized protocol greatly shortened scan time by around 60% without sacrificing tumor delectation sensitivity. The proposed optimization methodology and obtained results have a great potential in clinical applications. As a practical guideline, the optimal CE MPRAGE protocol after Gadavist administration is: TI 450 ms, FA 19^o and TR 1350 ms.