Introduction

Cerebral infarction (CI) is one of the worldwide leading causes of disability and death. An accurate measure of the severity of ischemic insult and the resulting prognosis is needed for personalized diagnosis and treatment evaluation. Routine MRI scanning protocols for CI generally consist of individual scans of T₁W or T₁-FLAIR, T₂W or T₂-FLAIR and diffusion weighted imaging (DWI) and susceptibility weighted imaging (SWI)¹, all except SWI are based on 2D spin echo (SE) acquisitions. These scans have been widely proven and accepted clinically, yet still have their own short planks. The major issue of 2D imaging methods lies in their thick slices with spatial gaps, which lead to strong partial volume effects and/or failure to capture small lesions located at the gaps. Second, different scans may have different spatial settings, and patient motion between scans may lead to lesion mismatch. Thirdly, total scan times of a complete CI protocol set, including the 3D SWI, can be too long for CI patients.
In this work, we conducted an initial test of a newly developed single-scan multi-parametric imaging method, i.e. MULTIPLEX², on a CI patient to evaluate its potential to reveal pathological and anatomical information.

Methods

With local IRB approval, one CI patient (male, 65/yro) with subacute stroke in the corona radiata region was recruited and scanned with CT and MRI 4 days after onset. CT results (not shown) reported no cerebral hemorrhage.
All MRI scans were performed on a 3T system (uMR780, United Imaging Healthcare, Shanghai ,China) with a 24-channel head coil. Routine scans included: 1) axial T₁-FLAIR (TI=870ms, TE/TR=10/2010ms, 2x interpolated in-plane resolution 0.45x0.45mm², TA=1’29” with 2.3x compress sensing (CS)); 2) axial T₂-FLAIR (TI=870ms, TE/TR=10/2010ms, 2x interpolated in-plane resolution 0.5x0.5mm², TA=1’20” with 2.3x CS); 3) an axial T₂-FSE (TE/TR=96/4000ms, 2x interpolated in-plane resolution 0.38x0.38mm², TA=40” with 2.2x CS); and 4) axial EPI-DWI (b = 0&1000s/m, TE/TR=78/4000ms, 2x interpolated in-plane resolution 0.72x0.72mm², TA=60” with 2x parallel imaging). All routine scans generated 20 slices with 5mm thickness and 1.5mm gap, and combined scan time of 4’29”.
For MULTIPLEX scan, the scanning protocol was: FA₁/FA₂=4°/16°, TR₁/TR₂=8.2/27.9ms, 5x echoes with TE=3.76~23.36ms, bandwidth=230Hz/px, in-plane resolution (not interpolated) 0.70x0.70mm², slice thickness=1.5mm (interpolated from 3mm), and TA=3’19” with 4x parallel imaging using wave readouts³. Virtual FLAIR (vFLAIR) images were generated via the MULTIPLEX tier-3 post processing² using proton density-weighted (PDW) images and T₁ map.

Results

Figure 1 compares the MULTIPLEX vFLAIR images to the routine T₂-FLAIR images on the capacity of revealing hyperintense infarct lesions in 3D views. Figure 2 compares the images between routine images (DWI/T₂-FSE/T₁- and T₂-FLAIR) and MULTIPLEX images on a typical slice, showing similarities in image contrast information between MULTIPLEX images vs. routine scans of T₂-FSE, T₁-FLAIR and T₂-FLAIR. Figure 3 shows another slice of images, where a cerebral microbleed (CMB) located in right temporal white matter was clear seen in MULTIPLEX images but was obscured in all routine scans.

Discussion

In this study, we have for the first time evaluated the potential of the MULTIPLEX² method for CI imaging, suggesting MULTIPLEX not only can offer similar information as with routine T₂-FSE, T₁- and T₂-FLAIR scans, but also offers additional capacity for CMB detection.
T₂-FLAIR is basically the gold standard for detecting hyperintense infarct lesions⁴. However, due to the long scan time in 3D mode⁵, T₂-FLAIR generally produce 2D slices that are thick (for sufficient SNR) and with gaps (to avoid RF crosstalk),for small lesions this not only may lead to reduced detectability due to partial volume effects, but also have the risk of non-detection for those physically located between slice gaps. Therefore some CI protocols would include a second or even a third T₂-FLAIR scan with orthogonal planes to ensure the capture of all lesions, at the cost of prolonged scan time. On the other hand, PDW images have also been demonstrated⁶ to reliably detect white matter lesions, and by combining the PDW contrast and CSF suppression effect via virtual inversion recovery, the resultant MULTIPLEX vFLAIR images was able to detect all lesions with better delineation in high resolution 3D views (Fig.1). However, it should be noted that the lesion contrast in MULTIPLEX vFLAIR was not as salient as T₂-FLAIR, making some of the lesions not easy for detection unless using T₂-FLAIR as reference. Therefore, MULTIPLEX vFLAIR can be used as supplementary information to T₂-FLAIR.
The MULTIPLEX’s T₁ map and aT₁W offered highly similar anatomical contrast with routine T₂-FSE and T₁-FLAIR (Fig.2), suggesting the possibility of substitution for these two scans. However, DWI remained the most sensitive and only method to detect the ischemic infarct core (Fig.2), the combination of its lack of SWI contrast suggested acute/sub-acute stage. With susceptibility sensitivity, several MULTIPLEX images detected a CMB in the right temporal lobe (Fig.3), which was not shown in any routine images nor the patient’s CT scan. The functionality of SWI can be fully realized with MULTIPLEX. And with state-of-the-art acceleration techniques³, the additional cost on scan time can be minimal.

Conclusion

In conclusion, we have demonstrated the potential of MULTIPLEX method for CI imaging, which may have the potential as the supplementary or replacement to some of the routine scans.

Acknowledgements

No acknowledgement found.