Diffusion imaging at 3.0 T with and without SENSE

Purpose: Diffusion imaging (DI) has become the cornerstone for the timely diagnosis of acute stroke. Still at 1.5T, diffusion imaging suffers from poor SNR. Moving from 1.5T to 3T significantly improves SNR, which may help improve sensitivity for subtle ADC changes. On the other hand, DI at 3T suffers from substantial image distortions, as susceptibility effects increase exponentially with field strength. An efficient approach to reduce susceptibility artifacts is by reducing echo train length in single shot DI by parallel imaging techniques such as SENSE (Sensitivity Encoding). We performed an intra-individual comparative trial to identify the optimal SENSE imaging technique for DI at 3T, and to investigate the impact of parallel imaging on image quality in patients presenting for work-up of suspected ischemic stroke.

Method and materials: A group of 85 patients underwent DI on a 3T whole body MR system twice, once with regular phase encoding (conventional DWI), and once with additional sensitivity encoding (SENSE DWI). A 6-element SENSE-compatible receive-only surface coil was used; the built-in body coil served for RF transmission and generation of the coil sensitivity profile. All patients underwent “regular“ DI (ssh SE-EPI, 128×128 matrix, 24 sections, 4mm thick, 2 b-values (0, 1000), with TE/TR 4283/79 ms), and DI with SENSE (the same ssh SE-EPI sequence with a SENSE factor of 3, yielding a TE/TR of 3141/69). Image quality, visibility and conspicuity of ischemic lesions were rated by 2 neuro-radiologists on a 5-point scale; results were compared for the SENSE and the regular DI pulse sequence. Overall SNR and lesion-to-background CNR of ischemic lesions were calculated for both sequences. Clinical film reading was evaluated regarding the confidence with which the presence of ischemic lesions could be demonstrated or excluded based on the cDWI and SENSE DWI data sets.

Results: Mean quantitative SNR in Sense DI compared to “regular“ DI was 70%. In spite of this numeric SNR reduction, SENSE DI scored consistently and significantly higher regarding image quality, owing to a substantial reduction of image distortions, particularly in areas close to the skull base or to the skull convexity, and a significant reduction of image blurring. Conspicuity of ischemic lesions was maintained at SENSE DI; in lesions close to the skull base or vertex, conspicuity was higher. In 4 patients, small microembolic lesions to fronto-dorsal and temporal cortex, were only diagnosed on the SENSE DI.

Conclusion: At 3.0T, DI with SENSE allows a substantial reduction of image artifacts. In fact, the degree of image distortions in the regions close to the skull base are comparable with DI at 1.5T, while the high sensitivity for ischemic lesion is maintained. SENSE may help to actually exploit the full potential of high field DWI. We conclude that SENSE (or parallel imaging in general) should be considered an integral part of clinical diffusion imaging at higher field strength.