Parameters optimization of diffusion tensor MR imaging of the human calf at 3 tesla

Objective: To demonstrate the feasibility of DTI in human calf with body phased-array coil and surface coil of spine as receiving coil on 3 T system, and to optimize the parameters of sequence, including slice thickness and b-value. Methods: Fifteen healthy volunteers were recruited in this study and randomly divided into three groups. The DTI sequence for head was performed on calf in the first group (5 cases), and the sequence parameters were optimized based on the deficits of the raw and the post-processed DTI images. Then, different slice thickness were applied in the senond group (5 cases) to optimize the slice thickness, and this optimized parameter with the highest score based on quality of the post-processed DTI images was applied in the next step. Finally, different b values were applied in the last group to optimize this parameters. The b value with the highest score based on the quality of the post-processed was the proper one. Results: Three problems existed in the raw and the post-processed images, when the DTI sequence for brain was used for the calf. First, the SNR of raw images is extremely low. Second, the muscle were unclear on the image with parts of signal lose, especially in the anterior tibialis muscle. Finally, the artifacts due to chemical shift and ghost are quite serious. The scores for muscle display quality with slice thickness of 4 mm, 5 mm and 6 mm were (7.0 ± 0.0), (8.6 ± 0.9) and (9.0 ± 0.0) score respectively, the signal loss scores were (5.0 ± 0.0) and (12.8 ± 2.6) and (13.8 ± 2.2) score respectively, and the general score were (22.0 ± 0.0) and (30.1 ± 3.8) and (31.0 ± 2.2) score respectively, and the general score were (22.0 ± 0.0) and (30.1 ± 3.8) and (31.0 ± 4.1) score respectively. The differences of above scores were significant among different slice thickness (F-value were 21.000 and 30.544 and 12.390 respectively, P<0.05). The muscle displaying quality, signal loss and general scores were lowest in group with 4 mm slice thickness (q-value were 4.896, 6.120, 6.327, 7.138, 3.863 and 4.043, P<0.05). The scores of muscle display quality, signal loss and general for b=400 s/mm2 were (9.0 ± 0.0), (14.0 ± 2.2) and (33.0 ± 2.2) score respectively, which were lower than those with b=800 s/mm2 [(7.0 ± 0.0), (6.2 ± 2.2), (21.8 ± 3.4) score] and b=1000 s/mm2 [(7.0 ± 0.0), (5.0 ± 0.0), (20.6 ± 2.2) score] (q-value were 3.873, 3.873, 6.650, 7.672, 7.101 and 5.917, P<0.05). The scores of muscle displaying quality, signal loss and general for b=600 s/mm2 were (8.2 ± 1.1), (13.0 ± 2.3) and (30.8 ± 3.8) score respectively, which were higher than those with b=800 s/mm2 and b=1000 s/mm2 (q-value were 3.873, 3.873, 5.797, 6.820, 5.326 and 5.917, P <0.05). There is no significant difference between b =600 s/mm2 and 400 s/mm2 (q-value were 2.582 and 0.852 and 1.775, P >0.05). Conclusion: Our preliminary findings indicate that it is feasible to perform DTI on human calf with 3 TMR. With body phased-array coil and surface coil of spine as receiving coil, the DTI sequence were optimized to acquire enough SNR with slice thickness of 5 mm and b-value of 400 s/mm2. (authors)