Purpose

We used CS-SENSE, a combination of compressed sense and parallel imaging, to accelerate a standard ankle MRI protocol with Dixon imaging and T2 mapping without losing image quality.

Methods and Materials

We scanned twenty healthy volunteers on a clinical 3.0T MRI scanner (Ingenia), equipped with a CS-SENSE software prototype extension (Philips Healthcare, Best, the Netherlands).

All CS-SENSE images were compared to standard SENSE images. The images were randomized, presented on two diagnostic monitors and evaluated by three blinded readers. To standardize the assessment, we used a five-point Likert scale consisting of 25 anatomical regions. Every answer on the Likert scale contributed one point in favor of the preferred type of scan. For the quantitative T2 mapping scans, we performed a pixel-wise comparison of T2 relaxometry measurements with accelerations of 1.5×, 1.7×, 1.8× and 1.9× faster than the normal acquisition time with SENSE.

Results

The readers found the quality of the Dixon images of SENSE and CS-SENSE comparable in 88.7% of the scorings as illustrated in figure 1. With CS-SENSE we were able to reduce the total acquisition time of the ankle Dixon scans by 45% from 13’38” to 7’37”.

The T2 relaxometry measurements on cartilage were highly dependent on the signal-to-noise ratios in the T2-mapping sequence. We found an acceleration of 1.8× faster than the SENSE scan to be the optimal acceleration factor based on the acquisition time and comparability with the SENSE scan. The spread in T2 values illustrated in the Bland-Altman plot (Figure 2a) was acceptable and the T2 values in talar and tibial cartilage using SENSE and CS-SENSE had a good correlation of ρ = 0.65 (figure 2b).

Conclusion

In conclusion, CS-SENSE could accelerate qualitative Dixon images and quantitative T2-mapping images by 45% with comparable image quality and T2 relaxometry measurements of the ankle.