Motion-Compensated MR Image Reconstruction

Magnetic resonance imaging (MRI) of the lung and abdomen is deteriorated by respiratory motion. This is because the acquisition times are usually much longer than the time of one respiratory cycle. This leads to motion blurring of reconstructed MR images. Standard motion mitigation strategies, such as breath-hold acquisition or gating, have limitations in either temporal or spatial resolution, SNR, CNR and artifact level or demand inappropriately long acquisition times.

We are developing new MR image reconstruction algorithms to overcome these limitations. They enable thoracic and abdominal respiratory motion-compensated 4D (3D + time) image reconstruction of radial MR data with very high sparsity acquired during free breathing. Since image reconstruction and motion estimation are carried out in an alternating manner, our algorithms exploit both the capabilities of a specifically-tailored iterative sparse data reconstruction employing parallel imaging and the advantages of motion compensation, i.e. each motion phase is reconstructed from 100% of the measured data. Thus, very short acquisition times in the order of 60 s or less become possible while achieving high temporal and spatial resolution and suppressing artifacts and noise effectively.

In the clinical environment, respiratory time-resolved 4D techniques might be employed for the analysis of lung function or for decision making and planning of surgical procedure for patients with lung tumors. In the context of radiation oncology, 4D MRI might be used for patient-specific definition of safety margins or even for motion-adapted radiotherapy planning.

back to main page