Incorporation of image data from a previous examination in 3D serial MR imaging

We aimed to demonstrate that follow-up scans in longitudinal examinations can be significantly accelerated by using images from previous scans as priors for constrained reconstruction.

In this work, we propose a method for incorporating a prior image to improve the reconstruction of a new acquisition with considerable k-space undersampling, which contains a two-level registration scheme with non-parametric transformation, an adaptive synthesis procedure, and a constrained reconstruction with weighted total variation constraint. The performance of the method is evaluated using simulations, as well as results from volunteer and patient examinations.

In vivo experiments with both volunteers and patients show that incorporating a prior image into the constrained reconstruction produces many fewer reconstruction errors compared to the conventional reconstruction using only the highly undersampled k-space data.

The redundant information in the prior image can be efficiently adopted to improve the reconstruction quality of the new acquisition. When maintaining the image quality, higher acceleration can be achieved with the incorporation of the prior image.