2668

View Abstract / View Presentation

Evaluation of a Deep Learning reconstruction framework for three-dimensional cardiac imaging

Gaspar Delso¹, Marc Lebel², Suryanarayanan Kaushik², Graeme McKinnon², Paz Garre³, Pere Pujol³, Daniel Lorenzatti³, José T Ortiz³, Susanna Prat³, Adelina Doltra³, Rosario J Perea³, Teresa M Caralt³, Lluis Mont³, and Marta Sitges³
¹GE Healthcare, Barcelona, Spain, ²GE Healthcare, Waukesha, WI, United States, ³Hospital Clínic de Barcelona, Barcelona, Spain

The Deep Learning framework was found to provide equivalent diagnostic information content as state-of-the-art 3D Cartesian reconstruction, with consistently superior image quality and processing time compatible with clinical routine.

Figure 1.- Long axis views of 3D MDE series, reconstructed with a standard 3D Cartesian method (left) and the proposed Deep Learning framework (right).

Figure 4.- Top: Logarithmic joint histograms of the voxel-wise relative standard deviation, in the reference Cartesian and DL reconstructions shown in figure 1. Notice how most voxels are located below the identity line, indicating SNR improvement. Bottom: Line profile illustrating the preservation of structure edges with the regularized reconstruction.