Evaluation of a direct motion estimation/correction method in respiratory-gated PET/MRI with motion-adjusted...

Respiratory motion compensation in PET/CT and PET/MRI is essential as motion is a source of image degradation (motion blur, attenuation artifacts). In previous work, we developed a direct method for joint image reconstruction/motion estimation (JRM) for attenuation-corrected (AC) respiratory-gated PET, which uses a single attenuation-map (μ-map). This approach was successfully implemented for respiratory-gated PET/CT, but since it relied on an accurate μ-map for motion estimation, the question of its applicability in PET/MRI is open. The purpose of this work is to investigate the feasibility of JRM in PET/MRI and to assess the robustness of the motion estimation when a degraded μ-map is used.

We performed a series of JRM reconstructions from simulated PET data using a range of simulated Dixon MRI sequence derived μ-maps with wrong attenuation values in the lungs, from -100% (no attenuation) to +100% (double attenuation), as well as truncated arms. We compared the estimated motions with the one obtained from JRM in ideal conditions (no noise, true μ-map as an input). We also applied JRM on 4 patient datasets of the chest, 3 of them containing hot lesions. Patient list-mode data were gated using a principal component analysis method. We compared SUV_max values of the JRM reconstructed activity images and non motion-corrected images. We also assessed the estimated motion fields by comparing the deformed JRM-reconstructed activity with individually non-AC reconstructed gates.

Experiments on simulated data showed that JRM-motion estimation is robust to μ-map degradation in the sense that it produces motion fields similar to the ones obtained when using the true μ-map, regardless of the attenuation errors in the lungs (< 0.5% mean absolute difference with the reference motion field). When using a μ-map with truncated arms, JRM estimates a motion field that stretches the μ-map in order to match the projection data. Results on patient datasets showed that using JRM improves the SUV_max values of hot lesions significantly and suppresses motion blur. When the estimated motion fields are applied to the reconstructed activity, the deformed images are geometrically similar to the non-AC individually reconstructed gates.

Motion estimation by JRM is robust to variation of the attenuation values in the lungs. JRM successfully compensates for motion when applied to PET/MRI clinical datasets. It provides a potential alternative to existing methods where the motion fields are pre-estimated from separate MRI measurements.