Improved MR to CT synthesis for PET/MR attenuation correction using Imitation Learning
This work addresses the specific challenge of optimizing pseudo-CT synthesis for PET/MRI attenuation correction in medical imaging, representing an incremental improvement by refining the loss function to better align with clinical goals.
The paper tackled the problem of synthesizing CT images from MRI for PET/MR attenuation correction by proposing a multi-hypothesis deep learning framework that minimizes both pixel-wise CT error and a novel metric-loss for PET residuals, resulting in improved PET reconstruction accuracy (115 a.u. vs. baseline 140 a.u.) despite slightly worse CT synthesis (69.68 HU vs. baseline 66.25 HU).
The ability to synthesise Computed Tomography images - commonly known as pseudo CT, or pCT - from MRI input data is commonly assessed using an intensity-wise similarity, such as an L2-norm between the ground truth CT and the pCT. However, given that the ultimate purpose is often to use the pCT as an attenuation map ($μ$-map) in Positron Emission Tomography Magnetic Resonance Imaging (PET/MRI), minimising the error between pCT and CT is not necessarily optimal. The main objective should be to predict a pCT that, when used as $μ$-map, reconstructs a pseudo PET (pPET) which is as close as possible to the gold standard PET. To this end, we propose a novel multi-hypothesis deep learning framework that generates pCTs by minimising a combination of the pixel-wise error between pCT and CT and a proposed metric-loss that itself is represented by a convolutional neural network (CNN) and aims to minimise subsequent PET residuals. The model is trained on a database of 400 paired MR/CT/PET image slices. Quantitative results show that the network generates pCTs that seem less accurate when evaluating the Mean Absolute Error on the pCT (69.68HU) compared to a baseline CNN (66.25HU), but lead to significant improvement in the PET reconstruction - 115a.u. compared to baseline 140a.u.