Cross noise level PET denoising with continuous adversarial domain generalization

Objective.Performing positron emission tomography (PET) denoising within the image space proves effective in reducing the variance in PET images. In recent years, deep learning has demonstrated superior denoising performance, but models trained on a specific noise level typically fail to generalize well on different noise levels, due to inherent distribution shifts between inputs. The distribution shift usually results in bias in the denoised images. Our goal is to tackle such a problem using a domain generalization technique.Approach.We propose to utilize the domain generalization technique with a novel feature space continuous discriminator (CD) for adversarial training, using the fraction of events as a continuous domain label. The core idea is to enforce the extraction of noise-level invariant features. Thus minimizing the distribution divergence of latent feature representation for different continuous noise levels, and making the model general for arbitrary noise levels. We created three sets of 10%, 13%-22% (uniformly randomly selected), or 25% fractions of events from 9718F-MK6240 tau PET studies of 60 subjects. For each set, we generated 20 noise realizations. Training, validation, and testing were implemented using 1400, 120, and 420 pairs of 3D image volumes from the same or different sets. We used 3D UNet as the baseline and implemented CD to the continuous noise level training data of 13%-22% set.Main results.The proposed CD improves the denoising performance of our model trained in a 13%-22% fraction set for testing in both 10% and 25% fraction sets, measured by bias and standard deviation using full-count images as references. In addition, our CD method can improve the SSIM and PSNR consistently for Alzheimer-related regions and the whole brain.Significance.To our knowledge, this is the first attempt to alleviate the performance degradation in cross-noise level denoising from the perspective of domain generalization. Our study is also a pioneer work of continuous domain generalization to utilize continuously changing source domains.

Accelerated positive contrast MRI of interventional devices using parallel compressed sensing imaging

Susceptibility-based magnetic resonance imaging (MRI) method can image small MR-compatible devices with positive contrast. However, the relatively long data acquisition time required by the method hinders its practical applications. This study presents a parallel compressive sensing technique with a modified fast spin echo to accelerate data acquisition for the susceptibility-based positive contrast MRI. The method integrates the generalized autocalibrating partially parallel acquisitions and the compressive sensing techniques in the reconstruction algorithm. MR imaging data acquired from several phantoms containing interventional devices such as biopsy needles, stent, and brachytherapy seeds, used for validating the proposed technique. The results show that it can speed up data acquisition by a factor of about five while preserving the quality of the positive contrast images.

Parallel compressive sensing in a hybrid space: Application in interventional MRI

Recently, susceptibility based positive contrast MRI technique emerged as an effective method of visualizing the small MR compatible devices, such as brachytherapy seeds. One of the challenges associated with this method is the long scan time. In this work, we present an accelerated susceptibility based positive contrast MR imaging method, in which the susceptibility map can be generated from an under-sampled data. We use a combination of parallel imaging (GRAPPA) and compressive sensing (CS) technique in a hybrid k-space. The results in brachytherapy seeds imaging show that 3-D high-quality images can be achieved at an acceleration factor up to 4, which can decrease the scan time from 4 minutes to 1.4 minutes.