Deep Evolutionary Networks with Expedited Genetic Algorithms for Medical Image Denoising.

Deep convolutional neural networks offer state-of-the-art performance for medical image analysis. However, their architectures are manually designed for particular problems. On the one hand, a manual designing process requires many trials to tune a large number of hyperparameters and is thus quite a time-consuming task. On the other hand, the fittest hyperparameters that can adapt to source data properties (e.g., sparsity, noisy features) are not able to be quickly identified for target data properties. For instance, the realistic noise in medical images is usually mixed and complicated, and sometimes unknown, leading to challenges in applying existing methods directly and creating effective denoising neural networks easily. In this paper, we present a Genetic Algorithm (GA)-based network evolution approach to search for the fittest genes to optimize network structures automatically. We expedite the evolutionary process through an experience-based greedy exploration strategy and transfer learning. Our evolutionary algorithm procedure has flexibility, which allows taking advantage of current state-of-the-art modules (e.g., residual blocks) to search for promising neural networks. We evaluate our framework on a classic medical image analysis task: denoising. The experimental results on computed tomography perfusion (CTP) image denoising demonstrate the capability of the method to select the fittest genes for building high-performance networks, named EvoNets. Our results outperform state-of-the-art methods consistently at various noise levels.

Local release of dexamethasone from macroporous scaffolds accelerates islet transplant engraftment by promotion of anti-inflammatory M2 macrophages.

Transplant-associated inflammatory responses generate an unfavorable microenvironment for tissue engraftment, particularly for cells susceptible to inflammatory stress, such as pancreatic islets. The localized delivery of anti-inflammatory agents, such as glucocorticoids, offers a promising approach to minimize the detrimental side effects associated with systemic delivery; however, the dosage must be carefully tailored to avoid deleterious responses, such as poor engraftment. Herein, we employed a polydimethylsiloxane (PDMS)-based three-dimensional scaffold platform for the local and controlled delivery of dexamethasone (Dex). Incorporation of 0.1% or 0.25% Dex within the scaffold was found to significantly accelerate islet engraftment in a diabetic mouse model, resulting in improved control of blood glucose levels during the early transplant period. Investigation into the mechanism of this impact found that local Dex delivery promotes macrophage polarization towards an anti-inflammatory (M2) phenotype and suppresses inflammatory pathways during the first week post-implantation. Alternatively, higher Dex loadings (0.5% and 1%) significantly delayed islet engraftment and function by impairing host cell migration into the implanted graft. Our results demonstrate the dose-dependent impact of local glucocorticoid delivery on the modulation of inflammatory responses at the implant site in vivo. Outcomes highlight the potential of this platform for generating favorable host responses that improve overall cellular transplant outcomes.