LASSNet: A Four Steps Deep Neural Network for Left Atrial Segmentation and Scar Quantification.

Accurate quantification of left atrium (LA) scar in patients with atrial fibrillation is essential to guide successful ablation strategies. Prior to LA scar quantification, a proper LA cavity segmentation is required to ensure exact location of scar. Both tasks can be extremely time-consuming and are subject to inter-observer disagreements when done manually. We developed and validated a deep neural network to automatically segment the LA cavity and the LA scar. The global architecture uses a multi-network sequential approach in two stages which segment the LA cavity and the LA Scar. Each stage has two steps: a region of interest Neural Network and a refined segmentation network. We analysed the performances of our network according to different parameters and applied data triaging. 200+ late gadolinium enhancement magnetic resonance images were provided by the LAScarQS 2022 Challenge. Finally, we compared our performances for scar quantification to the literature and demonstrated improved performances.

Optical forward-scattering for identification of bacteria within microcolonies.

The development of methods for the rapid identification of pathogenic bacteria is a major step towards accelerated clinical diagnosis of infectious diseases and efficient food and water safety control. Methods for identification of bacterial colonies on gelified nutrient broth have the potential to bring an attractive solution, combining simple optical instrumentation, no need for sample preparation or labelling, in a non-destructive process. Here, we studied the possibility of discriminating different bacterial species at a very early stage of growth (6 h of incubation at 37 °C), on thin layers of agar media (1 mm of Tryptic Soy Agar), using light forward-scattering and learning algorithms (Bayes Network, Continuous Naive Bayes, Sequential Minimal Optimisation). A first database of more than 1,000 scatterograms acquired on 7 gram-negative strains yielded a recognition rate of nearly 80%, after only 6 h of incubation. We investigated also the prospect of identifying different strains from a same species through forward scattering. We discriminated, thus, four strains of Escherichia coli with a recognition rate reaching 82%. Finally, we show the discrimination of two species of coagulase-negative Staphylococci (S. haemolyticus and S. cohnii), on a commercial selective pre-poured medium used in clinical diagnosis (ChromID MRSA, bioMérieux), without opening lids during the scatterogram acquisition. This shows the potential of this method--non-invasive, preventing cross-contaminations and requiring minimal dish handling--to provide early clinically-relevant information in the context of fully automated microbiology labs.