Deep Learning for Classification and Localization of COVID-19 Markers in Point-of-Care Lung Ultrasound

Abstract

Deep learning (DL) has proved successful in medical imaging and, in the wake of the recent COVID-19 pandemic, some works have started to investigate DLbased solutions for the assisted diagnosis of lung diseases. While existing works focus on CT scans, this paper studies the application of DL techniques for the analysis of lung ultrasonography (LUS) images. Specifically, we present a novel fully-annotated dataset of LUS images collected from several Italian hospitals, with labels indicating the degree of disease severity at a frame-level, videolevel, and pixel-level (segmentation masks). Leveraging these data, we introduce several deep models that address relevant tasks for the automatic analysis of LUS images. In particular, we present a novel deep network, derived from Spatial Transformer Networks, which simultaneously predicts the disease severity score associated to a input frame and provides localization of pathological artefacts in a weakly-supervised way. Furthermore, we introduce a new method based on uninorms for effective frame score aggregation at a video-level. Finally, we benchmark state of the art deep models for estimating pixel-level segmentations of COVID-19 imaging biomarkers. Experiments on the proposed dataset demonstrate satisfactory results on all the considered tasks, paving the way to future research on DL for the assisted diagnosis of COVID-19 from LUS data.

A Learning-Based Microultrasound System for the Detection of Inflammation of the Gastrointestinal Tract

“…Computer-aided US diagnosis methods for breast cancer imaging have been researched, incorporating features relating to shape, margin, orientation, echo patterns and acoustic shadowing [10]. A more recent development has been reported to use deep learning techniques for extracting nonlinear features from US images at frame-level and videolevel using convex and linear probes with central frequency below 5 MHz [11]. Although this study demonstrated the feasibility of using deep learning to enable the characterisation of the state of high permeability and advanced disease, the issues of how to incorporate deep learning techniques into the US imaging chain from data acquisition to post image processing was ignored.…”

Deep learning and its role in COVID-19 medical imaging

“…Lung ultrasound further carries the advantage of not imposing any radiation and eliminating the need for transferring the patient, since the examination can be performed bedside[ 21 ]. Roy et al[ 22 ] developed and tested several deep learning models for detection of COVID-19 associated patterns on lung ultrasound scans. Their models concurrently output the severity of lung disease on a 4-point scale and segmented the pathological area on each lung ultrasound scan.…”

“…Their models concurrently output the severity of lung disease on a 4-point scale and segmented the pathological area on each lung ultrasound scan. Roy et al argue that their model could be beneficial in a point-of-care assessment of disease severity and also be used for triaging patients with COVID-19[ 22 ].…”

Ultrasound Elastography for Lung Disease Assessment

“…Clinical interpretation of 2-D lung images is subjective, and at times, inaccurate, and irreproducible. Roy et al [68] developed a deep learning framework for classification and localization of COVID-19 markers in LUS. This framework derived from a spatial transformer network was able to predict the disease severity score associated with the input frame and provide localization of pathological artifacts in a weakly supervised manner.…”