Abstracts

RETURN TO ABSTRACT LISTING


E1650. Role of Edge Device as Screening Point-of-Care Solutions Using Imaging Diagnostics for Population Screening
Authors
  1. Amit Kharat; DeepTek Imaging Private Limited
  2. Priyam Choudhury; DeepTek Imaging Private Limited
  3. Viraj Kulkarni; DeepTek Imaging Private Limited
Objective:
Internet accessibility differs across regions in countries creating the so called 'digital divide', between those who have access to digital information and communications technologies (ICT) and those who don't. Edge devices are revolutionizing diagnostics for low resource settings. It can reside within or adjacent to imaging tools, key advantages being they are faster, non-dependent on internet bandwidth, can perform a pre-screening before the radiologist can review the film and make a final diagnosis. Hence it can be peripherally deployed where experts are lacking. Edge devices are also called nano device which can operate in settings without an air conditioner or harsher weather conditions. These are either central processing units (CPUs) or Graphics processing units (GPUs) with advanced processing deep and machine learning (artificial intelligence) algorithms. It processes images and assists in classification and triage solutions, to flag studies as either normal or abnormal, without a need to push the images into cloud system, improving workflow and can be deployed as a screening point-of-care solution.

Materials and Methods:
The screening device enabled with deep learning tools should ideally be portable and ready for quick deployment in the suburbs and rural areas. Key features for edge devices to have an impact on diagnostics include operability of the diagnostic devices on a single-phase connection or on battery, and ability to deliver good quality images in Digital Imaging and Communications (DICOM) format. Any device that can generate DICOM data can potentially be used to assist in augmenting healthcare solutions for experts. In such an environment, the edge solutions process data, in real-time with no dependency on network breakdown or congestion, making healthcare accessible in every kind of setting.

Results:
Edge devices can screen the following conditions using chest X-rays: Tuberculosis (TB): active or old (latent TB); Pneumonia related to COVID-19, Severe Acute Respiratory Illness Syndrome (SARS), and H1N1 influenza or swine flu; Nodules for lung cancer; Pneumothorax. It can screen micro calcification and masses from mammography images. Edge solutions, when deployed on CT, can be used for the following conditions for CT Chest like Pneumonia detection (COVID-19, community acquired pneumonias etc.); Tuberculosis detection; Interstitial lung disease detection (ILD); Lung mass and lung nodules and for CT Brain: Stroke detection and excluding bleed and detecting acute infarcts. For MRI Brain, the following solutions can benefit from edge tools: Stroke detection (acute infarct or bleed detection).

Conclusion:
If the entire triage solution is available as an edge solution, it can be added in the workflow to raise alerts and enable smart notifications reducing significant workload of radiologists. It helps them to focus on important studies and emergencies first, before moving on to normal studies, therefore becoming a game changer in population health.