Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Feb 26.
doi: 10.1007/s10278-025-01444-3. Online ahead of print.

Explained Deep Learning Framework for COVID-19 Detection in Volumetric CT Images Aligned with the British Society of Thoracic Imaging Reporting Guidance: A Pilot Study

Shereen Fouad  1 , Muhammad Usman  2 , Ra'eesa Kabir  2 , Arvind Rajasekaran  3 , John Morlese  3 , Pankaj Nagori  3 , Bahadar Bhatia  3
Affiliations

Explained Deep Learning Framework for COVID-19 Detection in Volumetric CT Images Aligned with the British Society of Thoracic Imaging Reporting Guidance: A Pilot Study

Shereen Fouad et al. J Imaging Inform Med. .

Abstract

In March 2020, the British Society of Thoracic Imaging (BSTI) introduced a reporting guidance for COVID-19 detection to streamline standardised reporting and enhance agreement between radiologists. However, most current DL methods do not conform to this guidance. This study introduces a multi-class deep learning (DL) model to identify BSTI COVID-19 categories within CT volumes, classified as 'Classic', 'Probable', 'Indeterminate', or 'Non-COVID'. A total of 56 CT pseudoanonymised images were collected from patients with suspected COVID-19 and annotated by an experienced chest subspecialty radiologist following the BSTI guidance. We evaluated the performance of multiple DL-based models, including three-dimensional (3D) ResNet architectures, pre-trained on the Kinetics-700 video dataset. For better interpretability of the results, our approach incorporates a post-hoc visual explainability feature to highlight the areas of the image most indicative of the COVID-19 category. Our four-class classification DL framework achieves an overall accuracy of 75%. However, the model struggled to detect the 'Indeterminate' COVID-19 group, whose removal significantly improved the model's accuracy to 90%. The proposed explainable multi-classification DL model yields accurate detection of 'Classic', 'Probable', and 'Non-COVID' categories with poor detection ability for 'Indeterminate' COVID-19 cases. These findings are consistent with clinical studies that aimed at validating the BSTI reporting manually amongst consultant radiologists.

Keywords: British Society of Thoracic Imaging; COVID-19; Deep learning; Explainable AI; Medical image analysis; Multi-class classification.

PubMed Disclaimer

Conflict of interest statement

Declarations. Ethics Approval: In this study, the dataset was collected in 2020 at the Sandwell and West Birmingham National Health System Trust (SWBNHST) hospitals in the UK. This study was performed in line with the principles of the Declaration of Helsinki. The study was reviewed and approved by the ethics committee at Aston University (No. EPS21006 and 234700-51). All procedures were conducted in compliance with relevant guidelines and regulations. Consent to Participate: No recruitment of human participants took place in this project. This project consists of processing of data already being routinely collected and archived. Data subjects are adult non-vulnerable patients at Sandwell & West Birmingham NHS Trust. It is fully anonymised so that individuals cannot be identified by anyone without access to the identifiable data in the hospital systems. The format of hospital data is DICOM imaging data and HL7 electronic patient records storage on hospital information systems. Access to anonymised data via the Aston University server was protected and restricted to authorised users. Consent for Publication: No recruitment of human participants took place in this project. The participating site is Sandwell & West Birmingham NHS Trust and this artificial intelligence (AI) research project will use public interest task to manage the data for improving clinical management of patients. The legal basis for processing is GDPR Article 6(1)(e) and Article 9(2)(j). Conflict of Interest: The authors declare no competing interests.

Similar articles

See all similar articles

References

    1. Ahmad, M., Bajwa, U.I., Mehmood, Y., Anwar, M.W.: Lightweight resgru: a deep learning-based prediction of sars-cov-2 (covid-19) and its severity classification using multimodal chest radiography images. Neural Computing and Applications 35(13), 9637–9655 (2023)
    1. Atim, P., Fouad, S., Yu, S.T., Fratini, A., Rajasekaran, A., Nagori, P., Morlese, J., Bhatia, B.: Explainable deep learning framework for ground glass opacity (ggo) segmentation from chest ct scans. In: Medical Imaging and Computer-Aided Diagnosis. Springer (2024)
    1. Barzas, K., Fouad, S., Jasa, G., Landini, G.: An explainable deep learning framework for mandibular canal segmentation from cone beam computed tomography volumes. In: The 12th International Conference on Computational Advances in Bio and Medical Sciences. Springer (2023)
    1. Borakati, A., Perera, A., Johnson, J., Sood, T.: Diagnostic accuracy of x-ray versus ct in covid-19: a propensity-matched database study. BMJ open 10(11), e042946 (2020) - DOI - PubMed
    1. Cardoso, M.J., Li, W., Brown, R., Ma, N., Kerfoot, E., Wang, Y., Murrey, B., Myronenko, A., Zhao, C., Yang, D., et al.: Monai: An open-source framework for deep learning in healthcare. arXiv preprint arXiv:2211.02701 (2022)

Grants and funding