{
  "abstract": "Aim To assess the diagnostic accuracy of a fully automated deep learning (DL) model in segmenting the coronary arteries and detecting the calcifications using unenhanced and ungated CT thorax scans in patients with pulmonary hypertension.Methods A two-stage 3D segmentation pipeline was developed using 69 CT scans to identify coronary artery calcifications in the right coronary artery (RCA), left anterior descending artery (LAD), and left circumflex artery (LCX). The model was trained with anatomically refined labels and region-based optimisation to improve structural coherence. Model outputs were visually assessed in a separate cohort of 100 scans by two independent, experienced observers. Segmentation and detection performance was evaluated against manually annotated ground truth using a binary analysis in 473 scans. Volumetric measurements of calcifications were compared with baseline severity gradings derived from an experienced radiologists’ reports. Interobserver variability between DL and manual ground truth was calculated using Cohen’s kappa and p-value.Results The model was able to segment and detect the coronary arteries and calcifications with high performance. The majority of the scans were rated as excellent in the visual assessment with kappa values ranging from 0.68 to 0.81 for coronary artery segmentation and from 0.79 to 0.85 for calcification segmentation. The AI model accurately detected the presence of calcifications in the RCA (κ = 0.82, p < 0.001), LAD (κ = 0.93, p < 0.001), and LCX (κ = 0.82, p < 0.001). The model detected coronary calcifications with a sensitivity of 95%, specificity of 98%, positive predictive value of 99%, and negative predictive value of 88%. Calcification volume increased progressively with disease severity.Abstract P281 Figure 1Graphical abstract including three stages: segmentation model development stage, visual assessment stage, and calcification detection and diagnostic accuracy stageConclusion The DL model trained on unenhanced and ungated CT scans accurately segmented and detected the presence of calcifications as well as the severity with high accuracy, highlighting its potential for future clinical practice in the detection of incidental coronary artery disease.",
  "authors": [
    {
      "affiliations": [
        "School of Medicine and Population Health, The University of Sheffield, Sheffield, UK",
        "Radiological Sciences Department, College of Applied Medical Science, King Saud bin Abdulaziz University for Health Science, Riyadh, Saudi Arabia",
        "King Abdullah International Medical Research Center (KAIMRC), Riyadh, Saudi Arabia"
      ],
      "name": "TN Alnasser"
    },
    {
      "affiliations": [
        "School of Medicine and Population Health, The University of Sheffield, Sheffield, UK",
        "Insigneo Institute, Faculty of Engineering, The University of Sheffield, Sheffield, UK"
      ],
      "name": "A Hokmabadi"
    },
    {
      "affiliations": [
        "School of Medicine and Population Health, The University of Sheffield, Sheffield, UK",
        "3D Imaging Lab, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK"
      ],
      "name": "M Sharkey"
    },
    {
      "affiliations": [
        "School of Medicine and Population Health, The University of Sheffield, Sheffield, UK",
        "Department of Clinical Radiology, Sheffield Teaching Hospitals, Sheffield, UK"
      ],
      "name": "K Dwivedi"
    },
    {
      "affiliations": [
        "School of Medicine and Population Health, The University of Sheffield, Sheffield, UK",
        "Department of Clinical Radiology, Sheffield Teaching Hospitals, Sheffield, UK"
      ],
      "name": "M Salehi"
    },
    {
      "affiliations": [
        "School of Medicine and Population Health, The University of Sheffield, Sheffield, UK",
        "Insigneo Institute, Faculty of Engineering, The University of Sheffield, Sheffield, UK",
        "Sheffield Pulmonary Vascular Disease Unit, Sheffield Teaching Hospitals NHS Trust, Sheffield, UK",
        "National Institute for Health and Care Research, Sheffield Biomedical Research Centre, Sheffield, UK"
      ],
      "name": "DG Kiely"
    },
    {
      "affiliations": [
        "School of Medicine and Population Health, The University of Sheffield, Sheffield, UK",
        "Insigneo Institute, Faculty of Engineering, The University of Sheffield, Sheffield, UK",
        "Department of Clinical Radiology, Sheffield Teaching Hospitals, Sheffield, UK",
        "National Institute for Health and Care Research, Sheffield Biomedical Research Centre, Sheffield, UK"
      ],
      "name": "S Alabed"
    },
    {
      "affiliations": [
        "School of Medicine and Population Health, The University of Sheffield, Sheffield, UK",
        "Insigneo Institute, Faculty of Engineering, The University of Sheffield, Sheffield, UK",
        "Department of Clinical Radiology, Sheffield Teaching Hospitals, Sheffield, UK",
        "National Institute for Health and Care Research, Sheffield Biomedical Research Centre, Sheffield, UK"
      ],
      "name": "AJ Swift"
    }
  ],
  "title": "P281 Diagnostic accuracy of a deep learning model for coronary calcification detection on unenhanced CT in pulmonary hypertension patients",
  "uid": "1e6bbacf-96ec-595a-aff1-d89d2c6f0046"
}