International Journal of Scientific Research and Engineering Development

Cardiomegaly refers to enlargement of the cardiac silhouette on chest radiography and is commonly evaluated using the cardiothoracic ratio (CTR), defined as the ratio of maximal cardiac width to maximal internal thoracic width. On an appropriately acquired posteroanterior radiograph, a CTR greater than 0.50 may indicate cardiomegaly; however, its interpretation is influenced by projection type, patient rotation, inspiratory effort, positioning, and incomplete pulmonary expansion [1–3]. These technical and anatomical factors may produce apparent cardiac enlargement, thereby limiting the reliability of automated image-based screening. Machine-learning models trained on public chest-radiograph datasets may also be affected by uncertainty associated with report-derived labels. In small-sample studies, label noise, sampling bias, and heterogeneity in image acquisition may disproportionately influence estimated model performance and generalisability [4,5]. This study proposes an interpretable screening workflow that identifies relevant anatomical structures and derives cardiothoracic measurements rather than relying exclusively on raw-image classification. Because radiographic assessment of cardiomegaly is fundamentally based on anatomical proportions, this feature-based approach was designed to support transparent preliminary screening in settings with limited positive training data [1,2].

Rungphob Lertvilaivithaya,"Segmentation-Derived Cardiothoracic Features for A Pilot Cardiomegaly Screening Workbench on Chest Radiographs" International Journal of Scientific Research and Engineering Development, V9(3): Page(2601-2639) May-June 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.