FedStenoNet X-ray Coronary Angiography Dataset for Stenosis Detection

This dataset is released to support research on the automatic detection of coronary stenosis in X-ray coronary angiography (XCA). It contains 1565 XCA images acquired at Ospedali Riuniti (Ancona, Italy) using Azurion Clarity IQ (Philips) and Allura Xper FD10 (Philips) scanners, from 244 patients who underwent XCA procedures during 2020. All patients included in the dataset were diagnosed with significant coronary stenosis, as reported in their examination records. According to the European Society of Cardiology Guidelines, a stenosis was considered significant if presenting an area reduction ≥70% measured with quantitative coronary angiography (QCA). When necessary, diagnosis was confirmed by intravascular ultrasound (IVUS), where a stenosis was considered significant if the minimum lumen area was ≤4 mm². Images were selected considering high-contrast dye visibility, varied viewpoints, and the diastolic phase of the cardiac cycle. Three expert clinicians supervised the acquisition and provided manual annotations of stenotic regions along the left coronary artery. Data collection complied with the Helsinki Declaration, was approved by the Local Ethics Committee (CET 59/2024), and informed consent was obtained from all patients. This dataset corresponds to Dataset A described and used in the article:Di Cosmo, M., Migliorelli, G., Villani, F.P., Francioni, M., Muçaj, A., Frontoni, E., Moccia, S., Fiorentino, M.C. "FedStenoNet: Tackling domain shift in x-ray coronary angiography through a personalized federated detection framework", Computers in Biology and Medicine (2025).