Published March 6, 2026 | Version 1.0
Conference proceeding Open

A Dataset for Benchmarking Machine Learning Models for Autonomous Deep Vein Thrombosis Detection Based on Compression Ultrasound Videos

  • 1. ROR icon Democritus University of Thrace
  • 2. ROR icon Athena Research and Innovation Center In Information Communication & Knowledge Technologies

Description

Deep vein thrombosis (DVT) is a major vascular condition associated with substantial morbidity, mortality and healthcare burden. Compression ultrasonography, performed and interpreted by medical experts, is the primary diagnostic method. Advances in machine learning (ML) and deep learning (DL) offer promising
opportunities to support automated and real-time DVT assessment by non-experts. However, existing approaches rely on pixel-wise vessel annotations, which are costly to generate, posing difficulties in developing models that generalize across devices and acquisition protocols. To address these limitations, we introduce
the ThrombUS+ Dataset #1. The Dataset is based on 2919 segmentation-free compression ultrasound videos from 742 patients suspected of DVT, acquired from a multicenter cohort study across 5 European hospitals.

The work has been officilay published: In Proceedings of the 19th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 4: HEALTHINF, ISBN 978-989-758-802-0, ISSN 2184-4305, pages 853-862

Full Citation:

Didaskalou, S.; Portokallidis, N.; Tzatzimaki, K.; Liapi, M.; Moustakidis, N.; Moustakidis, T.; Balciuniene, N.; Macas, A.; Kijauskas, R.; Aladaitis, A.; Sotiriadou, A.; Sarafis, F.; Kynigopoulos, G.; Potoupnis, M.; Grandone, E.; Mastrangelo, G.; Maresca, S.; Gautier, M.; Chaiba, D.; Boussaha, H.; Goulvent, S.; Stylianou, C.; Oglou, E. N.; Chouchos, K.; Deftereos, S.; Papatheodorou, K.; Drougka, I.; Anagnostopoulou, P.; Yu, H. Q.; Kaldoudi, E. and ThrombUS+, (2026). A Dataset for Benchmarking Machine Learning Models for Autonomous Deep Vein Thrombosis Detection Based on Compression Ultrasound Videos.  In Proceedings of the 19th International Joint Conference on Biomedical Engineering Systems and Technologies - https://doi.org/10.5220/0014741500004070 

Link to training dataset: https://zenodo.org/records/17659415

Link to testing dataset: https://zenodo.org/records/17664207

Files

ThrombUS_Proceedings_HEALTHINF_2026_DVT_US.pdf

Files (1.3 MB)

Name Size Download all
md5:7cdc400089d07f0cc92e055cfe642ca4
1.3 MB Preview Download

Additional details

Funding

European Commission
ThrombUS - ThrombUS+: Wearable Continuous Point-of-Care Monitoring, Risk Estimation and Prevention for Deep Vein Thrombosis 101137227