QUEST-EF is a research website for testing a deep learning model that estimates both left and right ventricular ejection fraction from a single apical four-chamber echocardiography video. It was developed to make automated assessment of heart function faster, simpler, and more accessible in routine echocardiography workflows.

The model is vendor-independent, segmentation-free, and designed for research use, not clinical decision-making. It is intended to help users explore how a single ultrasound view can support rapid screening for LV and RV systolic dysfunction.

References

• Szijártó, Á., Merkely, B., Kovács, A., & Tokodi, M. (2025). Deep learning-enabled echocardiographic assessment of biventricular ejection fractions: the dual-task QUEST-EF model. European Heart Journal-Cardiovascular Imaging, 26(8), 1402-1405.
• Szijártó, Á., Magyar, B., Szeier, T. Á., Tolvaj, M., Fábián, A., Lakatos, B. K., ... & Tokodi, M. (2024, October). Masked autoencoders for medical ultrasound videos using ROI-aware masking. In International Workshop on Advances in Simplifying Medical Ultrasound (pp. 167-176). Cham: Springer Nature Switzerland.