Researchers in the Department of Radiology and Department of Cardiology at Fuwai Hospital, a leading hospital in Beijing specializing in the treatment of complex cardiovascular diseases, conducted a study to evaluate the performance of deep learning-based CT-FFR in detecting hemodynamic changes of stenosis [1]. Keya Medical provided its non-invasive CT-FFR software, DeepVessel FFR, for use in the study. The performance results were published in the paper “Additional Value of Deep Learning Computed Tomographic Angiography-based Fractional Flow Reserve in Detecting Coronary Stenosis and Predicting Outcomes,” available in the peer-reviewed radiology journal, Acta Radiologica*.

In this post, we first describe the methods used to evaluate the performance of CT-FFR in detecting hemodynamic changes of stenosis. We then summarize the findings that suggest CT-FFR is superior to coronary computed tomographic angiography (CCTA) in differentiating functional myocardial ischemia and has the potential to differentiate prognoses of patients with coronary artery disease (CAD).

CT-FFR Supplements CCTA for Diagnosing Coronary Ischemia

Coronary computed tomographic angiography is recognized as a preferred non-invasive technique for screening patients with CAD. The digital images created during the CCTA procedure can be further processed to create a simulation of Fractional Flow Reserve (FFR). Recent studies have demonstrated that CCTA alone may not be effective as a standalone diagnostic tool due to its low specificity and positive predictive value (PPV) [2, 3]. Deep learning-based fractional flow reserve derived from CCTA (CT-FFR) could deliver additional value in detecting coronary stenosis and predicting patient outcomes.

This retrospective study conducted by researchers at Fuwai Hospital included 73 patients suspected of CAD who received CCTA followed by invasive FFR within 90 days. Thirty-nine patients who received drug therapy instead of revascularization were followed for up to 31 months for major adverse cardiovascular events (MACE), unstable angina, and rehospitalization. Researchers compared the diagnostic performance of the deep learning-based CT-FFR technique to conventional CCTA for predicting myocardial ischemia.

Study results demonstrate that the performance of CT-FFR exceeded that of CCTA at both the patient and vessel levels in terms of accuracy, sensitivity, and specificity. The CT-FFR approach also offered improvements in specificity and PPV without losing sensitivity and negative predictive value (NPV).

Findings: CT-FFR Demonstrates Superiority to CCTA

Deep learning-based CT-FFR could be an effective non-invasive tool for imaging myocardial ischemia in patients with CAD. This retrospective study revealed two important findings:

  • The diagnostic performance of CT-FFR was superior to conventional CCTA for determining myocardial ischemia in patients with CAD
  • CT-FFR predicted clinical outcomes of patients, including MACE and rehospitalization.
These study results demonstrate the value of CT-FFR in detecting coronary stenosis and predicting outcomes. Future studies with a larger sample size will be needed to verify the value that CT-FFR can add to cardiovascular imaging.

Collaborate with us

We are honored to collaborate with leading healthcare institutions like Fuwai Hospital to advance the detection and diagnosis of cardiovascular disease. If you are interested in collaborating with Keya Medical on future projects, please contact us.


[1] Li Y, Qiu H, Hou Z, et al. Additional value of deep learning computed tomographic angiography-based fractional flow reserve in detecting coronary stenosis and predicting outcomes. Acta Radiologica. January 2021. doi:10.1177/0284185120983977

[2] Raff GL, Gallagher MJ, O’Neill WW, et al. Diagnostic accuracy of noninvasive coronary angiography using 64- slice spiral computed tomography. J Am Coll Cardiol 2005;46:552–557.

[3] Miller JM, Rochitte CE, Dewey M, et al. Diagnostic performance of coronary angiography by 64-row CT. N Engl J Med 2008;359:2324–2336.

*Acta Radiologica is a peer-reviewed radiology journal published in association with the Nordic Society of Medical Radiology. The journal covers all aspects of radiology, from clinical radiology to experimental work.