Validation of the MedHub AutocathFFR Software Device for Non-Invasive Fractional Flow Reserve (FFR)

• MedHub AutocathFFR (Device), also known as AutocathFFR: MedHub AutocathFFR is a software device for the analysis of previously acquired angiography data in Digital Imaging and Communications in Medicine (DICOM) format. The AutocathFFR value is computed from standard 2D angiography images, taken from two different views, retrieved from the C-arm angiography system. The FFR value is calculated using a machine-learning and computer vision algorithm. The result is intended to support the functional evaluation of coronary artery disease.

Primary outcomes

• Per-Vessel Sensitivity of AutocathFFR vs. Invasive FFR [Day 1 (the day of the index invasive coronary angiography procedure)]: Sensitivity of the dichotomously scored AutocathFFR per vessel compared to the gold standard, invasive FFR measurements, where an FFR ≤ 0.80 is scored "positive" and an FFR > 0.80 is considered "negative". Performance goal: lower bound of one-sided 97.5% confidence interval must exceed 80%.
• Per-Vessel Specificity of AutocathFFR vs. Invasive FFR [Day 1 (the day of the index invasive coronary angiography procedure)]: Specificity of the dichotomously scored AutocathFFR per vessel compared to the gold standard, invasive FFR measurements, where an FFR ≤ 0.80 is scored "positive" and an FFR > 0.80 is considered "negative". Performance goal: lower bound of one-sided 97.5% confidence interval must exceed 80%.

Secondary outcomes

• Per-Patient Sensitivity and Specificity [Day 1 (the day of the index invasive coronary angiography procedure)]: Per-patient sensitivity and specificity of the dichotomously scored AutocathFFR per vessel compared to gold standard, invasive FFR, where an FFR ≤ 0.80 is scored "positive" and an FFR > 0.80 is considered "negative".
• Sensitivity, Specificity, PPV, and NPV Including "No FFR" Output [Day 1 (the day of the index invasive coronary angiography procedure)]: Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of AutocathFFR, including cases where the device produces a "No FFR" output (failure to compute), compared to the gold standard, invasive FFR.
• Device Success Rate [Through study completion, up to 8 months]: Device success, measured as the ratio of completed versus initiated AutocathFFR index calculations.
• Device-Related Malfunctions [Through study completion, up to 8 months]: Number of device-related malfunctions recorded throughout the study, summarized descriptively.

Sex: All, ages: 23 Years to any

Inclusion Criteria:

* Age > 22 years
* Subjects with stable angina pectoris, unstable angina pectoris, or NSTEMI and in whom invasive FFR was performed to assess a non-culprit stenosis in at least one coronary artery
* Subject underwent a clinically indicated Invasive Coronary Angiogram (ICA)
* Subject underwent invasive FFR with intravenous (IV) or intracoronary adenosine, or adenosine triphosphate (ATP), as hyperemic stimulus

Exclusion Criteria:

* Angiogram not obtained on a United States Food and Drug Administration (FDA)-cleared C-arm angiography system
* Invasive FFR not obtained on an FDA-cleared FFR catheter tip pressure transducer (also known as a pressure wire)
* Vessel size less than 2 mm
* Subject presents with an ST-elevation myocardial infarction (STEMI)
* Chronic total occlusion (CTO) in target vessel
* Prior coronary artery bypass graft (CABG) surgery, heart transplant or valve surgery, prior transcatheter aortic valve implantation/replacement (TAVI/TAVR), or severe left-sided valvular disease
* Known left ventricular ejection fraction (LVEF) ≤ 30%
* Arteries supplying akinetic or severe hypokinetic territories on target lesion, if already known based on prior imaging
* Thrombolysis in Myocardial Infarction (TIMI) flow grade equal to or lower than 2 at baseline
* Target lesion involves Left Main (stenosis >50%)
* Percutaneous coronary intervention (PCI) with stent in target vessel in past 12 months, or target vessel involves in-stent restenosis
* Target coronary vessel provides collaterals to a chronically occluded vessel
* Coronary angiography was not performed per standard of care at a cine frame rate of at least 7 frames per second
* Presence of aortic damping
* Distortion of the waveform
* Less than two different angiographic views for each lesion of interest in the Left Circumflex Artery (LCX), Left Anterior Descending Artery (LAD), or Right Coronary Artery (RCA)
* Lesion is not clearly visible

• Hector M. Garcia-Garcia, MD, Principal investigator, Medstar Health Research Institute