Key result
An AI-based arterial input function model predicted myocardial blood flow with no significant difference compared to the reference dual-sequence method (2.79 vs 2.77 mL/min/g, P=0.33).
Why the study?
Quantification of myocardial blood flow from stress perfusion CMR is challenged by arterial input function estimation due to signal saturation from the non-linear relationship between gadolinium concentration and MR signal.
Does an AI-based arterial input function (AI-AIF) accurately quantify myocardial blood flow in patients undergoing stress perfusion CMR compared to dual-sequence acquisition?
Population
201 patients from centre 1 and a test set of consecutive and external patients (n = 44)
Comparison
AI-predicted unsaturated AIF vs reference dual-sequence acquisition AIF
Design
Retrospective deep learning model development and validation study
Authors
Loading...
May enable single-sequence quantitative stress perfusion CMR; leaves open prospective validation before routine use.
Observational (n=245)
Yes
Does an AI-based arterial input function (AI-AIF) accurately quantify myocardial blood flow in patients undergoing stress perfusion CMR compared to dual-sequence acquisition?
Mean Difference: -0.11
Absolute Event Rate: 2.79% vs 2.77%
p-value: p=0.33
An AI-based deep learning model can accurately predict unsaturated arterial input function from standard CMR images, enabling quantitative stress perfusion CMR with a single-sequence acquisition.
Scannell et al. (2022) conducted an observational in Patients undergoing stress perfusion cardiac magnetic resonance (n=245). AI-based arterial input function (AI-AIF) vs. Reference dual-sequence acquisition AIFs (DS-AIFs) was evaluated on Fully-automated myocardial blood flow (MBF) (bias of -0.11 mL/min/g, p=0.33). An AI-based arterial input function model predicted myocardial blood flow with no significant difference compared to the reference dual-sequence method (2.79 vs 2.77 mL/min/g, P=0.33).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: