Infarct quantification with cardiovascular magnetic resonance using "standard deviation from remote" is unreliable:...

Area of Science:

• Cardiovascular Imaging
• Medical Image Analysis
• Cardiac MRI

Background:

• Late gadolinium enhancement cardiovascular magnetic resonance (LGE-CMR) with the "n-SD from remote" method is a standard for infarct quantification.
• This method defines infarct based on signal intensity thresholds relative to remote myocardium.
• Lack of consensus on thresholding parameters and potential impact of region of interest (ROI) placement necessitates validation.

Purpose of the Study:

• To investigate the variability and agreement of the "n-SD from remote" method for myocardial infarction quantification.
• To assess the impact of ROI size, position, and the "n-SD" threshold on infarct size measurements.
• To evaluate the reliability of this commonly used LGE-CMR analysis technique.

Main Methods:

• Analysis of data from 214 patients across two multi-center trials.
• Comparison of infarct sizes derived from "n-SD from remote" with core lab reference delineations.
• Systematic variation of remote ROI position, ROI size, and "n-SD" thresholds.

Main Results:

• Variability in infarct size measurements reached up to 47% (ROI position), 48% (ROI size), and 40% ("n-SD").
• Low agreement was observed between the "n-SD from remote" method and core lab reference standards.
• Optimal "n-SD" thresholds exhibited high slice-by-slice variability (mean 5.3 ± 2.2).

Conclusions:

• The "n-SD from remote" method demonstrates significant unreliability for infarct quantification in LGE-CMR.
• Variability is influenced by ROI placement, size, "n-SD" threshold, and image acquisition parameters.
• The study recommends against using the "n-SD from remote" method and advocates for validated alternative approaches.