Comparison of the within-reader and inter-vendor agreement of left ventricular circumferential strains and volume indices derived from cardiovascular magnetic resonance imaging

Abstract

Purpose: Volume indices and left ventricular ejection fraction (LVEF) are routinely used to assess cardiac function. Ventricular strain values may provide additional diagnostic information, but their reproducibility is unclear. This study therefore compares the repeatability and reproducibility of volumes, volume fraction, and regional ventricular strains, derived from cardiovascular magnetic resonance (CMR) imaging, across three software packages and between readers.

Methods: Seven readers analysed 16 short-axis CMR stacks of a porcine heart. Endocardial contours were manually drawn using OsiriX and Simpleware ScanIP and repeated in both softwares. The images were also contoured automatically in Circle CVI42. Endocardial global, apical, mid-ventricular, and basal circumferential strains, as well as end-diastolic and end-systolic volume and LVEF were compared.

Results: Bland-Altman analysis found systematic biases in contour length between software packages. Compared to OsiriX, contour lengths were shorter in both ScanIP (-1.9 cm) and CVI42 (-0.6 cm), causing statistically significant differences in end-diastolic and end-systolic volumes, and apical circumferential strain (all p<0.006). No differences were found for mid-ventricular, basal or global strains, or left ventricular ejection fraction (all p<0.007). All CVI42 results lay within the ranges of the OsiriX results. Intra-software differences were found to be lower than inter-software differences.

Conclusion: OsiriX and CVI42 gave consistent results for all strain and volume metrics, with no statistical differences found between OsiriX and ScanIP for mid-ventricular, global or basal strains, or left ventricular ejection fraction. However, volumes were influenced by the choice of contouring software, suggesting care should be taken when comparing volumes across different software.

Fig 1. Schematic representation of workflow from imaging through to strain analysis.

16 short-axis cardiac MRI stacks were used from one specimen. Endocardial contours were traced in three software packages ScanIP, OsiriX, and Circle CVI42. Short-axis slices were grouped into three regions apex, mid-ventricle, and base, and regional circumferential strains were calculated as the average over slices in the region. Global circumferential strain, end-diastolic and end-systolic volumes, and left ventricular ejection fraction were also calculated.

Fig 2. Visual comparison of contouring approaches in ScanIP and Osirix.

A. Contour representation in ScanIP. Pixels within the endocardium are selected (based on thresholding or manual selection), to be exported in a binary image format. B. Contour representation in OsiriX. Contours markers are placed at points around the endocardium, and a spline curve is interpolated between these points.

Fig 3. Boxplots of peak data for the four circumferential strain variables investigated.

a) apical circumferential strain, b) mid-ventricular circumferential strain, c) basal circumferential strain, d) global circumferential strain. The circles represent data from individual readers. T1 and T2 are the first and second repetition of contouring in a given software. Equivalent data from CVI42 is shown in the fifth column.

Fig 4. Boxplots of peak data for the three clinical volume variables investigated.

a) end diastolic volume, b) end systolic volume, c) left ventricular ejection fraction. The circles represent data from individual readers. Equivalent data from CVI42 is shown in the fifth column.

Fig 5. Bland-Altman analyses of differences in endocardial contour lengths measured in ScanIP, OsiriX, and CVI42.

Results from test 1 and test 2 from ScanIP and OsiriX are shown. The CVI42 data was assumed to be 100% repeatable, and thus the same data was used for both tests 1 and 2, as previously described.