Agreement between cardiac cycle time intervals and myocardial velocities measured by automated analysis of color tissue Doppler cine loops and by manual tracing of spectral pulsed-wave tissue Doppler waveforms of sheep fetal heart under normal and pathological conditions: An experimental study.

INTRODUCTION: This study evaluated agreement between automated analysis of color tissue Doppler (cTDI) cine loops and manual spectral pulsed-wave tissue Doppler (pwTDI) measurements of fetal cardiac cycle time intervals and myocardial velocities under normal and pathological conditions in an experimental model using fetal sheep.

MATERIAL AND METHODS: In 10 sheep fetuses, 25%-50% partial ascending aorta constriction was performed, and in nine, utero-placental blood flow was reduced by maternal uterine artery occlusion. Four fetuses served as controls. Experiments were performed after 4 to 5 days recovery, and cTDI cine loops and pwTDI waveforms were sequentially obtained during baseline and during fetal hypoxemia induced by maternal hypo-oxygenation. cTDI cine loops were processed using an automated, ECG-independent algorithm, whereas pwTDI waveforms were manually traced to measure cardiac cycle time intervals and myocardial velocities.

RESULTS: Fetal heart rate (FHR) showed excellent agreement between methods (intraclass correlation coefficient, ICC = 0.94). In contrast, myocardial velocities and cardiac time intervals demonstrated poor-to-moderate absolute agreement (ICC range 0.20-0.63). Bland-Altman analysis revealed systematic modality-specific offsets: the automated method yielded consistently lower velocities and longer isovolumic intervals. These offsets were stable across physiological states, with no metric showing phase-dependent changes in bias after Benjamini-Hochberg false discovery rate (FDR) correction (all q > 0.05). Despite limited absolute agreement, consistency and correlation (ICC(C,1) and Pearson r) were markedly higher than absolute agreement, indicating strong preservation of relative ranking across methods. Lin's concordance correlation coefficient (CCC) showed a similar pattern, with high concordance for FHR (CCC = 0.936) and lower concordance for most other metrics.

CONCLUSIONS: Automated cTDI is not directly interchangeable with manual pwTDI for absolute measurements of sheep fetal myocardial velocities or time intervals due to persistent modality-specific offsets. However, the automated method provides reproducible relative tracking and stable performance across different physiological conditions. Automated cTDI is therefore well suited for longitudinal assessments when interpreted using method-specific baselines or calibration. FHR was the only parameter demonstrating near-interchangeable performance.