Intra- and Inter-Observer Variability of Quantitative Parameters Used in Contrast-Enhanced Ultrasound of Kidneys of Healthy Cats

SIMPLE SUMMARY: Contrast-enhanced ultrasound is a technique which enables assessment of tissue blood flow by injecting gas-filled micro-bubbles in the bloodstream. As a diagnostic and monitoring tool, this technique has shown great promise in multiple clinical conditions, while offering many advantages over other imaging techniques. A major challenge with this technique, however, is the fact that the imaging results can vary significantly, and there is no comprehensive understanding of their cause. Improving the understanding of the sources of variation could contribute to the development of guidelines to reduce variation and promote the application of CEUS in practice. The aim of this study was to assess variability in blood flow measurements in the kidney due to observers. As regions of interests are manually delineated using a specialized software, positioning of the regions may differ between observers and within the same observer at different occasions. This study found that a number of evaluated parameters had a good inter- and intra-observer agreement, while other parameters had a considerably lower inter-observer agreement compared to intra-observer-agreement. Therefore, it may be advisable that quantitative assessment of renal perfusion is not carried out by different researchers, especially if their experience levels differ. ABSTRACT: Contrast-enhanced ultrasound (CEUS) is a non-invasive imaging technique which allows qualitative and quantitative assessment of tissue perfusion. Although CEUS offers numerous advantages, a major challenge remains the variability in tissue perfusion quantification. This study aimed to assess intra- and inter-observer variability for quantification of renal perfusion. Two observers with different levels of expertise performed a quantitative analysis of 36 renal CEUS studies, twice. The CEUS data were collected from 12 healthy cats at 3 different time points with a 7-day interval. The inter- and intra-observer agreement was assessed by the intraclass correlation coefficient. Within and between observers, a good agreement was demonstrated for intensity-related parameters in the cortex, medulla, and interlobular artery. For some parameters, ICC(inter) was considerably lower than ICC(intra), mostly when the ROI encompassed the entire kidney or medulla. With the exception of time to peak (TTP) and mean transit time (mTTI), time-related and slope-related parameters showed poor agreement among observers. In conclusion, it may be advised against having the quantitative assessment of renal perfusion performed by different observers, especially if their experience levels differ. The cortical mTTI seemed to be the most appropriate parameter as it showed a favorable inter-observer agreement and inter-period agreement.