Bench top model for comparing flow rate measurements from a thermodilution catheter to a novel software based method of flow quantification using digital subtraction angiography

Purpose Intra-access blood flow rates are used for monitoring dialysis access function. We hypothesize that flow can be calculated from Digital Subtraction Angiograms (DSA) during fistulography. The purpose of this study is to compare DSA-based flow quantification algorithms with thermodilution catheters. Materials and Methods Using a flow model with a 37 °C water bath and 3/8′ silicone tubing similar to prior studies (1), thermodilution and 8 DSA flow algorithms were used at rates from 400-1200 mL/min. Power injectors were used for room temperature saline and contrast injections. Thermodilution was repeated 6 times and DSA was repeated 3 times at 3, 6, 7.5, and 10 frames per second (FPS). Actual flow rate was estimated by measuring the volume displaced in 1 minute. Mean and standard deviation (SD) of each algorithm was determined and the algorithm with the closest flow to estimated flow and smallest standard deviation was correlated with the thermodilution catheter measurements. Repeatability coefficient (RC) for DSA and thermodilution methods was estimated for each flow rate and the limits of agreement (LOA) were assessed between the 2 methods at each frame rate. Results Cross Correlation (CC) had the smallest SD and closest values to actual flow. The mean difference between CC and thermodiluton was 44.2±75; 144.7±37; 230.6±52; and 232.6±25 mL/min at 3, 6, 7.5, and 10 FPS. The SD of all other algorithms at all frame rates were >100 mL/min. RC increased from 149.2 to 709.7 with increasing flow rate for CC and from 46.9 to 191.2 for thermodilution. Higher RC indicates more variability in measurements. LOA increased with increasing frame rate from 52±209 to 232±91 mL/min. Conclusion Given that flow rates from the DSA method were consistently higher and more variable than thermodilution values, further refinement of the software algorithms is needed. Reproducibility and accuracy need to be improved prior to clinical use of this method for flow estimation. Reference 1. Heerwagen ST, Lonn L, Schroeder TV, Ladefoged SD, Hansen MA. Catheter-based flow measurements in hemodialysis fistulas - bench testing and clinical performance. J Vasc Access. 2012 Jan-Mar;13(1):45-50. doi: 10.5301/JVA.2011.8443. PubMed PMID: 21725955.