Power law as a method for ultrasound detection of internal bleeding: in vivo rabbit validation

New detection methods for vascular injuries can augment the usability of an ultrasound (US) imager in trauma settings. The goal of this study was to evaluate a potential-detection strategy for internal bleeding that employs a well-established theoretical biofluid model, the power law. This law characterizes normal blood-flow rates through an arterial tree by its bifurcation geometry. By detecting flows that deviate from the model, we hypothesized that vascular abnormalities could be localized. We devised a bleed metric, flow-split deviation (FSD), that quantified the difference between patient and model blood flows at vessel bifurcations. Femoral bleeds were introduced into ten rabbits (∼5 kg) using a cannula attached to a variable pump. Different bleed rates (0% as control, 5%, 10%, 15%, 20%, 25%, and 30% of descending aortic flow) were created at two physiological states (rest and elevated state with epinephrine). FSDs were found by US imaging the iliac arteries. Our bleed metric demonstrated good sensitivity and specificity at moderate bleed rates; area under receiver-operating characteristic curves were greater than 0.95 for bleed rates 20% and higher. Thus, FSD was a good indicator of bleed severity and may serve as an additional tool in the US bleed detection.