Endovascular mechanical interventions and local vascular drug delivery (LVDD) affect vascular healing but differently across patients. Device design, formulation and deployment influence but do not strictly predict, clinical response. Two implications emerge. The first is that repair biology is inherently variable and clinical outcomes can never be predicted. The second is that response variability can be categorized but depends on more proximate physiological forces, which, if identified, can allow better definition of procedural success and prediction of adverse events. We embrace the latter view, that dimensional and anatomic parameters can speak to procedural success but only over a wide range and with little discrimination, and it is physiologic consequences of vascular manipulation that are the strongest predictors of performance. Guided by our past funded work, the current proposal examines the hypothesis that induced patterns of flow and drug distribution are the proximate drivers of biological response to vascular interventions, and can better predict the therapeutic consequences of interventions in animals and now in humans as well. In past cycles we created analytical, immunohistochemical and imaging technologies to characterize repair and flow disruptions after vascular intervention in animal systems, and a quantitative framework to predict pharmaco-kinetics and -dynamics of LVDD. We extend this work by using these resources in more controlled animal models, in silico models and with human clinical data.
Three specific aims will: (1) define the limits of traditional descriptors of complex interventions no matter how precise, (2) develop and validate computational models that predict near-wall flow patterns and drug distributions in real-world settings in complex animal models and provide statistical tools to determine predictive roles of these forces relative to procedural variables alone, and (3) investigate whether the variables determined predict clinical outcomes in humans. Innovation exists in the tools we employ, data we analyze, approach we take, implications of the work and assembly of a pandisciplinary group of investigators with a legacy of collaboration. OCT imaging quantifies biologic effect in situ, providing high-resolution images of stent-vessel geometry in animals and humans in an identical manner. In-house computer algorithms extract procedural geometries, which create 3D computational models of physiological flow disruption and drug distribution. MALDI corroborates drug distribution after local delivery. Animal experiments use custom-made drug delivery devices to allow precise control in defining complex interventional procedures in vivo. Access to the University Hospital's extensive OCT databank of clinical images offers a rich test bed for clinical validation. Statisticl innovation will accurately describe the effects of stent characteristics, flow, and drug distributin on biological outcomes in a setting where the data have multilevel, longitudinal, and spatial structure. The lessons learned may extend our understanding of basic vascular biology, local drug delivery in stents and other combination devices, tissues, or pathologic conditions.
Our past work explained the value and risks of local vascular drug delivery (LVDD) - creating a quantitative framework to predict pharmacokinetics and dynamics of LVDD, and animal and benchtop models to characterize tissue repair after vascular intervention and drug delivery. As direct extension we now meld non- clinical (animal and in silico) with human data to extract mechanistic insight into performance from clinical trials with LVDD. Clinical observations provide model inputs, model predictions in silico are tested in animal models using clinical devices, and clinical events with these same devices are examined to determine if the determinants of LVDD effects in animals are borne out in human clinical trials.
Showing the most recent 10 out of 158 publications
