My research brings together mechanistic modeling and data-driven AI to understand, predict, and ultimately improve outcomes in vascular disease, solid-organ transplantation, and oncology. I develop multiscale “digital twin” frameworks—combining agent-based models, differential-equation systems, computational fluid dynamics, and finite element methods with high-throughput data, 3D micro-CT–derived geometries, and modern machine-learning—to connect molecular and cellular mechanisms to clinical decision-making.

In vascular biology, I study how hemodynamics, inflammation, and gene programs shape maladaptive remodeling across settings such as vein-graft bypass, atherosclerosis, and in-stent restenosis. My group builds predictive models that couple local flow metrics to cellular dynamics and transcriptional signatures, with the goal of explaining lesion progression and guiding procedure design (e.g., stent deployment strategies) to reduce intimal hyperplasia and restenosis.

In transplantation, I focus on cardiac allograft vasculopathy and chronic rejection, integrating agent-based immuno-vascular modeling, CFD, and quantitative morphometry to dissect the interplay between geometry, flow, and immune-mediated injury. In parallel, we model the physiology of brain-dead donors to test hypotheses about inflammatory control and donor management, aiming to improve post-transplant graft survival. On the population side, we apply machine learning to large referral and evaluation cohorts to quantify risk and uncover inequities in kidney-transplant access, translating insights into actionable indices for earlier intervention.

In cancer systems biology, I build and validate computational models of prostate cancer/renal cancer bone metastasis that capture tumor–bone-vessel crosstalk and spatial dose effects, enabling in silico experimentation on single therpaies and combination strategies. Across projects, we pair simulations with imaging and experimental data to accelerate hypothesis generation and therapy optimization.

Finally, we develop deep-learning pipelines for microscopy image analysis (e.g., multiphoton tissue imaging) and unsupervised clinical phenotyping. Current efforts include stratifying patients with bloodstream infections and hypermobile Ehlers-Danlos syndrome, and extracting platelet-based inflammatory signatures to predict adverse outcomes—work that supports precision diagnostics and tailored care.

Keywords: multiscale modeling; agent-based modeling; computational fluid dynamics; systems biology; hemodynamics; transplant immunology; cardiac allograft vasculopathy; vein-graft remodeling; in-stent restenosis; digital twins; deep learning for microscopy; clinical machine learning; patient stratification; prostate cancer bone metastasis; micro-CT.