Research

Pancreatic ductal adenocarcinoma (PDAC) remains lethal because it is not a static molecular target but an evolving tissue ecosystem. Tumor cells occupy spatially structured habitats shaped by stroma, immune surveillance, vascular and metabolic constraint, and therapy-induced selection. Genomic alterations initiate and constrain malignant behavior, but clinical outcomes emerge from how cancer cells compete, cooperate, disperse, and adapt within these niches. My research program seeks to define the measurable rules governing PDAC ecosystem organization and evolution, then translate those rules into predictive and adaptive treatment strategies.

Quantitative PDAC Ecosystem Mapping

The program is organized around a Measure → Model → Perturb framework. First, we quantify PDAC ecosystem states using digital pathology, spatial biology, and landscape ecology. This pillar currently focuses on residual disease after neoadjuvant therapy, where tumor architecture may reveal which ecological states survive treatment and seed recurrence. We will develop locked, clinically interpretable Tumor Landscape Scores that improve recurrence prediction beyond conventional pathology, CA 19-9 dynamics, radiographic response, and treatment exposure.

---

Evolution-Guided Therapeutic Control

We model treatment resistance as an evolutionary process. Using orthotopic, immunocompetent, DNA-barcoded PDAC models, we will compare fixed-intensity therapy with adaptive and evolutionary dosing strategies for chemotherapy and emerging KRAS-directed therapies. The goal is to determine when treatment control is improved by preserving competitive drug-sensitive populations, limiting resistant clone expansion, and reducing toxicity without sacrificing disease control.

---

Dynamic Disease-State Sensing

Adaptive treatment requires feedback at a temporal resolution that conventional imaging cannot provide. We will study high-frequency blood-based biomarker kinetics, including point-of-care aptamer-based biosensing and exploratory label-free spectral approaches, as surrogate measures of tumor burden, response, regrowth, and impending escape. These platforms will be evaluated not as devices in isolation, but as control systems for adaptive oncology.

---

Stress-Biased Mutational Accessibility

As a high-risk, high-reward basic science pillar, we will investigate stress-biased mutation supply. The central hypothesis is that PDAC cells do not sample genomic alterations uniformly: oncogenic stress, therapeutic pressure, and microenvironmental constraint may remodel chromatin topology, replication timing, nuclear architecture, and chromosome-segregation vulnerability, shifting the probability distribution of structural variation before selection acts. This work will test whether recurrent catastrophic events such as chromothripsis reflect selection alone, constitutive biophysical vulnerability, or environmentally biased mutational accessibility.

---