Single cell-resolution in situ sequencing reveals spatial dynamics of H3K27M-mutated diffuse midline gliomas and identifies myeloid-tumor cell interactions
In preparation - 2024
High-resolution CosMx sequencing ties myeloid neighborhoods to H3K27M tumor evolution during GD2-CAR therapy.
Role: Co-led spatial inferences and CAR T response analytics.
Pre-infusion biomarker classifier predicts lymphoma response and toxicity after CAR T cell
In preparation - 2024
Multi-cohort biomarker signature that stratifies outcomes prior to CAR T infusion in lymphoma.
Role: Designed the machine learning stack and prospective validation pipeline.
CCL8/CCL13-producing tumor-associated macrophages linked to poor outcomes after CAR T cell therapy for LBCL
Blood 144(Suppl. 1) · Cancer Cell (submitted) - 2024
Connects chemokine-driven macrophage states with post-infusion CAR T failure and charts intervention points.
Role: Led the single-cell integration and interactome modeling.
View ->
Unraveling temporal dynamics of GD2-CAR T cell therapy in H3K27M-mutant diffuse midline glioma through single-cell and multi-omic profiling
In preparation - 2024
Joint single-cell, spatial, and serum proteomic monitoring across GD2-CAR therapy.
Role: Architected longitudinal data model and multi-omic harmonization.
Humanized anti-CAR antibodies affects durable GD2 CAR T cell response in diffuse midline glioma
Neuro-Oncology (submitted) - 2025
Characterizes anti-CAR immune responses that influence durability of GD2-CAR therapy in pediatric brain tumors.
Role: Led computational analysis of antibody repertoires and survival associations.
Multiple instance learning identifies perturbation targets poised to improve clinical outcomes after CD19-CAR therapy for large B cell lymphoma
In preparation - 2024
Weakly supervised learning reveals cell states to target for durable CD19-CAR responses.
Role: Developed the MIL framework and interpretability tooling.
Patient-level prediction from single-cell data using attention-based multiple instance learning
NeurIPS 2025 (in review) - 2025
Attention-based MIL links diagnostic single-cell data to clinical outcomes for CAR T therapy.
Role: Driving methodology and benchmarking across multi-site datasets.
Overcoming T cell tolerance to tumor self-antigens through catch-bond engineering
Science (in review) - 2025
Catch-bond receptor engineering strategy that rescues self-antigen recognition for solid tumors.
Role: Contributed translational analytics on engineered T cell persistence.
View ->
Pan-hematologic cancer predictive biomarkers for chimeric antigen receptor T-cell therapy
Nature Biomedical Engineering (in review) - 2024
Cross-indication biomarker atlas forecasting efficacy across hematologic CAR programs.
Role: Built the harmonized feature store and calibration pipelines.
Spatial transcriptomics analysis of glioblastoma reveals three distinct regional programs of T-cell infiltration
Neuro-Oncology 26(Suppl. 8) viii12 - 2024
Maps immune niches across glioblastoma and highlights infiltration programs for intervention.
Role: Delivered spatial statistics and visual analytics for the abstract.
View ->
CD22-targeted CAR T cell single cell multiomic features linked to immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome
Blood 144(Suppl. 1) - 2024
Multiomic signatures that anticipate IEC-HLH-like toxicities after CD22-CAR therapy.
Role: Co-developed the multiomic analysis and toxicity classifier.
View ->
Biomechanical analysis of swallow-induced primary peristalsis in patients with EGJOO
Gastroenterology 160(6 Suppl. 1) S-102 - 2021
Quantified esophageal mechanics to differentiate EGJOO phenotypes using high-resolution manometry.
Role: Implemented the biomechanics pipeline and visualization suite.
View ->