About

Rajan Jain, MD, is the William Wikoff Smith Associate Professor in Cardiovascular Research at the University of Pennsylvania Perelman School of Medicine. He is also the Founding Co-Director of the Penn Measey Scholars in Molecular Medicine, the Developmental Biology Interest Group Leader at the Penn Epigenetics Institute, and the Associate Director of the Penn Cardiovascular Institute. His research focuses on understanding how genome organization influences organogenesis, with particular interest in how the three-dimensional folding and spatial organization of the genome in the nucleus establish and maintain cell identity. Jain's work leverages models of stem cell biology and interdisciplinary approaches to dissect the mechanisms underlying gene regulation, chromatin structure, and epigenetics, especially in the context of heart development and disease. His studies aim to elucidate the rules governing genome architecture and lamina-chromatin interactions, which have implications for congenital and adult cardiac diseases, as well as regenerative therapies.

Research topics

• Artificial Intelligence
• Computer Science
• Medicine
• Machine Learning
• Data Mining
• Internal medicine
• Pathology
• Intensive care medicine
• Biology
• Data science
• Surgery
• Psychology
• Pediatrics
• Emergency medicine

Selected publications

• ASFNR Current State of Practice in Neuroimaging of Distal Medium Vessel Occlusion Stroke

  American Journal of Neuroradiology · 2026-05-14

  article

  BACKGROUND: Distal medium vessel occlusions (DMVO) constitute approximately 25%-40% of acute ischemic stroke. These potentially disabling strokes remain diagnostically challenging due to vessel caliber, tortuosity, and low sensitivity of standard CT angiography. RECENT DEVELOPMENTS: in the MT arm. PURPOSE: Summarize current evidence and provide a "state of practice" guide for radiologists on DMVO detection, workflow standardization, triage, and imaging-based prognostication. KEY POINTS: Optimized CTA (including multiphase), CT Perfusion (territorial Tmax), and MRI DWI/SWI improve diagnostic confidence for DMVO. Certain perfusion parameters indicative of collateral status, for instance, rCBV index and hypoperfusion intensity ratio, have prognostic value. Structured reporting of important positive and negative radiologic findings can guide neurologic triage despite neutral trials. CONCLUSION: Radiologists play a central role in DMVO diagnosis and prognostication. Standardized imaging workflows are essential in the post-trial landscape.

  PublisherDOI

• Categorical and phenotypic image synthetic learning as an alternative to federated learning

  Nature Communications · 2025-10-23

  articleOpen access

  Multi-center collaborations are crucial in developing robust and generalizable machine learning models in medical imaging. Traditional methods, such as centralized data sharing or federated learning (FL), face challenges, including privacy issues, communication burdens, and synchronization complexities. We present CATegorical and PHenotypic Image SyntHetic learnING (CATphishing), an alternative to FL using Latent Diffusion Models (LDM) to generate synthetic multi-contrast three-dimensional magnetic resonance imaging data for downstream tasks, eliminating the need for raw data sharing or iterative inter-site communication. Each institution trains an LDM to capture site-specific data distributions, producing synthetic samples aggregated at a central server. We evaluate CATphishing using data from 2491 patients across seven institutions for isocitrate dehydrogenase mutation classification and three-class tumor-type classification. CATphishing achieves accuracy comparable to centralized training and FL, with synthetic data exhibiting high fidelity. This method addresses privacy, scalability, and communication challenges, offering a promising alternative for collaborative artificial intelligence development in medical imaging.

  PublisherOA PDFDOI

• Table S8 from Clinical Efficacy of ONC201 in H3K27M-Mutant Diffuse Midline Gliomas Is Driven by Disruption of Integrated Metabolic and Epigenetic Pathways

  2025-02-18

  preprintOpen access

  <p>Expression values, Spearman’s correlation coefficient, and GSEA of genes associated with positive or negative correlation with radiographic response to ONC201 treatment.</p>

  PublisherDOI

• Supplementary Data S2 from Clinical Efficacy of ONC201 in H3K27M-Mutant Diffuse Midline Gliomas Is Driven by Disruption of Integrated Metabolic and Epigenetic Pathways

  2025-02-18

  preprintOpen access

  <p>Protocol for NCT03134131, ONC201-018: Expanded Access to ONC201 for Patients with H3K27M-mutant and/or Midline High Grade Gliomas.</p>

  PublisherDOI

• Table S3 from Clinical Efficacy of ONC201 in H3K27M-Mutant Diffuse Midline Gliomas Is Driven by Disruption of Integrated Metabolic and Epigenetic Pathways

  2025-02-18

  preprintOpen access

  <p>Individual cases from historical control datasets.</p>

  PublisherDOI

• Supplementary Figures 1-15 from Clinical Efficacy of ONC201 in H3K27M-Mutant Diffuse Midline Gliomas Is Driven by Disruption of Integrated Metabolic and Epigenetic Pathways

  2025-02-18

  preprintOpen access

  <p>Contains supplementary data and table titles and as well as supplementary figures with associated titles and legends.Fig. S1. Selection method for planned efficacy analysis of ONC201 in patients with H3K27MDMG.Fig. S2. Progression-free survival from diagnosis of trial patients with non-recurrent H3K27MDMG treated with ONC201.Fig. S3. Swimmers’ plot of ONC201 response by recurrence status, tumor location, and ONC201 trial.Fig. S4. ONC201 efficacy is independent of TP53 mutation status and chromosomal instability.Fig. S5. Cox proportional hazard regression to assess the effect of ONC201 after adjusting for confounders.Fig. S6. Survival of patients with H3K27M-DMG treated with ONC201 versus ONC201untreated historical controls.Fig. S7. Survival of patients with H3K27M-DMG treated with ONC201 versus ONC201untreated patients (PNOC003 or HERBY Phase II trials).Fig. S8. Molecular attributes of patients with H3K27M-DMG treated with ONC201.Fig. S9: Survival of H3K27M-DMG mice models treated with ONC201.Fig. S10. Integrative analysis of in vitro and human tumor metabolic gene expression in response to ONC201.Fig. S11. ONC201-mediated L2HG production increases H3K27me3 in H3K27M-DMG cells.Fig. S12. ONC201-induced increase in H3K27me3 is mediated by lactate dehydrogenase.Fig. S13. ONC201 alters genomic chromatin accessibility and H3K27ac distribution in H3K27M-DMG cells.Fig. S14. ONC201 increases global H3K27me3 in patient samples.Fig. S15. ONC201 does not cause hypermethylation leading to a glioma CpG island methylator phenotype.</p>

  PublisherDOI

• Table S2 from Clinical Efficacy of ONC201 in H3K27M-Mutant Diffuse Midline Gliomas Is Driven by Disruption of Integrated Metabolic and Epigenetic Pathways

  2025-02-18

  preprintOpen access

  <p>Individual patients with H3K27M-mutant DMG treated with ONC201.</p>

  PublisherDOI

• Table S6 from Clinical Efficacy of ONC201 in H3K27M-Mutant Diffuse Midline Gliomas Is Driven by Disruption of Integrated Metabolic and Epigenetic Pathways

  2025-02-18

  preprintOpen access

  <p>Radiographic response versus gene expression Spearman's correlation coefficient for all protein-coding genes.</p>

  PublisherDOI

• IMG-33. Unveiling glioblastoma heterogeneity with deep learning derived MRI subtyping: insights from the global ReSPOND consortium

  Neuro-Oncology · 2025-11-01

  articleOpen access

  Abstract PURPOSE This study aims to identify distinct imaging subtypes of glioblastoma using multi-modal MRIs from the ReSPOND consortium, providing insights into tumor heterogeneity to inform personalized treatment approaches. METHODS We analyzed 3,145 subjects with multi-modal MRIs (T1, T2, T1Gd, FLAIR) from 16 geographically distinct institutions across North America, Europe, and Asia. Radiomic features were extracted using Masked Auto-Encoder (MAE), a deep learning architecture trained through self-supervised learning on 23,608 MRIs across 11-dimensional MRI-derived imaging measures (conventional, diffusion, and perfusion protocols). For clustering analysis, we used features extracted from the four structural modalities available in all subjects. Following feature extraction, we applied ComBat harmonization to remove institutional batch effects, performed dimensionality reduction using cross-validated PCA, and employed K-medoids clustering with consensus analysis (100 random data partitions with a 90/10 split for clustering/validation) to identify stable subtypes, with optimal cluster number determined using adjusted rand index. Additionally, we extracted morphological, intensity, and textural features to characterize the identified subtypes in an interpretable manner. RESULTS Our analysis revealed three reproducible glioblastoma subtypes. Statistical analysis demonstrated significant differences between subtypes in tumor morphology and spatial location features. Subtype 1 showed centrally located, spherical tumors; Subtype 2 exhibited peripheral, irregular morphologies; Subtype 3 had mixed features (p < 0.001). Key features included right hemisphere ratio, centroid Z position, sphericity, and shape convexity ratio (p < 0.001). Survival analysis indicated median overall survival of 12.3, 14.5, and 16.8 months for Subtypes 1, 2, and 3, respectively (p = 0.042, borderline). Preliminary genomic analysis showed frequent TP53 and RB1 co-mutations, with distinct molecular patterns per subtype. CONCLUSION This study demonstrates that multi-modal MRIs can successfully identify glioblastoma subtypes characterized by deep learning derived radiomic features. Future work will focus on validation with comprehensive molecular and tissue-based profiles, survival outcomes, and treatment responses to support personalized medicine.

  PublisherOA PDFDOI

• Supplementary Data S1 from Clinical Efficacy of ONC201 in H3K27M-Mutant Diffuse Midline Gliomas Is Driven by Disruption of Integrated Metabolic and Epigenetic Pathways

  2025-02-18

  preprintOpen access

  <p>Protocol for NCT03416530, ONC201-014: ONC201 in Newly Diagnosed Diffuse Intrinsic Pontine Glioma and Recurrent/Refractory Pediatric H3K27M Gliomas.</p>

  PublisherOA PDFDOI

Recent grants

• NIH Grant K08HL119553

  NIH · $607k · 2018

Frequent coauthors

• Mark R. Gilbert

  University of Missouri

  467 shared

• Daniel T. Chang

  400 shared

• Erik P. Sulman

  New York University

  335 shared

• Patrick Y. Wen

  316 shared

• Michael Weller

  University Hospital of Zurich

  305 shared

• Olivier Gevaert

  302 shared

• Ivan Smirnov

  University of California, San Francisco

  300 shared

• J. Costello

  Centre National de la Recherche Scientifique

  300 shared

Labs

• Rajan Jain LabPI

Awards & honors

• William Wikoff Smith Associate Professor in Cardiovascular R…
• Founding Co-Director, Penn Measey Scholars in Molecular Medi…
• Developmental Biology Interest Group Leader, Penn Epigenetic…
• Associate Director, Penn CVI