Sean Khozin, MD, MPH — Physician-Scientist, Oncologist & AI Innovator

About

At the Nexus of Medicine, Technology & AI

Dr. Sean Khozin is a physician-scientist and oncologist, internationally recognized for his leadership in leveraging data science, AI, machine learning, and real-world evidence to drive innovation in biomedical research, with a particular emphasis on oncology drug development.

A founding member of the FDA Oncology Center of Excellence, Dr. Khozin escaped the revolution in Iran and found his way to the United States, harmonizing his passion for patients and data into a career spanning startups, the FDA, Johnson & Johnson, and pioneering AI ventures, all while pursuing his love of music.

His work bridges the gap between cutting-edge technology and clinical care, with a focus on harnessing advanced analytics and AI to accelerate drug discovery, optimize therapeutic development, and improve patient outcomes.

79+
Publications

5,800+
Citations

18,000+
Research Reads

Current Roles

Leading at the Frontier

CEO Roundtable on Cancer & Project Data Sphere

Chief Executive Officer

Directing public health programs and pioneering initiatives focused on the development of AI foundation models in oncology.

Phyusion Bio

Founder

Subsidiary of Phyusion

Advancing innovations at the nexus of biology, technology, and AI to accelerate drug discovery and optimize therapeutic development.

Massachusetts Institute of Technology

Research Affiliate

Focusing on novel applications of AI and machine learning technologies to accelerate drug discovery and optimize therapeutic development strategies.

Precision Signals

Host

A podcast from the CEO Roundtable on Cancer, examining the hard problems in oncology and exploring practical solutions that connect science to patient impact.

YouTube

Career

Professional Journey

Present

CEO, CEO Roundtable on Cancer & Project Data Sphere

Directing public health programs and pioneering AI foundation model initiatives in oncology.

Present

Founder, Phyusion Bio & Research Affiliate, MIT

Novel applications of AI/ML to accelerate drug discovery and optimize therapeutic development.

Led the restructuring and acquisition of the precision oncology enterprise leveraging real-world data and analytics.

Implemented cutting-edge data science solutions supporting the development of innovative medicines and vaccines.

Established the FDA's first data science and technology incubator under special federal authorities.

Helped build and launch the FDA's Oncology Center of Excellence, shaping the agency's approach to oncology drug regulation and review.

Conducted clinical research in oncology at the U.S. National Cancer Institute.

Pioneering TechBio company developing telemedicine, point-of-care data visualization, and advanced analytical systems.

Training

Fellowship, National Cancer Institute

Board-certified oncologist. Fellowship training following internal medicine residency.

Education

MD & MPH

Doctor of Medicine, University of Maryland School of Medicine. Master of Public Health, George Washington University. BS, University of Maryland.

Initiative

FDA INFORMED

Information Exchange and Data Transformation (INFORMED) was launched in 2015 at the U.S. FDA with special authorities from the Department of Health and Human Services.

As a technology and data science incubator, INFORMED was designed as a public-private collaboration sandbox to de-risk innovations with transformative potential in biomedical research and therapeutic development.

The effort was instrumental in establishing the foundation for the use of real-world evidence, digital health solutions, and AI/ML in drug development and regulatory decision-making.

Real-World Evidence

Pioneered frameworks for integrating real-world data into regulatory decision-making.

Digital Health

Advanced digital health solutions, including smart devices and decentralized clinical trials.

AI & Machine Learning

Established foundations for AI/ML adoption across the product life cycle.

References

From Big Data to Smart Data: FDA's INFORMED Initiative Nature Reviews Drug Discovery, 2017 INFORMED: An Incubator at the US FDA for Driving Innovations in Data Science and Agile Technology Nature Reviews Drug Discovery, 2018

Research

Dr. Khozin's research spans regulatory science, clinical trial methodology, and data science, with a focus on modernizing how evidence is generated and applied in cancer drug development.

Precision Oncology and Regulatory Science

As a clinical investigator at the National Cancer Institute, Dr. Khozin conducted clinical research in thoracic oncology and rare tumors, contributing to early-phase trials of molecularly targeted agents and novel combination regimens in lung cancer and thymic epithelial tumors. This work included studies of IGF-1R-targeted therapy in thymoma, survivin suppression in NSCLC, the identification of ALK rearrangements in rare lung cancer histologies, characterization of germline EGFR T790M mutations, and the use of FDG-PET imaging as a predictive and prognostic tool in thymic malignancies. At the FDA, Dr. Khozin led several landmark accelerated and regular approvals that defined the modern paradigm of biomarker-driven oncology, including ceritinib (ALK-positive NSCLC), erlotinib (EGFR-mutant NSCLC), osimertinib (EGFR T790M NSCLC), and atezolizumab (metastatic NSCLC).

Clinical Cancer Research Osimertinib for the Treatment of Metastatic EGFR T790M Mutation-Positive Non-Small Cell Lung Cancer Clinical Cancer Research FDA Approval: Ceritinib for the Treatment of Metastatic ALK-Positive Non-Small Cell Lung Cancer The Oncologist U.S. FDA Approval Summary: Erlotinib for the First-Line Treatment of Metastatic NSCLC with EGFR Mutations Clinical Cancer Research U.S. FDA Approval Summary: Atezolizumab for Metastatic Non-Small Cell Lung Cancer The Lancet Oncology Cixutumumab for Patients with Recurrent or Refractory Advanced Thymic Epithelial Tumours: A Multicentre, Open-Label, Phase 2 Trial Clinical Cancer Research ¹⁸F-FDG PET in the Management of Patients with Thymic Epithelial Tumors Annals of Oncology A Phase I/II Study of Sepantronium Bromide (YM155, Survivin Suppressor) with Paclitaxel and Carboplatin in Patients with Advanced Non-Small-Cell Lung Cancer Journal of Thoracic Oncology Hepatoid Carcinoma of the Lung with Anaplastic Lymphoma Kinase Gene Rearrangement Clinical Lung Cancer Concurrent Molecular Alterations in Tumors with Germ Line Epidermal Growth Factor Receptor T790M Mutations

Real-World Evidence and Clinical Trial Design

Dr. Khozin’s research has helped define the methodological foundations for using electronic health records and other real-world data to generate validated clinical endpoints, construct robust external control arms, and support regulatory decision-making in oncology. His work demonstrated that real-world endpoints can meaningfully correlate with overall survival and be implemented at scale with scientific rigor. He has also advanced hybrid trial designs that integrate randomization with external data sources, developed approaches to improve the interpretation of clinical trials for underrepresented populations, and established practical frameworks for incorporating real-world evidence into both regulatory and treatment decisions.

Journal of the National Cancer Institute Real-World Data for Clinical Evidence Generation in Oncology Cancer Real-World Progression, Treatment, and Survival Outcomes During Rapid Adoption of Immunotherapy for Advanced NSCLC The Oncologist Real-World Outcomes of Patients with Metastatic NSCLC Treated with PD-1 Inhibitors in the Year Following U.S. Regulatory Approval JCO Clinical Cancer Informatics Characterizing the Feasibility and Performance of Real-World Tumor Progression End Points Nature Communications The Design and Evaluation of Hybrid Controlled Trials That Leverage External Data and Randomization Biochimica et Biophysica Acta Reviews on Cancer A Novel Approach to Reducing Disparities in Health Outcomes by Enhancing Interpretation of Cancer Clinical Trials for Underrepresented Patient Groups Advances in Therapy Real-World Evidence Acceptability and Use in Breast Cancer Treatment Decision-Making Frontiers in Medicine Real-World Evidence for Regulatory Decision-Making: Updated Guidance from Around the World Health Affairs Real-World Evidence in Support of Precision Medicine: Clinico-Genomic Cancer Data as a Case Study

Artificial Intelligence and Machine Learning in Drug Development

Dr. Khozin's work in AI spans foundation models for oncology, LLM-based clinical trial data analysis, and regulatory frameworks for AI/ML adoption across the drug development life cycle. At the FDA, he established Information Exchange and Data Transformation (INFORMED), the agency's first data science and technology incubator, which laid the groundwork for integrating AI and digital health into drug development and regulatory decision-making.

A central focus of Dr. Khozin’s current work is modernizing how tumor burden is measured in clinical trials. In a large FDA-led meta-analysis of more than 14,000 patients enrolled in registration trials, he found that two independent radiologists reviewing the same scans agreed on tumor response classification only about 70% of the time. In ovarian and pancreatic cancers, discordance reached as high as 45%, with striking cases in which one reader determined a complete response while the other concluded progressive disease in the very same patient.

Inter-Reader Response Classification

Primary Investigator vs. Independent Radiologist · N = 13,677

		Independent Radiologist
		CR	PR	SD	PD	NE	Total
Primary Investigator	CR	305	250	116	14	25	710
	PR	240	3,366	908	126	33	4,673
	SD	64	849	3,522	668	74	5,177
	PD	9	58	507	1,890	134	2,598
	NE	1	13	31	49	425	519
	Total	619	4,536	5,084	2,747	691	13,677

Concordant Discordant

Response categories per RECIST 1.1 (Response Evaluation Criteria in Solid Tumors): CR = Complete Response · PR = Partial Response · SD = Stable Disease · PD = Progressive Disease · NE = Not Evaluable. The highlighted diagonal cells are the only instances where both readers agreed on the same response classification. Overall discordance across tumor types was approximately 30%, with notably higher rates in ovarian and pancreatic cancers.

FDA meta-analysis of studies submitted to the agency · Khozin et al., Unpublished, 2017.

AutoRECIST

AutoRECIST applies AI to automate and standardize tumor response assessment, eliminating inter-reader variability that undermines the reliability of clinical trial endpoints. By replacing subjective radiologist interpretation with algorithmic consistency, AutoRECIST produces reproducible, auditable response classifications at scale.

Total Tumor Burden Index (TTBI)

TTBI replaces conventional categorical RECIST endpoints with a continuous volumetric measure of total disease burden. By integrating 3D tumor segmentation across all lesions, TTBI enables earlier and more granular detection of treatment effect, capturing changes that binary response categories miss.

NEJM AI The Use of Artificial Intelligence for Cancer Therapeutic Decision-Making Clinical and Translational Science AI for Cancer Classification medRxiv Artificial Intelligence in Biomedical Data Analysis: A Comparative Assessment of Large Language Models for Automated Clinical Trial Interpretation and Statistical Evaluation NPJ Digital Medicine Artificial Intelligence and Machine Learning in Clinical Development: A Translational Perspective Nature Reviews Drug Discovery INFORMED: An Incubator at the US FDA for Driving Innovations in Data Science and Agile Technology Nature Reviews Drug Discovery From Big Data to Smart Data: FDA's INFORMED Initiative Clinical Pharmacology and Therapeutics Decentralized Trials in the Age of Real-World Evidence and Inclusivity in Clinical Investigations Advances in Therapy Generating Real-World Tumor Burden Endpoints from Electronic Health Record Data: Comparison of RECIST, Radiology-Anchored, and Clinician-Anchored Approaches Annals of Oncology Analysis of Time-to-Treatment Discontinuation of Targeted Therapy, Immunotherapy, and Chemotherapy in Clinical Trials of Patients with NSCLC

Press

The Wall Street Journal Clinical Drug Trials May Be Coming to Your Doctor's Office The Wall Street Journal Open Your App and Say Ahh

Google Scholar Profile ResearchGate Profile

Music

Jazz Meets Classical in Suspended Time

Open in Spotify

Featured Essay

Physiological Harmony and Dissonance

The Confluence of Math, Music, and Medicine

How physiological state changes can be represented musically — the composition Perpetual Drift proposes that transitions in human physiology, detectable through computational methods, can also be heard as shifts in tonality and dissonance.

Read on Substack →

Leadership

Board Appointments

Society for Translational Oncology

Alliance for AI in Healthcare

Digital Medicine Society

Cytometric Therapeutics

Connect

Get in Touch

Interested in collaborating on research, speaking engagements, or advancing AI-driven innovation in healthcare?

LinkedIn X (Twitter) Phyusion Bio Substack Google Scholar

Sean Khozin