Introduction
Segmed is a real-world imaging data company founded in 2019 by a team of Stanford engineers and physicians, headquartered in the Stanford, California area. David Gascoigne serves as Chief Executive Officer per Segmed’s about page reviewed in July 2026. Co-founders include Martin Willemink, MD, PhD, Chief Scientific Officer; Jie Wu, PhD, Chief AI Officer; Adam Koszek; and Cailin Hardell — reflecting origins in Stanford radiology and machine learning research translated into a commercial data platform for medical AI and life-sciences R&D.
Segmed provides de-identified medical imaging studies at scale from more than two thousand healthcare partner sites across five continents — a dataset infrastructure layer supporting FDA clearances, foundation models, and pharmaceutical real-world evidence programs referenced in company materials.
What the company does
Segmed acquires, de-identifies, standardizes, and distributes medical imaging data — including X-rays, CT, MRI, ultrasound, and digital breast tomosynthesis — alongside reports and multimodal datasets for AI training, algorithm validation, and clinical research. Its Openda platform (formerly Segmed Insight), rebranded July 2024, offers self-serve access tools for researchers needing fit-for-purpose imaging cohorts without months-long manual hospital contracting.
The March 5, 2026 release of the largest open-source annotated breast tomosynthesis dataset — in collaboration with iMerit and Advocate Health — exemplifies Segmed’s strategy of lowering research barriers through curated public datasets while commercializing broader proprietary libraries for enterprise customers. The dataset includes biopsy-confirmed cases interpreted by U.S. board-certified, MQSA-certified radiologists with specialist segmentation.
Segmed cites participation in more than thirty-five FDA clearances and multiple foundation model projects — positioning real-world imaging data as regulatory-grade infrastructure rather than ad hoc research convenience samples.
Who it serves
Segmed serves medical device and AI companies pursuing FDA submissions, pharmaceutical teams building external control arms or imaging biomarker programs, and academic researchers needing diverse populations. Life-sciences R&D customers use Segmed for oncology, neurology, and cardiology imaging cohorts; medical AI startups use Openda for rapid prototyping and validation; healthcare providers partner as data contributors in Segmed’s network spanning fifty U.S. states and international sites.
Company background
Willemink co-founded Segmed while directing cardiovascular imaging research at Stanford Radiology, joining full-time in April 2022 after leading the company through early funding including a $10.4 million Series A. Wu co-founded in May 2019 and joined full-time after completing his Stanford PhD in 2021. Segmed participated in Y Combinator W20 and evolved leadership from Willemink’s early CEO tenure to Gascoigne as CEO per current about page — a transition reflecting commercial scaling beyond founder-led scientific leadership.
Gascoigne leads go-to-market and operations while Willemink and Wu remain co-founders driving scientific and AI strategy — a structure common among Stanford-spinout data platforms.
Product and AI capabilities
Segmed’s platform emphasizes regulatory-grade de-identification, structured metadata, and multimodal linkage between imaging and reports — capabilities AI developers need for reproducible training and prospective validation studies. Openda product features announced with the 2024 rebrand include expanded database integration and self-serve cohort selection tools aimed at reducing time from research question to dataset access.
Blog and resources content on segmed.ai documents FDA draft guidance implications for AI-enabled devices and real-world evidence requirements — Segmed markets expertise at the intersection of imaging data governance and regulatory submission strategy, not raw pixel delivery alone.
Key developments
2019: Segmed co-founded in Stanford area. July 1, 2024: Openda rebrand and Insight Platform evolution (official news). March 5, 2026: Open-source annotated breast tomosynthesis dataset with iMerit and Advocate Health (official news). David Gascoigne CEO per about page.
Why it matters
Segmed matters because medical imaging AI quality ceilings are often set by training data diversity and labeling rigor — not model architecture alone. Open-source dataset releases like DBT-2026 demonstrate Segmed’s role in community infrastructure while commercial datasets fund scaled operations. Hospital AI teams and device manufacturers evaluating data vendors should assess population diversity, annotation provenance, and regulatory documentation depth — Segmed’s cited FDA clearance support track record is the credibility signal Brel tracks for analytics infrastructure companies.
Sector context
Segmed belongs in Brel’s AI in healthcare coverage under healthcare analytics and real-world imaging data, upstream from algorithm vendors such as Lunit or Kheiron Medical Technologies. Readers should distinguish data platform purchases from diagnostic AI products — complementary but separate budget lines in most health systems.
Sources and references
Segmed open-source mammography dataset release (March 5, 2026) and co-founder leadership profile.
- Segmed — 3D mammography dataset release (segmed.ai, March 5, 2026)
- Segmed — Martin Willemink co-founder profile (segmed.ai)
Reviewed July 10, 2026. David Gascoigne is CEO. Cansu Safak CSV error corrected. diagnosticimaging.com removed per batch 3 rules.