Segmed – Weekly Recap

Segmed spent the week underscoring its role as a data infrastructure partner for medical imaging AI, rather than a pure algorithm developer, through new international and ecosystem collaborations. The company highlighted research trends in multimodal imaging AI, precision oncology data, and foundation models as it seeks deeper integration into life sciences and healthcare AI workflows.

A key development was Segmed’s partnership with Japan’s Medical AI Promotion Institute, or MAPI, aimed at giving Japanese AI developers streamlined access to de-identified imaging data from more than 2,800 healthcare institutions. MAPI will handle local regulatory guidance, licensing, and curation in Tokyo, addressing historical bottlenecks around compliance and fragmented access to overseas datasets.

The MAPI collaboration is positioned to meet rising demand in Japan for diverse imaging data to train and validate algorithms intended for global use, potentially expanding Segmed’s recurring revenue from data access and services. While no financial terms were disclosed, the agreement strengthens Segmed’s presence in a strategically important healthcare AI market and supports its broader data-as-a-service model.

Segmed also continued to spotlight its collaboration with Verily Health across the Verily Pre platform, Exchange, and Workbench tools, emphasizing its contribution of de-identified longitudinal breast and lung cancer imaging cohorts. The company is supplying a breast cancer cohort that integrates 3D imaging, biopsy-confirmed outcomes, and longitudinal EHR data in a research-ready, governed environment on Verily Exchange.

By aligning with Verily’s AI-native workflow and secure co-analysis infrastructure, Segmed aims to reduce data procurement and curation friction for pharma, medtech, and academic researchers. This positioning could increase demand for Segmed’s multimodal, analysis-ready datasets as real-world evidence and imaging-intensive oncology research gain prominence in drug development and clinical AI.

Through its Decoded newsletter and social media, Segmed highlighted a multicenter PLOS Medicine study on knee osteoarthritis that used a multimodal model combining MRI radiomics, biochemical biomarkers, and clinical variables. Reported AUCs of 0.880–0.913 and improved physician prediction accuracy underscored the performance benefits of integrated data over single-modality models.

This focus on multimodal imaging AI and decision-support aligns Segmed with broader trends toward personalized risk prediction and clinically contextualized radiology tools. It reinforces the company’s strategic emphasis on supplying standardized, de-identified imaging data that underpin foundation models and advanced analytics, rather than commercializing stand-alone algorithms.

Segmed further emphasized thought leadership in data quality, governance, and real-world adoption of AI, including participation by co-founder and CSO Martin Willemink in a foundation-model data session at the 2026 IEEE International Symposium on Biomedical Imaging. The company’s messaging stressed time-to-treatment and workflow impact as key metrics for clinical AI success, beyond headline accuracy figures.

Although the week did not bring announcements of new financings, specific product launches, or explicit revenue milestones, Segmed’s activities advanced its ecosystem positioning in Japan and within Verily’s precision health platforms. Overall, the company continued to strengthen its profile as a key enabler of data-centric, multimodal medical imaging AI with growing relevance for oncology, musculoskeletal research, and global life sciences R&D.