The following seven projects were selected by the expert review panel for their innovative approaches to advancing health equity. 

AFRICA

Project: AI-assist for MRI safety and image quality: A study to develop AI tools to enhance patient safety, workflow compliance, and captured image quality for MRI in resource-limited settings.

Project team: 

• Abdulrazaq Zubair (Lead), Federal University of Health Sciences, Azare, Medserve Kano Diagnostic Center/AKTH, 
• Nafiu Musa Muhammad, B.Rad Medserve Kano/Diagnostic center/AKTH
• Mubaraq Yakubu, MSc King’s College London/AKTH
• Abbas Rabiu Muhammad MBBS, MSc Bayero University/Aminu Kano Teaching Hospital
• Mohammed Abba, PhD Bayero University/Aminu Kano Teaching Hospital
• Abba Muhammad Baze University Abuja, Nigeria
• Zulyadaini Muhammad Aminu Omdena Kano State AI for Smart Farming Project
• Yunusa Mohammed Garba, PhD Gombe State University, Nigeria
• Charles Delahunt, PhD University of Washington, USA
• Tarisiro Matiza, Mphil Market Access Africa

Location: Nigeria

Executive Summary:

Magnetic Resonance Imaging (MRI) is indispensable in modern diagnostics, and expanding the effective use of MRI in Africa will yield both improvements to individual care and the means to create Africa-centric MRI datasets necessary to develop AI-assisted diagnostics suitable for African populations. But patient-safety incidents and poor image quality remain persistent challenges in much of Africa. Many MRI centres operate under heavy workloads, with limited staff training, outdated systems, and paper-based screening forms, leading to incomplete safety checks, inconsistent image quality, and a lack of structured patient metadata. These gaps hinder both patient care and the creation of usable Africa-centric MRI datasets. 

Therefore, this project proposes an AI-assisted decision-support system at the pre-acquisition stage, to enhance MRI safety compliance, image quality, and complete metadata collection in resource-limited settings through three integrated modules: 

1. Automated Pre-MRI Safety Screening, which standardizes safety protocols and assists the MR radiographers/technologists in verifying implant safety and contraindications. 
2. Image Quality Optimization, a recommendation engine suggesting patient-specific acquisition parameters to minimize artefacts and repeat scans; and 
3. Digital Data Recording, which replaces paper forms with a structured electronic system for reliable documentation and metadata capture. A pilot study currently underway in Nigerian imaging centres, assessing MRI safety knowledge and compliance, has revealed wide variations in practice and reliance on paper-based methods, underscoring the need for standardized, data-driven safety procedures.

Building on this preliminary evidence, this project aims to develop and test AI tools to improve safety-screening compliance, reduce repeat scans, enhance captured image quality, and ensure complete patient metadata collection. The project is clinician-led, and includes both clinical and AI teams, to ensure that the AI development is informed by and serves clinical needs. The project will establish the first structured MRI safety database in Nigeria, and will develop AI tools to advance the ability of hospitals and clinics in low-resource countries to generate MRI datasets of high quality for future use in AI research.

ASIA

Project: ASEAN-HeartTwin: Validating Cardiac Digital Twin Models for Sudden Cardiac Death Risk Prediction of Hypertrophic Cardiomyopathy Across Diverse Healthcare Systems

Project team: 

• Lei Li, (lead) National University of Singapore
• Ching Hui Sia, National University Heart Centre Singapore
• Pham Hieu, College of Engineering and Computer Science, VinUniversity 
• Marni Azira Markom, Universiti Malaysia Perlis

Location: Singapore, Vietnam, Malaysia

Executive Summary

Cardiovascular disease remains the leading cause of mortality across ASEAN countries, where access to advanced diagnostic imaging and personalized risk prediction tools remains highly unequal. Sudden cardiac death (SCD) accounts for a significant portion of this burden, yet risk stratification methods are often based on data and models developed in Western populations, with limited validation in Southeast Asia. 

The ASEAN-HeartTwin project aims to establish and validate a cardiac digital twin (CDT) framework for SCD risk prediction across Singapore, Vietnam, and Malaysia, using multi-source clinical and imaging data to ensure cross-regional generalizability and fairness. We plan to use hypertrophic cardiomyopathy (HCM) as a typical example to predict its SCD risk. By integrating patient-specific anatomical and electrophysiological characteristics into a validated digital twin model, this project will lay the foundation for equitable cardiac risk prediction of HCM in ASEAN countries. The proposed work aligns with MICCAI’s focus on geographical health equity, by addressing regional disparities in healthcare technology and demonstrating scalable, data-driven solutions that are both clinically relevant and resource-sensitive. 

EUROPE

Project: MICCAI 2026 Open Data Initiative: Enabling Equitable Access and Global Participation in Medical Imaging Research

Project team: 

• Martijn P.A. Starmans (lead), Erasmus University Medical Center
• Apostolia Tsirikoglou, Karolinska Institutet
• Lidia Garrucho Moras, University of Barcelona
• Kaouther Mouheb, Erasmus University Medical Center

Location: Europe, Global

Executive Summary:

The MICCAI 2026 Open Data Initiative seeks to advance health equity in medical imaging by addressing global disparities in data availability and participation. Most public medical imaging datasets originate from high-income countries, limiting the diversity and generalizability of AI models. Building on two successful MICCAI Open Data Sessions, this project will provide targeted financial and logistical support to researchers, particularly from low- and middle-income countries, to help them share datasets addressing underrepresented diseases and populations. Support will cover key barriers such as limited infrastructure, legal or ethical constraints, and lack of expertise. The initiative will culminate in the 3rd MICCAI Open Data Session, recognizing exemplary datasets and fostering community exchange. To ensure long-term sustainability, we will establish a MICCAI Open Data Special Interest Group (SIG) dedicated to open and inclusive research practices. Together, these efforts will strengthen global collaboration, increase data diversity, and promote more equitable AI-driven healthcare innovation.

NORTH AMERICA

Project: Enhancing Translational Neuronavigation Capacity in Honduras

Project team: 

• Ena Isabel Miller (Co-lead) Hospital Mario Catarino Rivas and Universidad Nacional 
• Gabor Fichtinger (Co-lead), Queen’s University
• Víctor Berríos, MD, Hospital Mario Catarino Rivas
• Norman Cubilla, MSc, Universidad Nacional Autónoma de Honduras
• Alexandra Golby, MD, Harvard University / Brigham and Women's Hospital
• Sonia Pujol, PhD, Harvard University / Brigham and Women's Hospital
• Ron Kikinis, MD, Harvard University / Brigham and Women's Hospital
• Carla Chong Lara, MD, Hospital Mario Catarino Rivas
• Francisco Amador, MD, Hospital Hondureño de Seguridad Social
• Eduardo Nassar, MD Hospital Escuela Universitario
• Oscar Panameno, MD, Hospital Escuela Universitario
• Fabiola Cortez, MD, Hospital del Caribe, Puerto Cortés

Location: Honduras

Executive Summary:

The project aims to build sustainable translational neuronavigation research and training capacity in Honduras. Our initiative responds to severe and persisting disparities in neurosurgical care across Mesoamerica, where advanced technologies, such as computer-assisted neuronavigation systems, are limited to private hospitals due to exorbitant costs. In Honduras, public hospitals lack such technology despite having experienced neurosurgeons who previously have used commercial neuronavigation systems that became unsustainably expensive to maintain. The country’s centralized healthcare system and cohesive neurosurgical community create favorable conditions for reintroducing a locally maintainable neuronavigation technology.

The project leverages NousNav (www.nousnav.org), a low-cost open source neuronavigation platform offering over 95% cost reduction compared to commercial alternatives while maintaining clinically appropriate spatial accuracy. To support hands-on practical experience and training, the project integrates NousNav with a novel hybrid simulation environment using reusable manikin-based phantoms and public de-identified MRI data. Through diffeomorphic registration, MRI volumes are morphed to the manikin’s 3D mesh to produce anatomically realistic, patient-specific models. This setup will enable complete workflow practice, from segmentation and planning to targeting, without requiring imaging scanners or expensive hardware or elaborate fabrication processes. Tests on twelve MRI datasets yielded submillimeter accuracy and rapid case generation, validating its technical feasibility.

The one-year project will proceed in four phases: (1) building local technical competence in NousNav installation and maintenance; (2) developing a modular training curriculum; (3) delivering and evaluating training for neurosurgeons and residents using structured performance and usability assessments; and (4) consolidating sustainability, preparing publications, and disseminating results through open platforms.

By combining expertise in neurosurgery, computing, engineering, training, education, and open-source software, this project addresses critical barriers to neurosurgical capacity building in low-resource settings. Drawing on the successful practices and experience of the Train the Trainers program supporting the establishment of locally sustainable centers of translational research into ultrasound-guided interventions in West Africa, we aim to foster the development of a self-sustaining community of neurosurgeons in Honduras skilled in neuronavigation while establishing a reproducible model for capacity-driven, locally led neurosurgical innovation.

OCEANIA

Project: Automated Anomaly Detection using Ultra-Low-Field MRI for Remote Patients

Project team: 

• Zhaolin Chen (lead), Monash University, Australia,
• Kh Tohidul Islam, Monash University,
• Karen Caeyenberghs, Deakin University, Australia
• Meng Law, Alfred Hospital, Australia
• Jianfei Cai, Monash University, Australia

Location: Oceania

Executive Summary

The World Health Organization reports that over two-thirds of the global population lacks access to even basic medical imaging services, and many remote and indigenous Australian communities face inequality in accessing medical imaging. There is a pressing need to improve the affordability and accessibility to current and future medical imaging analysis software, especially for socio-economically disadvantaged patients. Many of these patients live remotely and lack adequate access to medical imaging, resulting in delays in treatments. 

Ultra-low-field (ULF) MRI (64 mT) provides a portable, low-cost neuroimaging solution critical for improving diagnostic access in resource-limited and remote regions. However, the inherent low Signal-to-Noise Ratio (SNR) and lack of large, labelled pathology datasets impede the development of robust, automated AI screening tools for clinical utilisation of these systems.

This project addresses this barrier by developing an automated anomaly detection toolkit for the ULF MRI. We will leverage datasets acquired from existing funded projects including a 64 mT and 3 T paired dataset with both healthy and disease to develop an anomaly detection tool, termed Normative and UnsupervisedAnomaly Detection (NUAD) to identify both healthy and abnormal brains, facilitating timely diagnosis and triage of patients. The seed funding will deliver a clinically interpretable, open-source UAD pipeline compatible with portable ULF scanners, providing an immediate, scalable mechanism to enhance timely diagnostics for brain injuries and other acute neurological conditions in rural and remote hospitals across Oceania, fulfilling the mandate for geographical health equity.

SOUTH AMERICA

Project Title: MSI-based Skin Analysis for All in Brazil

Project team: 

• Leticia Rittner (lead), School of Electrical and Computer Engineering (FEEC), UNICAMP
• Simone Appenzeller, MD/PhD: School of Medical Sciences (FCM), UNICAMP
• Bruna Alice Gomes de Melo, PhD: Center for Biomedical Engineering (CEB), UNICAMP
• Ana Clara Caznok Silveira, MSc: School of Electrical and Computer Engineering (FEEC), UNICAMP
• Nathan Shen Baldon, B.Eng.: School of Electrical and Computer Engineering (FEEC), UNICAMP

Location: Brazil

Executive Summary

Dermatological diagnosis in patients with darker skin tones remains a persistent challenge, rooted in both clinical and systemic inequalities. Conditions such as melanoma, diabetic foot ulcers and lupus exemplify how the progress of critical lesions often goes unnoticed in darker skin. These diagnostic inequalities highlight the urgent need for objective and skin-tone-independent evaluation methods, raising questions of whether we should look beyond the visible spectrum in skin assessment. Spectral imaging does exactly that, but, despite its proven diagnostic potential, it remains largely inaccessible in low- and middle-income countries. Specifically in Brazil, a country with a diverse population that could benefit from a technology like this, acquiring and maintaining this equipment would be financially unfeasible to the Unified Health System (SUS) — the largest public health system in the world. To address this gap, focusing on future implementation on SUS, our project seeks to develop a universal multispectral imaging (MSI) platform for clinical dermatology, by integrating low-cost locally manufactured hardware with artificial intelligence-driven analytical software, and performing inclusive pre-clinical and clinical validation. 

To implement this, we foresee nine milestones to be achieved in the next years: 

• low-cost spectral imaging hardware; 
• pre-clinical validation with skin phantoms; 
• open-access local dataset; 
• artificial intelligence algorithm for assisted skin assessment; 
• comparison of our method with other commercial options; 
• clinical validation; intellectual property; 
• regulatory plans; implementation on SUS. 

For Year 1, the first milestone (low-cost spectral imaging hardware) will be targeted, with activities divided in three phases: preparation, hardware development, and documentation and reports. With this is mind, we will need materials for both manufacturing, characterization, and calibration of the low-cost hardware, with a budget of $9869,67. 

We hope that this project will promote health equity both in the near and distant future, bridging long-standing technological and social gaps in dermatological diagnosis. In the short term, this project will provide a practical response to racial bias in dermatology, with the development of low-cost multispectral imaging hardware and personalized software that objectively analyzes skin lesions across different skin tones. In a medium to long term, this work will also be a call for change in how medical technologies are conceived and developed, helping to set a path for inclusive biomedical innovation. 

Furthermore, by releasing open-access datasets, in vitro skin phantoms, and software frameworks, we aim to encourage researchers to build upon, adapt, and expand this technology for their own local contexts. Specifically in Brazil, we believe that custom-made hardware and software will better capture the diversity of skin presentations found in our population, ensuring fairer and more accurate diagnosis. And by aiming this technology for implementation on SUS, which serves more than 200 million people, this project will strengthen early detection for all and improve patient outcomes while reducing systemic disparities in access to healthcare.

CLIMATE AND ENVIRONMENTAL HEALTH

Project: LEINA: Low-Compute Vision–Language Model for Equitable Tuberculosis Detection in Indonesian Chest X-Rays

Project team: 

• Vanya Valindria (lead), Monash University Indonesia
• Reyhan Eddy Yunus, RSCM Hospital / Universitas Indonesia
• Grace Wange, Monash University Indonesia

Location: Indonesia

Executive Summary

Tuberculosis (TB) remains one of Indonesia’s most urgent health challenges, with over 800,000 new cases annually. Diagnostic inequity persists across Indonesia’s primary healthcare network (Puskesmas community health centers) where radiologists and modern imaging infrastructure are limited. Climate-related factors—air pollution, urban overcrowding, and environmental stress - further exacerbate respiratory disease severity and significantly alter chest imaging appearances.

Most existing AI models for chest X-ray–based TB detection are developed using clean datasets from high-income countries. These models often underperform when applied to Indonesian patients due to environmental, demographic, and imaging-quality differences. Despite a population exceeding 270 million, Indonesia has no widely deployed, locally trained AI models for chest X-ray screening - creating a substantial equity gap.

LEINA aims to close this gap through a low-compute, data-efficient Vision–Language Model (VLM) trained using paired Indonesian chest X-rays and radiology reports. Designed for climate-impacted and resource-constrained healthcare environments, LEINA will:

• Leverage real-world data reflecting pollution-related variations in lung imaging.
• Align visual features with clinical text for better interpretability and generalizability; and
• Enable deployment without GPU requirements, as a lightweight web-based screening tool suitable for Puskesmas and district hospitals.

Expected Outcomes

• Curated Indonesian chest X-ray dataset with paired annotations / radiology report
• LEINA model performance benchmarked against global datasets to quantify geographic bias
• Prototype web-based screening application enabling equitable TB triage in low-resource settings, ready to integrate with Indonesia’s national TB initiatives

As one of the first MICCAI-aligned medical imaging AI efforts from Indonesia, LEINA will advance climate-conscious, locally relevant, and affordable AI innovation - supporting health equity across Indonesia and the broader Southeast Asian region.