We propose to extend the recently developed Medical Imaging Informatics Bench to Bedside (mi2b2) Workbench software to make medical images collected during routine clinical care available to clinical translational investigators. Today, this gold mine of patient information is not easily available. Well characterized medical images from clinical repositories would be an extremely valuable resource for clinical translational investigators who have specific ideas that require testing in large sets of images from many patients with different diseases. Our hospitals are international leaders in the development and deployment of advanced biomedical imaging technologies (MRI, high speed CT, ultrasound, PET and others) for clinical practice. The Picture Archive and Communication Systems (PACS) within each of the Departments of Radiology in our participating hospitals (Massachusetts General Hospital, Brigham and Women's Hospital, and Children's Hospital Boston) contain a wealth of medical images that often equal the quality of available clinical research imaging data, and greatly exceed its volume in terms of the number of patients and disease types. Though immediately extensible to other research applications and diagnostic scenarios, we apply our approach first to the difficult task of interpreting pediatric brain magnetc resonance images (MRI). The proposed Pediatric MRI module and Harvard Catalyst Radiological Decision Support toolkit that host it are targeted to address the immediate clinical challenge of understanding """"""""normal"""""""" because normal is a constantly moving target, changing rapidly with brain development. These free and open source software tools are natural extensions of currently funded government projects. This includes software from two National Centers for Biomedical Computing;Informatics for Integrating Biology &the Bedside (i2b2) which provides tools that extract and integrate data in a secure, HIPAA compliant manner from electronic medical records, laboratory data, billing information systems and the National Alliance for Medical Image Computing (NAMIC), which provides tools for medical image analysis including registration, visualization and computation of individual subject and large cohort data sets and will form the foundation for the Harvard Catalyst Radiological Decision Support Toolkit.

Public Health Relevance

The software tools developed by this proposal will enable monumental steps forward by making available to research scientists expensive medical images collected during routine clinical care and image analysis tools customized for large scale cohorts along with a wealth of data about associated patient illness that can be used for studies of nearly all medical diseases. To provide an explicit example of the profound impact this software can have to transform healthcare delivery we will develop and deploy the Harvard Catalyst RDS toolkit with a Pediatric MRI module based on neurodevelopmental MRI atlases for structural and diffusion weighted imaging data. These software tools will improve clinical management of babies with acute and chronic brain disorders. This grant provides the way for tens of thousands of images to be used for research with careful protection of patient privacy while working harmoniously within the culture of most radiology departments that are oriented around managing the images strictly for clinical care.

National Institute of Health (NIH)
National Institute of Biomedical Imaging and Bioengineering (NIBIB)
Research Project (R01)
Project #
Application #
Study Section
Biomedical Computing and Health Informatics Study Section (BCHI)
Program Officer
Pai, Vinay Manjunath
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Massachusetts General Hospital
United States
Zip Code
Yu, Zhi; Kim, Seoyoung C; Vanni, Kathleen et al. (2018) Association between inflammation and systolic blood pressure in RA compared to patients without RA. Arthritis Res Ther 20:107
Ou, Yangming; Zöllei, Lilla; Retzepi, Kallirroi et al. (2017) Using clinically acquired MRI to construct age-specific ADC atlases: Quantifying spatiotemporal ADC changes from birth to 6-year old. Hum Brain Mapp 38:3052-3068
Bernal-Rusiel, Jorge L; Rannou, Nicolas; Gollub, Randy L et al. (2017) Reusable Client-Side JavaScript Modules for Immersive Web-Based Real-Time Collaborative Neuroimage Visualization. Front Neuroinform 11:32
Ou, Yangming; Gollub, Randy L; Retzepi, Kallirroi et al. (2015) Brain extraction in pediatric ADC maps, toward characterizing neuro-development in multi-platform and multi-institution clinical images. Neuroimage 122:246-61
de Macedo Rodrigues, Katyucia; Ben-Avi, Emma; Sliva, Danielle D et al. (2015) A FreeSurfer-compliant consistent manual segmentation of infant brains spanning the 0-2 year age range. Front Hum Neurosci 9:21
Murphy, Shawn N; Herrick, Christopher; Wang, Yanbing et al. (2015) High throughput tools to access images from clinical archives for research. J Digit Imaging 28:194-204