SUMMARY
Steve G. Langer, Ph.D., has a long-standing interest in tools that allow research analytics to be brought to bear in clinical practice within and beyond radiology. He studies performance profiling, adherence to standards required in healthcare IT, and scalable architectures for workflow automation that allow higher throughput and accuracy. Thoroughput is the amount of material or items passing through a system or process. With the growth of artificial intelligence (AI) in medical imaging, Dr. Langer works on monitoring AI drift. AI drift occurs when the correctness of predictive results change. This is a problem that requires data to communicate among multiple systems using standard ontologies that promote data mining on a strong and performant workflow orchestration platform.
Focus areas
- Automated methods for medical imaging including quality control, processing and display.
- Automated image selection, preparation and per-processing to train deep learning models for both classification and diagnosis.
- AI-based workflow orchestration tools to assure that clinical studies are properly routed to the right AI models and results are available in clinically relevant time frames.
- Prevention of drift in AI models over time.
Significance to patient care
An aging population introduces an increasing number of people to the healthcare environment. This healthcare environment is challenged to handle that load while trying to limit costs with fewer personnel.
The technologies that Dr. Langer assists in developing promote efficiency, reliability, data mining ability and trust in AI results through cross-checking with other data sources and monitoring drift. Such abilities will be essential as AI is used in more places, often ahead of certainty that the results are always correct.
Professional highlights
- Chair, Mayo Clinical Digital Imaging Architecture Sub-Committee, Mayo Clinic, 2019-present.
- Society of Imaging Informatics in Medicine:
- Chair, Machine Learning Committee, 2016-present.
- Co-chair, Hackathon Committee, 2010-2016.
- Chair, Mayo Enterprise Radiology Architecture and Standards Committee, Mayo Clinic, 2012-2021.
- Author, Follow-Up of Non-Critical Actionable Findings (FUNC), an international standard for transmitting medical information, 2017.
- Author, Standardized Operational Log of Events (SOLE), an international standard for transmitting medical information, 2017.
- Co-author, Informatics in Medical Imaging, CRC Press, 2012.
- Co-architect, Radiological Society of North America Image Share Network, National Institutes of Health, 2009-2012.