Hiro Yoshida, PhD, director of the 3D Imaging Research Group in the department, and team member Janne Näppi, PhD, received two honors at this year’s SPIE Medical Imaging meeting.
First, the team received the Cum Laude Award for their “Feasibility of the automated detection of colorectal polyps in photon-counting CT colonography” (Näppi JJ, Hironaka T, Wu D, Yoshida RS, Gupta R, Tachibana R, Taguchi K, and Yoshida H.). This award acknowledges the best scientific poster presented at the Imaging Informatics Conference at the SPIE Medical Imaging 2023.
Also, Yoshida has been elected to serve as the Chair of the Imaging Informatics for Healthcare, Research, and Applications Conference at SPIE Medical Imaging 2024, scheduled to be held in San Diego in February next year.
We checked in with Yoshida and Näppi both to hear more about these honors and to learn more about the work they do generally.
Can you describe the work you do in your group?
Our 3D Imaging Research group concentrates on the research and development of innovative AI technologies for three-dimensional medical imaging and their diagnostic applications. Our mission is to create AI-driven imaging technologies that enhance image-based diagnosis and integrate seamlessly into clinical diagnostic workflows. Established in 2005 at the MGH, our group has a foundation in our pioneering research on AI, machine learning, advanced image processing and visualization, and computer-aided detection (CADe), dating back to our work at the University of Chicago since 2000.
Can you describe the work presented in your poster “Feasibility of the automated detection of colorectal polyps in photon-counting CT colonography”?
Although CT colonography provides a nearly ideal screening strategy for addressing the issues and problems inherent with the other colorectal cancer screening tests, the opponents of CT colonography keep citing the ionizing radiation of the underlying CT technology as a concern. Photon-counting CT (PCCT) can be used to address that concern. Meanwhile, computer-aided detection (CADe) has been shown to be essential in implementing an effective laxative-free CT colonography examination, which is very important for improving patient adherence to colorectal screening guidelines. However, there have been no studies on the performance of AI in PCCT colonography. The most novel aspect of our approach was the integration of AI for automated polyp detection in laxative-free PCCT colonography. As far as we know, this was the first study of its kind, and our findings demonstrated the effectiveness of our AI methods on this problem.
What is the next step in this research?
The next step in our research is to extend our preliminary study, which was based on an anthropomorphic phantom, to clinical PCCT colonography cases. In addition to lower radiation, PCCT also has other advantages over conventional CT, such as higher spatial image resolution and native multi-energy imaging, which can be utilized to further enhance the performance of our AI methods. Moreover, we plan to investigate the application of our AI system to CADe to evaluate the improvements it brings to clinical practice.
Dr. Yoshida, what are you most looking forward to as Chair of the Imaging Informatics for Healthcare, Research, and Applications Conference at the SPIE Medical Imaging 2024?
As Chair of the Imaging Informatics conference, I am thrilled about the ongoing groundbreaking advancements in generative AI technologies, such as large language models like GPT-4 or innovative latent diffusion models like Stable Diffusion, as well as in their emerging applications in imaging informatics. Those game-changing models are poised to revolutionize the field of imaging informatics, as they have the potential to improve healthcare outcomes through more accurate and efficient management of the growing imaging data volumes and the development of personalized medical treatments.
In response to this trend, the Call for Papers emphasizes the applications of generative AI in managing, analyzing, and visualizing imaging informatics data across healthcare, research, and related domains. The conference will showcase cutting-edge research and clinical applications, offering a vibrant platform for experts to exchange groundbreaking ideas on enhancing big data analytics, predicting imaging outcomes, and performing diagnosis and clinical decision-making. The attendees can anticipate trailblazing presentations and discussions on this transformative technology and its applications in imaging informatics.