Assistant Professor Luca Giancardo, PhD recently co-authored an article that Nature Journal published in its Scientific Report in late October. The article is titled “MRI Compatibility: Automatic Brain Shunt Valve Recognition using Feature Engineering and Deep Convolutional Neural Networks.”
“The process of identifying implantable devices in x-rays, and establishing their MRI safety profile, is a very time-consuming process that leads to suboptimal patient care,” noted Giancardo. “This technology has the potential to be applied to other medical devices, thereby enabling a fast and precise MRI clearance process for busy radiology departments. It can potentially decrease the number of patients that experience denials or delays for their MRI exams.”
Giancardo collaborated with four McGovern Medical School radiology researchers to write the article: Octavio Arevalo, MD, Andrea Tenreiro, MD, Roy Riascos-Castaneda, MD and Eliana Bonfante-Mejia, MD. The mutual collaboration between SBMI and McGovern Medical School demonstrates the importance of multidisciplinary innovation in the health science field.
“This research stems from a clinical need identified by the radiologists Dr. Bonfante-Mejia directs,” Giancardo continued. “I met Dr. Riascos-Castaneda and he is leading an effort to facilitate the integration of medical imaging and data science. Those focuses allowed us to form a multidisciplinary team and we were able to design an automated solution that has the potential to significantly improve this process.”
Currently, there are no immediate benefits for clinicians or patients but the research in the article proves the feasibility of the system. Once implemented in clinical practice, this system could be utilized in smaller centers where the expertise or time to identify implantable devices before MRI scans might not be available. “The algorithm proposed in our research can be used for other implanted devices as well, like cardiac and vascular devices such as stents, inferior vena cava filters, coils and clips,” stated Giancardo.
Scientific Reports is an open-access, multidisciplinary journal. To learn more about Giancardo and his partners’ research, please read the online.
published on 11/12/2018 at 11:30 a.m.