UCLA Researchers in the Department of Radiology have developed an automated computer program that is capable of identifying implanted medical devices and safe zones for their placement: potentially offering a solution to the high volume of imaging exams performed daily to check placement and avoiding fatalities arising from misplaced devices that are detected too late.
BACKGROUND:
Chest radiography is not uncommon to represent the majority of cases encountered by a radiologist in a hospital. Due to an increasing number of intensive care units and the increased ability to treat those that are gravely ill with modern medicine, chest radiography is often standard practice prior to undertaking treatments and checking the placement of tubes and lines. Failure to detect misplacement of medical devices and immediately correct them could lead to life threatening complications. Therefore, radiologists are indispensable during medical implementations. Unfortunately, many hospitals have a limited number of radiologists, and trained eyes still require significant time frames to accurately identify medical devices and safe zones surrounding them. Currently, this issue creates major resourcing limitations and increases the risk of complications from medical procedures. Therefore, there exists a major unmet need in the ability to increase the throughput of reading and providing safe zones of medical treatments in patient chest radiography.
INNOVATION:
UCLA Researchers in the Department of Radiology have developed an automated computer program that is capable of identifying medical devices and the safe zones of medical implementation surrounding them: potentially offering a solution to the limiting number of chest surgeries that can be performed daily. The software is an automated overlay of a safe zone on chest radiography, that is capable of issuing an alert if the implemented device falls outside the safe zone. Additionally, the software can be used to identify the zone of safety that medical devices can be placed upstream of the surgical process. This automated platform could allow the ability to address the issue of throughput in chest radiography and increase access of patient care across the increasing number of intensive care units around the world.
POTENTIAL APPLICATIONS:
• The upstream identification of margins of safety zones for medical device placement
• The downstream identification of correct placement of medical devices, lowering the incidences of associated mortality
ADVANTAGES:
• Direct interface with chest radiography images
• Ease of identification of the safety zone for surgeons
• Currently the only system that has a direct output interface to chest radiography images
DEVELOPMENT-TO-DATE:
A fully developed and working prototype has been developed. The software is continually being refined for increased access to surgeons
RELATED PAPERS:
Wang, X., Teng, P., Ontiveros, A., Goldin, J. G. & Brown, M. S. High throughput image labeling on chest computed tomography by deep learning. JMI 7, 024501 (2020).