Reading medical ultrasound images is challenging, as technicians must convert 2D visuals into a mental 3D model of tissue. MIT researchers introduced a new method that enables users to view a 3D augmented-reality image of scanned objects. With a virtual-reality headset, users can see an accurate 3D digital representation, facilitating easier identification and analysis. This innovation could accelerate training for ultrasound technicians and healthcare professionals using ultrasound. It may also be applied in hospitals for tasks like guiding needle placement for biopsies.<\/p>\n
“For training, this could make ultrasound more intuitive and more understandable. On the clinical side, it could be less time-consuming, more accurate, and also give health care providers more peace of mind. They wouldn\u2019t have to wonder if they missed anything,” said Canan Dagdeviren, an MIT associate professor and lead author of the study. MIT graduate students Jason Hou and Shrihari Viswanath led the research, published in Nature Communications Engineering. Additional contributors include Bowen Wu and MIT Summer Research Program students Cinay Dilibal and Tanisha Shende.<\/p>\n
Ultrasound imaging operates by sending high-frequency sound waves into body tissues, which reflect back to an ultrasound transducer. This transducer translates the sound into electrical signals, creating a 2D image, from which technicians generate a 3D mental picture. “It’s a difficult skill to master, and there are long learning curves,” Hou commented. “The hardest thing is this mental tomography bottleneck where you\u2019re trained to reconstruct the 2D slices in your 3D mental space. That is a cognitive burden that can lead to inaccuracies in scanning.”<\/p>\n
To ease this cognitive demand, the MIT team combined 3D ultrasound imaging with augmented reality (AR). Although 3D ultrasound is sometimes used in fetal imaging and echocardiography, common systems are costly and limited in availability. The team utilized their recently developed real-time 3D system for breast-cancer detection. This system features an ultrasound probe, smaller than a deck of cards, using a chirped data acquisition system (cDAQ). The probe’s ultrasound array, arranged in a square, captures 3D images of underlying tissue. With fewer components than standard 3D systems, it is less power-intensive and cheaper to produce.<\/p>\n
The probe’s data is compressed and fed into a 3D graphics engine called Unreal Engine, which translates the ultrasound image voxel data into a direct 3D representation without loss. Users wearing AR\/VR headsets can view this 3D rendering superimposed over the object\u2019s actual location, similar to X-ray vision. By moving their perspective, users can view the object from different angles, aiding in identification.<\/p>\n
The researchers tested the AR-VIU (augmented real-time volumetric imaging in ultrasound) system with 18 participants, including nine ultrasound experts and nine novices. They performed identification tasks using four ultrasound technologies, comparing 2D and 3D images on screens and in augmented reality conditions. Participants identified objects in gelatin and marked locations on “tissue phantom” materials. The AR-VIU system notably enhanced object identification and location accuracy for all users, with novices performing almost as well as experts.<\/p>\n
“Overlaying images with the anatomy and providing 3D visual context makes ultrasound significantly easier for novices to understand,” Viswanath stated. After the tests, most novices expressed a preference for the AR-VIU approach, citing ease of use. “The 3D system imposes less brain drain, it\u2019s more intuitive, and it\u2019s easier to understand what is happening in the targeted region,” Dagdeviren added. While many experts favored traditional 2D imaging due to familiarity, they acknowledged potential benefits of AR-VIU for specific tasks, such as biopsy needle placement or heart wall movement visualization.<\/p>\n
The MIT team is now focused on enhancing the resolution of their imaging system and conducting further tests to validate the accuracy of the AR-VIU technology. The research received funding from the MIT Media Lab Consortium, the National Science Foundation, an MIT HEALS graduate fellowship, and an MIT-Tata graduate fellowship.<\/p>\n