Published in Journal of Ultrasound in Medicine, 2020
Handheld ultrasound devices (HUDs) have previously been limited to grayscale imaging without options for left ventricle (LV) quantification. We aimed to study the feasibility and reliability of automatic measurements of mitral annular plane systolic excursion (MAPSE) by HUDs.
Published in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, 2019
The processing power in commercially available hand-held devices has improved dramatically in recent years. In parallel, techniques used in high-frame-rate medical ultrasound imaging, especially plane-wave (PW) imaging, have reduced the number of ultrasound transmissions and amount of data necessary to reconstruct an ultrasound image. In combination, the processing power and data reduction allow all of the processing steps in ultrasound image formation, from raw ultrasound channel data to final rendering, to be performed on a hand-held device. In this study, we send the raw ultrasound channel data from a research scanner wirelessly to an off-the-shelf hand-held device. The hand-held unit’s graphical processing unit is processing the raw ultrasound data into the final image, achieving real-time frame rates on the order of 60–90 frames per second (FPS) for a single-angle PW transmission. Higher quality images are achieved by trading off frame rate by coherently compounding multiple PW images, resulting in frame rates on the order of, e.g., 13 FPS when coherently compounding 7 PW transmissions. The presented setup has the potential of providing image quality which could be valuable for simple medical ultrasound diagnostic scans of, e.g., the carotid artery or thyroid. Also, since the computationally expensive beamforming can be done in off-the-shelf devices, this could reduce the price of hand-held ultrasound systems in the future.
Published in European Heart Journal, 2018
Mitral annular systolic plane excursion (MAPSE) is an easy and reliable measure of global longitudinal left ventricular (LV) function. We have developed an algorithm that automatically measures MAPSE from live grey scale recordings that can be implemented on pocket-size imaging devices (PSID). Automatic measurements and interpretations of findings can assist inexperienced users when evaluating LV function. This has never been evaluated on PSID, which so far have not been able to provide quantitative assessment of LV function.
Published in Mixed and Augmented Reality in Medicine, 2018
Echocardiography has become a widely used imaging technique for the diagnosis and follow-up of heart disease. However, the acquisition of high-quality echocardiographic images requires extensive training, which limits the adoption of the technique by nonexpert users. This chapter describes the main elements required to build an AR system that can be used for teaching purposes, which mimics the conditions encountered in the echocardiographic laboratory. Then, we describe a specific implementation of an AR system that combines a highly descriptive geometric model of the heart with US data that is acquired in real time and with a video feed containing the subject to be imaged. It will be shown that a low-cost AR system can be built and implemented on an off-the-shelf tablet device. The usability of the system has been tested on a group of medical students who had limited knowledge of US, and early results are shown.
Published in 2016 IEEE International Ultrasonics Symposium (IUS), 2016
This work describes the implementation of software plane-wave beamforming performed on the GPU of an Apple iPhone 6s and 6s Plus. The code can run on any current iOS device that supports the Metal API. The implementation is largely written in Swift, with some Objective-C, while the core processing component was written in Metal, Apple’s new GPU programming language which provides low-overhead compute shaders for exploiting the device’s GPU. We have demonstrated that ultrasound channel data recorded on a Verasonics Vantage system can be wirelessly transmitted to the iPhone using a simple networking implementation obtaining a frame rate up to 5 FPS including serialization and transmission, and easily 60+ FPS for on device processing depending on number of samples and output image size.
Published in 2015 IEEE International Ultrasonics Symposium (IUS), 2015
Ultrasound imaging is a highly valuable diagnostic tool. It is increasingly portable, provides real-time imaging of complex structures, and is considered safe. Yet, because ultrasound is a highly operator dependent modality the uptake of ultrasound within a broader range of medical contexts has been limited and hasn’t made major inroads within the offices of General Practitioners, Midwives, and other non-specialists. Learning to effectively use ultrasound can easily take up to 12 months with direct expert supervision.