Robot Uses Deep Learning, Ultrasound Imaging to Draw Blood

An autonomous, robotic tool used deep learning and ultrasound imaging to draw blood or insert catheters on par with or better than human providers, according to a study published in Nature Machine Intelligence.

Over 90 percent of diagnostic and therapeutic procedures in the ER, intensive care unit, catheterization lab, and operating room require gaining vascular access, researchers noted. Gaining access to vessels is a critical first step in drawing blood, administering fluids and medications, introducing endovascular devices, monitoring physiological status, and several other important procedures.

The timely delivery of these interventions can significantly affect morbidity and mortality and accessing vessels can be a challenging task. The research team stated that failures occur in an estimated 20 percent of procedures, and difficulties can be exacerbated in patients with small or collapsed vessels, which are common among pediatric, elderly, and chronically ill patients.

Using medical robots, providers could possibly reduce injuries and improve the efficiency and outcomes of procedures, researchers said. Additionally, robots could carry out tasks with minimal supervision when resources are limited, allowing clinicians to focus on more critical aspects of care delivery.

“Unlike imaging-based methods, which rely on manual insertion, robotic strategies could altogether eliminate the dependence on practitioner experience and availability,” researchers said.

“Here, we present a portable robotic device driven by deep learning capable of steering needles and catheters into submillimetre vessels with minimal supervision.”

The tool combined a deep learning framework with near-infrared and ultrasound imaging to perform complex visual tasks, such as identifying blood vessels from surrounding tissue, classifying blood vessels and estimating their depth, and motion tracking.

After testing the tool, the researchers found that the device reduced the average number of failed access attempts from 1.8 per trial to 0.3 per trial. The device also increased first-stick success rates from 53 percent to 88 percent compared to blind manual access, with the largest gains seen in the most complex physiological conditions.

“Using volunteers, models and animals, our team showed that the device can accurately pinpoint blood vessels, improving success rates and procedure times compared with expert healthcare professionals, especially with difficult to access blood vessels,” said senior author Martin L. Yarmush, Paul & Mary Monroe Chair & Distinguished Professor in the Department of Biomedical Engineering in the School of Engineering at Rutgers University – New Brunswick.

Researchers pointed out that the tool can benefit areas beyond clinical care delivery.

“Not only can the device be used for patients, but it can also be modified to draw blood in rodents, a procedure which is extremely important for drug testing in animals in the pharmaceutical and biotech industries,” Yarmush said.

Within clinical settings, the group said that the device has the potential to improve workflows and facilitate more efficient medical processes, as well as increase patient access to care.

“The robotic paradigm may be extended to address clinical challenges in minimally invasive endovascular workflows, where accurate cannulation of major vessels (such as the common femoral artery) is a prerequisite to surgical success and where repeat punctures increase the risk of arterial trauma and hemorrhage,” the team said.

“Outside the hospital, robotic technologies could allow emergency medical providers to obtain rapid vascular access under time-critical conditions and bring advanced interventional and resuscitation capabilities to remote and resource-limited environments.”

The research team plans to test the device on a broader range of people, including those with normal and difficult blood vessels to access.

“We posit that, by lowering the likelihood of failed attempts, robotic cannulation could prevent injuries from multiple sticks, reduce complication rates and minimize the need for central catheter placement following unsuccessful peripheral access,” researchers concluded.

“The present work provides motivation for further assessment of clinical benefits and risks across a broader demographic spectrum that includes both normal and difficult populations.”