Photo By Steven Galvan | U.S. Army Institute of Surgical Research teammates Lt. Col. Brian Kirkwood, left, a...... read more read more
FORT DETRICK, Md. – The U.S. Army Institute of Surgical Research Organ Support and Automation Technologies team is working with the Massachusetts Institute of Technology’s Lincoln Laboratory to develop a new device for providing regional pain control in trauma patients while they are still at or near the point of injury.
The device uses advanced artificial intelligence technology to recognize and map the location of regional nerve bundles on the patient’s body and automatically guide a needle to the correct location for delivery of pain control medications. A safety mechanism prevents needle insertion until the appropriate anatomical target has been identified. These features will enable the device to be used by medics with little or no expertise in providing regional anesthesia, making it ideal for use in mass casualty events and other scenarios where such experts will likely be in short supply.
Clinical studies have shown that ultrasound guided femoral nerve blocks provide greater pain reduction than intravenous or intramuscular opioids. However, the technique requires a high level of dexterity to ensure accurate needle placement near the target nerve bundle when delivering the anesthetic. By utilizing a combination of AI and robotics, the device’s creators hope to be able to overcome this limiting factor in providing effective and timely combat casualty care. Not only that, but the interdisciplinary collaboration between engineers and military clinicians that led to the development and testing of the device suggests a possible pathway for more streamlined implementations of clinical AI applications in the military.
Origins of the Nerve Block Device
While systemic pain control drugs such as morphine or ketamine are effective at treating pain, they also render the patient drowsy or unconscious, making them incapable of communicating effectively, continuing to fight, or leaving the battlefield under their own power. Systemic pain medications also potentially introduce physiological risks, such as a reduced respiratory rate that increases the susceptibility to cardiac arrest. Furthermore, patients treated by systemic pain control methods require constant monitoring by a medic, and should they require evacuation, will need to be transported by litter – all of which pulls additional personnel out of the fight.
Lt. Col. Brian Kirkwood, a comprehensive dentist and chief AI officer with the USAISR team, known as CRT3, said the inspiration for the ultrasound nerve block device was a talk by Maj. Gen. Michael J. Talley, then the commanding general of MRDC, during a command-wide town hall meeting in October 2020. Talley, who mobilized MRDC’s laboratories to respond to the coronavirus pandemic, encouraged attendees to think about the kinds of military medical technologies that would be needed on the battlefield of tomorrow. “Think Stalingrad meets Star Wars,” Kirkwood recalls Talley saying.
“After the town hall meeting, I was thinking about that question as I was walking back to the office with Dr. Jose Salinas, our science lead,” recalled Kirkwood, the project’s overall principal investigator. “I told him, ‘I’m a dentist. I know how to administer anesthesia to manage pain. If you were to give me some type of technology that would enable me to jump in and help provide regional anesthesia in an area outside the mouth in a mass casualty event beyond just doing triage, I’d be more than willing to help get a Soldier of pain.’ And that idea eventually evolved into this device.”
To develop a prototype device that would be capable of allowing a non-expert to deliver a regional nerve block with pinpoint precision, USAISR partnered with the Metis Foundation, a nonprofit research organization, and MIT’s Lincoln Laboratory, which specializes in research, development, and rapid prototyping of advanced technologies for national security applications. The lab had developed a handheld ultrasound device called AI-GUIDE, which allows specialists to accurately place femoral vascular catheters and guide wires, that could serve as the basis for the new device. The team applied for and received support through a funding solicitation by the Medical Technology Enterprise Consortium, a nonprofit international affiliation of over 600 academic institutions, businesses, nonprofits and other organizations in the biomedical technology sector that operates through a contractual agreement with MRDC. The funding support for this research and development effort is supported by the Combat Casualty Care Research Program.
To learn how regional nerve blocks are administered in a clinical setting, Kirkwood observed several anesthesiologists as they worked and used that knowledge to help inform the design of the device and its operational technique. For example, he noted that, unlike AI-GUIDE, in which the needle is inserted out-of-plane – that is, perpendicular to the cross-sectional image produced by the ultrasound transducer – the prototype was developed with the needle inserted in-plane, parallel with the transducer, to mimic how the anesthesiologists he observed performed regional nerve blocks. Inserting the needle in-plane provides an opportunity to observe the needle during insertion to aid in development and adds to the safety system designed into the software. This important distinction needed to be taken into account when designing both the hardware and software used in the new device.
The prototype device developed by CRT3 and Lincoln Laboratory consists of a 3D-printed handheld frame that holds an ultrasound transducer. The AI software utilizes both segmentation and object detection techniques were trained on nerve, artery, and vein landmarks from ultrasound scans of swine lower-body neurovascular bundles. In addition to detecting anatomy, the software also controls the device guidance system, which is a robotic arm that adjusts needle angulation and insertion. Once the target location has been identified by the AI software and safety checks are cleared, the operator is prompted to press the trigger button to deploy the needle. Once the needle is inserted at the proper location, the anesthetic can then be delivered.
Tests Validate Concept, Point to Improvements
Over the course of a year, Kirkwood’s team conducted multiple rounds of tests of the prototype. For the tests, at least ten different operators – including health care providers, engineers and technicians – attempted to use the device to place the needle for a femoral nerve block. On the ex-vivo model, each attempt began at the knee, scanning proximally toward the inguinal crease between the leg and groin until the AI recognized the target location. The device would prompt the operator’s directional movements until the target location was identified, at which point the trigger button would be activated to allow for needle deployment. Insertion time and needle placement location were collected after each attempt to continuously improve the device.
Each round of tests identified areas requiring further work. The team at USAISR communicated the necessary hardware and software improvements to the Lincoln Laboratory team after each round of testing so that it could update the prototype. Kirkwood and his team decided to reduce the number of operators to the two who were most proficient with the device, in order to focus on the device’s performance.
“From the initial testing, we have shown that this type of technology can enable a minimally experienced person to place a needle for regional anesthesia in under 40 seconds,” said Kirkwood. “During the pilot study, we were also able to show that the AI system works in both normal and hypotensive conditions, which is important because we can expect to see a wide range of damage to limbs on the battlefield. We still need to conduct larger preclinical studies in animal models to validate some of the systems that guide the device and needle placement. The ultimate goal is a device that is self-contained to improve portability for a frontline battlefield application.”
Furthermore, as the prototype advances from the preclinical phase to clinical testing, the AI system will need to be thoroughly trained on human data, said Kirkwood. Eventually, usability testing is necessary to capture the insights and experience of clinical experts to refine the device for the end-user. Kirkwood is currently exploring funding options to support continued development and testing of the device in 2025 and beyond.
Project Demonstrates Value of Interdisciplinary Collaboration
Another important outcome of this project is its demonstration of the value of collaboration between clinicians and engineers in successfully developing and demonstrating an AI-based medical device for the future battlefield. This accomplishment spotlights one of CRT3’s core competencies.
“One of the main missions of CRT3 is to use advanced engineering technologies to address documented capability gaps – in this case, the need to provide regional pain control on the battlefield,” said Salinas. “To that end, we have put together heterogeneous and multidisciplinary teams that specialize in applying engineering technologies to develop medical solutions. It’s not an easy solution, because there aren’t any degrees out there that will teach you how to do this. An engineer isn’t taught how to speak medical. Physicians aren’t taught how to speak engineering. Creating these teams and getting them to speak the same language to the point where they can actually generate solutions is something that we have been very successful at doing.”
Kirkwood – who is now pursuing a doctorate in translational science – and Salinas will be discussing the importance of interdisciplinary collaboration in developing AI systems to assist in clinical decision-making in military medicine at the 2025 AMSUS Annual Meeting in March 2025, in a breakout session titled “Bridging the Gap: Engineers and Clinicians Working Together to Advance Expeditionary Medical and Dental Applications of Artificial Intelligent Systems.”
“It’s important that we have good communication between clinicians and engineers to shape the development of AI systems so that the end user has trust and confidence in the systems that we’re developing,” said Kirkwood.
(NOTE: A version of this story was originally published in the Winter 2024/2025 issue of Combat & Casualty Care.)
| Date Taken: | 02.13.2025 |
| Date Posted: | 02.13.2025 09:27 |
| Story ID: | 490705 |
| Location: | FORT DETRICK, MARYLAND, US |
| Web Views: | 83 |
| Downloads: | 1 |
This work, USAISR, MIT Team Up on AI Tool for Nerve Block Anesthetics, by Paul Lagasse, identified by DVIDS, must comply with the restrictions shown on https://www.dvidshub.net/about/copyright.
No keywords found.
