USAISR and MIT Collaborate on Advanced AI Tool for Nerve Block Anesthetics

FORT DETRICK, Md. – The U.S. Army Institute of Surgical Research Organ Support and Automation Technologies team (USAISR) is collaborating with the Massachusetts Institute of Technology’s Lincoln Laboratory to create a new device designed to provide regional pain control for trauma patients right at or near the site of injury.

This innovative device utilizes advanced artificial intelligence (AI) technologies to detect and map the locations of nerve bundles in a patient's body. It then guides a needle to the precise spot where pain control medications need to be delivered. To enhance safety, the device is equipped with a feature that prevents needle insertion until it confirms the correct anatomical target is located. This makes the device user-friendly for medics, even those with minimal training in regional anesthesia, making it useful during mass casualty situations or environments where specialists may not be available.

Clinical studies have indicated that ultrasound-guided femoral nerve blocks are more effective for pain control than traditional intravenous or intramuscular opioids. However, performing this technique requires significant skill to ensure accurate needle placement near the targeted nerve. By integrating AI and robotics, the team aspires to eliminate this barrier to timely and effective care for combat casualties. The partnership between engineers and military medical professionals that has driven the development of this device may pave the way for more effective implementations of clinical AI applications in military settings.

Background of the Nerve Block Device

While administering systemic pain medications such as morphine or ketamine can effectively treat pain, these drugs can also cause drowsiness or unconsciousness, hindering the patient's ability to communicate, defend themselves, or evacuate the battlefield. Additionally, systemic medications may lower respiratory rates, increasing the risk of cardiac arrest, and they require continuous monitoring. In an emergency situation, patients needing evacuation would rely on medics, taking critical personnel away from immediate combat duties.

Lt. Col. Brian Kirkwood, a dentist and chief AI officer of the USAISR team, known as CRT3, shared that the idea for the ultrasound nerve block device emerged from a discussion led by Maj. Gen. Michael J. Talley during a command-wide meeting in October 2020. Talley had mobilized MRDC's labs to address the pandemic and encouraged attendees to envision the future of military medical technology. "Think Stalingrad meets Star Wars," Kirkwood remembers Talley saying.

After the meeting, Kirkwood expressed to Dr. Jose Salinas, the project's science lead, a desire for technology that could assist non-specialist medics in providing regional anesthesia in mass casualty events. This conversation laid the groundwork for the development of the device.

To build a prototype that would facilitate precise nerve block administration by non-experts, USAISR collaborated with the Metis Foundation and MIT’s Lincoln Laboratory. This lab had already produced a handheld ultrasound device called AI-GUIDE, which assists experts in placing femoral vascular catheters. This technology formed the foundation for the new device. Support was obtained through a funding solicitation from the Medical Technology Enterprise Consortium, an organization encompassing over 600 academic institutions, businesses, and nonprofits in the biomedical technology field, working in partnership with MRDC. Funding for this project is provided by the Combat Casualty Care Research Program.

To understand how anesthesiologists administer regional nerve blocks, Kirkwood observed several procedures, and this experience guided the design of the device. He noted that the prototype required an innovative approach by inserting the needle in-plane, parallel to the ultrasound transducer, unlike the AI-GUIDE, which inserts the needle from an out-of-plane angle. This new method provides better visualization of the needle during placement and enhances safety.

The currently developed prototype features a handheld frame designed with 3D printing that holds an ultrasound transducer. The AI software employs segmentation and object detection techniques based on scans of swine lower-body neurovascular structures to identify anatomical landmarks. Once the target area is located, the guidance system, operated by a robotic arm, adjusts the needle's angle and depth for accurate insertion. The operator then presses a button to deploy the needle once safety checks are passed, after which anesthetic delivery can occur.

Testing and Future Improvements

Kirkwood’s team has conducted numerous tests on the prototype over the past year. At least ten operators, including healthcare providers and engineers, evaluated the device for placing a needle for a femoral nerve block. Each test involved scanning the area from the knee to the inguinal crease until the AI identified the target location, prompting operator adjustments until the needle could be deployed. Data on insertion time and accuracy were collected to refine the device.

Each round of testing provided insights into necessary improvements. After evaluations, USAISR communicated findings to the Lincoln Laboratory team for updates to the prototype. Eventually, the focus shifted to two proficient operators to enhance the device's performance.

"Our initial tests revealed that this technology can enable minimally experienced personnel to place a needle for regional anesthesia in under 40 seconds," stated Kirkwood. "We discovered that the AI system functions effectively under various conditions, including hypotensive situations, which is vital for battlefield scenarios. However, large-scale preclinical studies still need to be performed to validate the device's guiding systems and needle placement". The goal is to produce a self-contained device to improve portability for battlefield application.

As the prototype transitions from preclinical to clinical testing stages, there will be a crucial need to train the AI system based on human data. Comprehensive usability testing will also be needed to incorporate insights from clinical experts in order to fine-tune the device for its intended users. Kirkwood is actively seeking funding opportunities to support ongoing development and testing into 2025 and beyond.

Significance of Team Collaboration

This project exemplifies the successful collaboration between engineers and clinicians in creating an AI-based medical device intended for future military use. Such partnerships are a core mission of CRT3.

"One of CRT3's main objectives is to leverage advanced engineering technologies to address specific capability gaps, such as the need for effective pain management in military scenarios," Salinas explained. "We have formed multidisciplinary teams of engineers and medical professionals to find this solution. Achieving communication between these experts to generate practical outcomes is a significant accomplishment for us."

Kirkwood, who is pursuing a doctorate in translational science, along with Salinas, plans to present on the importance of interdisciplinary collaborations in developing AI for military clinical medicine during the 2025 AMSUS Annual Meeting. The session is titled "Bridging the Gap: Engineers and Clinicians Working Together to Advance Expeditionary Medical and Dental Applications of Artificial Intelligent Systems."

"Effective communication between clinicians and engineers is crucial to establish trust in the AI systems we are creating for real-world applications," Kirkwood emphasized.

(This narrative is an adaptation from the Winter 2024/2025 issue of Combat & Casualty Care.)

medical, technology, AI