Feasibility of AI-assisted Neuraxial-Specific Ultrasound Guidance for Bedside Lumbar Puncture
Matthew Rumschlag1, Kalina Misiolek2, Paul Sheeran3, Christopher Thom4, Karen Johnston4
1Trinity Health IHA Medical Group, 2Cleveland Clinic Foundation, 3Rivanna Medical, 4University of Virginia
Objective:
In this exploratory pilot study, we assessed the feasibility of using Accuro 3S, a new neuraxial-specific, AI-assisted ultrasound probe, to improve accuracy of needle placement and user confidence when preparing for bedside lumbar puncture (LP).
Background:
LP is an essential diagnostic technique; however, failures commonly lead to delayed diagnosis, prolonged hospitalizations, and repeat procedures. Specialty ultrasound guidance during LP is a potential way to mitigate the risk of failure.
Design/Methods:
Using randomized, paired trial design, four residents performed simulated LPs (no needle puncture) on 12 participants using either landmark (standard) technique or Accuro 3S guidance. An ultrasound-fellowship trained emergency medicine physician performed blinded outcome assessment based on location and trajectory of a simulated needle puncture. The primary outcome was likelihood of LP accuracy on a 4-point scale (4 being "highly likely to succeed on first pass"), and secondary outcome was resident confidence on a 5-point Likert scale (5 being "high confidence"). Results were analyzed using Wilcoxon signed-rank test with paired ordinal data. Secondary analyses included paired t-test to compare distance measurements from L3-L4 space between the two approaches and ground truth as well as paired t-test to compare differences in angle measurements.
Results:
Mean likelihood-of-success rating was significantly different between the two groups with 3.15 (SD 0.54) using Accuro 3S and 2.02 (SD 0.73) in the landmark group (p<0.0001). Confidence levels were also significantly different at 3.42 (SD 1.08) with Accuro 3S and 2.77 (SD 1.1) in landmark group (p=0.001). The vertical angle difference between each group and ground truth was also significantly different with the device group being closer to ground truth than landmark group (p=0.004).
Conclusions:
Overall, we demonstrated early feasibility, and our results suggest further studies are warranted to robustly assess whether use of neuraxial specific ultrasound devices has the potential to increase the rate of success of bedside LP.
10.1212/WNL.0000000000215043
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.