uSINE-PAMS Artificial Intelligence Driven, Ultrasound-Guided Lumbar Puncture to Improve Procedural Accuracy

Lumbar puncture (LP) is a routine invasive procedure performed for the diagnosis and treatment of central nervous system disorders. The traditional landmark-based method is associated with a high failure rate of up to 50%, leading to complications including back pain and epidural hematoma, diagnostic delay, and increased healthcare costs. The success rate has been improved with ultrasound guided LP but the use of ultrasound-guided LP has not been widespread due to inadequate training in ultrasound use and the inability to translate information derived from ultrasonography to the LP procedure. uSINE-PAMS was designed to address these challenges: uSINE is a machine-learning software to aid the operators in ultrasound-guided LP; PAMS is a two-part hardware to translate data from ultrasound to accurate needle insertion and angulation.

This study consists of a pilot phase to test the usability of uSINE-PAMs and allow for improvement in technology and clinical workflow for uSINE-PAMS. In the pilot phase, experienced NNI clinicians who have participated in a previous healthy volunteer study and have been trained to use uSINE will be the operators performing the uSINE-PAMS-guided lumbar puncture. During and upon completion of the study, challenges in the use of uSINE and PAMS to conduct LP as well as in the implementation of uSINE-PAMS in the clinical workflow will be determined and addressed prior to the main study, open-label single-arm phase 2 clinical trial. Ten patients will be recruited in this pilot phase.

Following the pilot phase, an open-label single-arm phase 2 clinical trial to test the effectiveness of uSINE-PAMS-guided LP against historical data of 55% from traditional landmark-based method. 50 patients who are planned for LP will be recruited to undergo LP using the uSINE-PAMS-guided technique. The overall hypothesis is that uSINE-PAMS-guided LP will achieve at least 20% improvement in first pass success rate of obtaining cerebrospinal fluid compared to historical data of 55% from traditional landmark-based technique, therefore leading to reduction in complication rates.