An Ultrasound Guided Automated Spinal Landmark Identification System

Neuraxial ultrasonography has been shown a safe and effective technique to enhance the overall success rate of lumbar puncture and reduce the number of injection attempts. However, the current blind palpation landmark technique is known to be highly inaccurate and may increase the risk of multiple insertion attempts, patient suffering and complication rates such as spinal cord injury. Various clinical studies have confirmed the effectiveness of ultrasound imaging compared with the traditional palpation method. However, none of the present system can achieve real-time guidance.

The overall aim of this proposal is to develop an ultrasound guided automated spinal landmark identification with real-time neuraxial needle insertion system (uSINE) to improve patient safety and efficacy of neuraxial procedure needle insertion success. This will be achieved in 4 phases of technology development and clinical trial phases. The investigators will recruit 20 subjects in a prospective cohort study to investigate the spinal needle first attempt success rate as a clinically relevant outcome. The secondary aim is to develop an automated spinal landmark identification algorithm using image processing to identify spinal landmarks in 50 obese patients. Third phase objective will be to obtain clinical data, and evaluation and annotation of the clinical data of spinal ultrasonography in 65 obese patients (BMI \> 30kg/m2), whereas the fourth phase will measure the uSINE identification accuracy and first-attempt puncture success rate of uSINE in a clinical study of 65 obese patients (BMI \> 30kg/m2).