Utilizing AI to Enhance Patient Selection and Treatment Response in Cervical Total Disc Arthroplasty Patients

Introduction: While many IDE studies for cervical total disc arthroplasties (cTDR) list strict inclusion criteria, real life clinical data often includes a wider patient population. Pre-operative motion and alignment metrics could provide insight on the likelihood of success with a cTDR, just as post-operative metrics may guide treatment response. However, it is difficult and time consuming for clinicians to analyze motion and alignment metrics with traditional by-hand methods. The purpose of this abstract is to investigate the use of AI which can independently report radiographical metrics that may indicate below-average outcomes.

Methods: Pre-operative and postoperative flexion-extension imaging of 32 subjects implanted with a cTDR was retrospectively analyzed using AI. Patient reported outcomes gathered include NDI, neck pain, and PROMIS General Physical Health. A threshold limit graphical approach was utilized to help identify motion and alignment metrics which may predict below average patient reported outcomes. A two tailed t-test was used to compare means of patients grouped by radiographical motion metrics.

Results: Both preoperative and postoperative metrics were identified that are correlated with below-average outcomes. Preoperative posterior spondylolisthesis was identified as a contraindication for cTDR, with a below average (0% versus > 80%) proportion of patients having a Patient Acceptable Symptom State (NDI <=21). Postoperative range of motion at the operated level was associated with worse improvement in pain from preop to postop of 1.8 versus 3.8 (p=0.0336).

Conclusion : Identifying preoperative and postoperative metrics that are associated with poorer clinical outcomes may lead to better patient selection and swifter treatment response in cTDR patients. It is important to acknowledge the limited sample size in this analysis. Future work surrounding the identified metrics and others may provide clinicians with actionable insights to improve cTDR patient outcomes .