The Role of Unid Virtual Surgical Planning Software and Patient-specific Rods in Short-segment Fusions for Lumbar Degenerative Disease

Introduction: UNiD has emerged as an artificial intelligence, surgical planning software used in spinal deformity surgery to improve spinopelvic alignment corrections. However, there is a lack of literature regarding the accuracy of UNiD planning software for short-segment lumbar fusions in degenerative lumbar disease. This is the first institution to use UNiD for this application. Our objective is to assess the accuracy and feasibility of the UNiD system in planning short-segment spinal fusions for degenerative conditions.

Methods: We conducted a retrospective analysis of a prospectively maintained database and identified 66 patients who underwent 1-3 level lumbar fusion for degenerative disease using UNiD planning software and patient-specific rods. The UNiD software plan focused on correcting the L4-S1 lordosis, global lumbar lordosis (LL), and lordotic distribution while maintaining pelvic tilt (PT) and sacral slope (SS). A comparative analysis was performed between the UNiD simulated plan and achieved spinopelvic parameters on 3-month follow-up radiograph.

Results: Amongst 66 patients with an average age of 59, 53% were female. When compared to the UNiD plan, the achieved LL, L4-S1 lordosis, and pelvic incidence (PI)-LL were significantly undercorrected with a mean difference of 4.474 degrees (p < 0.001), 2.892 degrees (p=0.005), and 4.271 degrees (p < 0.001), respectively. The following achieved spinopelvic parameters were statistically equivalent within 3 degrees: L1-L4 lordosis (p =0.004), lordotic distribution (p=0.044), PT (p= 0.003), and SS (p=0.002) within 3 degrees.

Conclusion : UNID planning software and patient-specific rods may facilitate spinopelvic parameter correction in short segment fusions for lumbar degenerative disease. Despite the undercorrection in L4-S1 lordosis and LL relative to predicted values, overall lordosis distribution goals were met and lordosis distribution index improved. This may be confounded by compensatory changes that were unaccounted for as the spine is still relatively mobile compared to long segment fusions for deformity corrections. With advancements in machine learning and software engineering, predictive capacities will continue to improve resulting in increased accuracy and anticipation of patient compensatory/reciprocal changes with the potential to revolutionize patient-specific care and outcomes in spine surgery.