Development and internal validation of a clinico-radiological classification system for lumbar adjacent segment degeneration after TLIF

<h2>Abstract</h2><h3>Background</h3> Adjacent segment degeneration (ASD) following transforaminal lumbar interbody fusion (TLIF) lacks a standardized classification system integrating both clinical and radiological parameters. We aimed to develop and internally validate a novel clinico-radiological classification system for lumbar ASD after TLIF and evaluate its association with surgical intervention. <h3>Methods</h3> We retrospectively analyzed 150 patients who underwent one- or two-level TLIF with a minimum 5-year follow-up. Clinical variables (low back pain [LBP], radiculopathy, claudication, neurological deficit) and radiographic/magnetic resonance imaging (MRI) findings were assessed. Latent class analysis identified five distinct classes, forming a five-grade system: Grade 0: Nil or occasional LBP with normal imaging; Grade 1: Continuous LBP with mild radiological degeneration without instability; Grade 2: Radicular pain with moderate degeneration and/or instability; Grade 3: Claudication with advanced degeneration and severe central stenosis; Grade 4: Neurological deficit and/or proximal junctional kyphosis or failure. Internal validation was performed in 30 independent cases using Fleiss' kappa. Firth's penalized logistic regression identified predictors of surgery. <h3>Results</h3> The classification demonstrated substantial inter-observer agreement (κ = 0.74) and excellent intra-observer reliability (κ = 0.82). Surgical intervention was strongly associated with higher grades: 87% of patients in Grades 3–4 underwent surgery compared to 6.7% in Grades 0–2 (p < 0.001). High-grade ASD (odds ratio [OR] 24.30, p < 0.001) and diabetes mellitus (OR 3.24, p = 0.005) were independent predictors of surgery. Average grading time was 2–3 minutes per case. <h3>Conclusions</h3> This novel clinico-radiological classification system for lumbar ASD after TLIF is reliable, practical, and strongly associated with surgical decision-making. It may assist in clinical stratification and follow-up planning. Prospective multicenter validation is warranted.