Distal fusion vertebra selection in neurofibromatosis type 1 scoliosis: integrating CT/MRI-detected atrophic changes reduces long-term mechanical complications

Purpose Distal fusion vertebra (DFV) selection in neurofibromatosis type 1 (NF1) scoliosis is traditionally guided by adolescent idiopathic scoliosis (AIS) criteria, which do not account for NF1-specific atrophic changes. We evaluated whether integrating preoperative CT/MRI-detected atrophic changes into DFV selection reduces long-term mechanical complications. Methods This retrospective cohort study included 156 NF1 scoliosis patients (Lenke 1–3, Cobb ≥ 45°) undergoing posterior spinal fusion (2010–2018) with a minimum 5-year follow-up. Patients were stratified into an optimized group (AIS criteria plus CT/MRI atrophic assessment) and a traditional group (AIS criteria alone). Propensity score matching (1:1) yielded 65 matched pairs. Primary outcomes were internal fixation failure (IFF) and distal adding-on. Multivariate logistic regression and ROC analysis were performed. Results The optimized group demonstrated lower IFF (12.3% vs. 32.3%; OR 0.29, 95% CI 0.12–0.68; p = 0.002) and distal adding-on (15.4% vs. 40.0%; OR 0.27, 95% CI 0.12–0.60; p < 0.001). Absolute risk reductions were 20.0% (IFF) and 24.6% (adding-on), corresponding to numbers needed to treat of 5 and 4, respectively. Correction magnitude and fusion length were comparable between groups. DFV adjacency to vertebral (OR 6.12), disc (OR 5.37), and paraspinal soft-tissue atrophy (OR 4.89) independently predicted complications. Combined CT/MRI assessment demonstrated good discrimination (AUC = 0.87; sensitivity 82.6%, specificity 80.3%). Conclusion Integration of CT/MRI-detected atrophic changes into DFV selection was associated with significantly reduced long-term mechanical complications without compromising deformity correction in NF1 scoliosis.