Abstract / Summary
Degenerative thoracolumbar kyphosis (DTLK) is a common form of adult degenerative spinal deformity, but postoperative symptom remission varies among patients. This study evaluated a pelvic-compensation-based DTLK phenotyping system and developed a preoperative prediction model for symptom remission. This retrospective study included an asymptomatic reference cohort and a surgical cohort of adults with degenerative spinal deformity. In the reference cohort, pelvic incidence-based targets for pelvic tilt and lumbar lordosis were derived by robust linear regression. In the surgical cohort, Δ PT was defined as actual pelvic tilt minus target pelvic tilt and combined with sagittal vertical axis to classify patients as aligned, compensatory, or decompensatory. Symptom remission was defined as an Oswestry Disability Index raw score of 11 or lower after at least 12 months of follow-up. Multivariable logistic regression was used to construct a nomogram. The reference cohort included 270 adults, and the surgical cohort included 267 patients. The surgical cohort included 66 aligned, 84 compensatory, and 117 decompensatory patients. Compared with the aligned type, the compensatory and decompensatory types were associated with lower odds of symptom remission. Older age and osteoporosis were also independently associated with reduced remission probability. The nomogram achieved an area under the curve of 0.808. The pelvic-compensation-based DTLK phenotype was associated with postoperative symptom remission. This model may support preoperative risk stratification and expectation management. Not applicable, as this is a retrospective observational study.