Agreement of CT versus MRI in reinvestigation of patients operated for lumbar spinal stenosis. a comparison of surgically and non-surgically treated levels

Surgical decision-making for lumbar spinal stenosis (LSS) depends on accurate imaging. MRI is preferred, but contraindications may require CT. This study compared CT‑ and MRI-based measurements of dural sac cross-sectional area (DSCA) at operated and non-operated levels two years after LSS surgery. Patients from the NORDSTEN Spinal Stenosis Trial (SST) were evaluated. Patients randomized to spinous process osteotomy (SPO) underwent both CT and MRI at two-year follow-up. Two independent readers measured DSCA on axial MRI at L2–L5. On CT, the inner ligamentous canal boundary was delineated to approximate the DSCA. Mean values were compared using Student’s t‑test, and interobserver reliability assessed with intraclass correlation coefficients (ICCs). Of 149 eligible patients, 96 (64%) completed both imaging modalities. Mean age was 65.9 years, and 66% were male. In total, 274 lumbar levels were analysed, including 123 surgically treated levels. CT produced slightly larger canal areas than MRI: +2 to + 4 mm² at non-operated levels and + 11 to + 14 mm² at operated levels (p < 0.05). Interobserver agreement was good for CT and comparable to MRI (ICC 0.80–0.96). CT and MRI provided broadly comparable measurements of the spinal canal, with small systematic differences. The modestly larger CT areas at surgically treated levels are likely of limited clinical relevance, though their importance in patients with very small canals remains uncertain. CT may serve as a supplementary postoperative modality, but further research is needed before it can be considered a consistent alternative to MRI.