The Lenke paradox: substantial agreement or methodological mirage?

Systematic review and meta-analysis.

To evaluate inter- and intra-observer reliability of the Lenke classification for adolescent idiopathic scoliosis (AIS) and identify methodological factors influencing agreement. The Lenke system is widely used in AIS surgical planning. While initial reports showed excellent reproducibility, subsequent studies have yielded inconsistent results, particularly for lumbar and sagittal modifiers.

Following PRISMA guidelines, five databases (PubMed, Embase, Cochrane CENTRAL, Scopus, and Web of Science) were searched to August 2025. Data on observer characteristics, imaging protocols, measurement methods, and statistical indices (κ, ICC) were extracted. Study quality was assessed with QAREL. Descriptive synthesis and random-effects meta-analysis were performed.

Eleven studies (528 radiographs) met inclusion. Inter-observer reliability averaged κ=0.76 (range 0.50-0.96); lumbar κ=0.75 (0.53-0.92); sagittal κ=0.79 (0.41-1.0). Pooled analysis of two studies with confidence intervals yielded κ=0.69 (95% CI 0.56-0.82; I²=21%). Intra-observer reliability was higher (κ≈0.82). Reliability improved with pre-measured radiographs, experienced observers, computer assistance, and short repeat intervals; it was lower with non-premeasured films, mixed cohorts, and longer intervals. Lumbar modifiers were relatively robust, whereas sagittal reproducibility was most vulnerable to methodological bias. One study illustrated the κ paradox, with high agreement but κ=0.41 due to skewed distribution.

The Lenke classification is reliable under controlled conditions but less consistent in everyday practice. Reported κ values should be interpreted within their methodological context. Standardised protocols, transparent reporting, and digital tools may improve reproducibility.