Transverse-plane descriptors and modifiers of the SRS-Lenke-Aubin 3D classification provide complementary and clinically informative 3D features beyond 2D assessment in Lenke 1 AIS curves

Abstract Purpose To investigate how the transverse-plane descriptors and modifiers of the SRS-Lenke-Aubin 3D classification—orientation of the regional plane of deformation (ORPD), apical vertebral rotation (AVR), and apical displacement (3D distance from the apex to the end-vertebrae line, DAEVL)—characterize Lenke 1 thoracic curves and complement conventional 2D Lenke modifiers and radiographic measures. Methods One hundred sixty-two Lenke 1 adolescent idiopathic scoliosis patients underwent 3D reconstruction from preoperative biplanar radiographs. Associations between transverse-plane descriptors (ORPD, AVR, DAEVL) and standard coronal and sagittal radiographic measures (Cobb angles, thoracic kyphosis [TK], lumbar lordosis [LL]) were analyzed across proximal thoracic (PT), main thoracic (MT), and thoracolumbar/lumbar (TL/L) regions. Relationships between classical Lenke modifiers and transverse-plane modifiers derived from the SRS-Lenke-Aubin 3D classification were evaluated using contingency tables, chi-square tests or Fisher-Freeman-Halton exact tests as appropriate, and Cramér’s V. Pearson correlations and multivariate linear regression were used to assess regional coupling and the independent contributions of coronal and sagittal parameters to ORPD. Subgroup analyses compared Lenke 1A, 1B, and 1C patterns. Results Across analyses, substantial inter-patient variability was observed, with wide dispersion of transverse-plane descriptors for similar 2D measurements. Relationships between coronal severity and transverse-plane geometry were region-dependent. ORPD showed no statistically significant association with Cobb magnitude in the PT and MT regions (p > 0.05), whereas a moderate and statistically significant association was observed in the TL/L region (r = 0.4792, p < 0.001). Associations between AVR and Cobb magnitude were weak in PT (r = 0.1688, p = 0.0317) and moderate in both MT (r = 0.4424, p < 0.001) and TL/L (r = 0.4085, p < 0.001). ORPD and AVR were not significantly associated in PT, showed a weak inverse association in MT (r = −0.2177, p = 0.0054), and were moderately associated in compensatory TL/L (r = 0.5342, p < 0.001). TK demonstrated a moderate-to-strong inverse association with MT ORPD (r = −0.6344, p < 0.001), while LL showed no significant association with TL/L ORPD (r = 0.1177, p = 0.1357). DAEVL showed moderate associations with Cobb magnitude across regions and selective coupling with ORPD and AVR. Strong associations were observed between the Lenke lumbar modifier and TL/L ORPD, and between the thoracic sagittal profile modifier and MT ORPD (both p < 0.001), indicating partial overlap but suggesting complementarity between 2D and 3D descriptors. Multivariate regression confirmed region-specific behavior, with MT ORPD primarily influenced by sagittal profile and TL/L ORPD predominantly driven by coronal magnitude. Subgroup analyses revealed distinct 3D signatures across Lenke 1A, 1B, and 1C, but the 1C subgroup should be interpreted as exploratory. Conclusion In Lenke 1 thoracic curves, given the weak to moderate associations found, ORPD, AVR, and DAEVL thus provide additional, complementary, region-specific information that is not fully explained by conventional 2D parameters. The predominance of weak-to-moderate associations reflects meaningful independence rather than lack of relationship, reinforcing the value of integrating transverse-plane descriptors to enhance 3D characterization of AIS.