DO CHANGES IN SURFACE TOPOGRAPHY MEASUREMENTS PREDICT CHANGES IN ANGLE OF TRUNK ROTATION PRE- AND POST-PSSE IN THORACIC ADOLESCENT IDIOPATHIC SCOLIOSIS?

Outcome measures such as angle of trunk rotation (ATR) and measurements from surface topography (ST) including apical deviation (AD), coronal imbalance (CI), and scoliosis angle (SA) are commonly used to assess conservative management of adolescent idiopathic scoliosis (AIS). Understanding how these measures interrelate and respond to Physiotherapeutic Scoliosis Specific Exercises (PSSE) validates their combined use in outcome assessment. We aimed to investigate relationships between ST measurements and ATR in a thoracic AIS cohort. Including sub-group analysis in adult (≥18y/o), paediatric (<18y/o), braced paediatric and non-braced paediatric patients and establish whether PSSE influences these measurements and relationships.

Consecutive patients with single thoracic AIS were analysed retrospectively (n=54, mean age 17.04 (SD 6.64), 72.2% female, mean known major Cobb angle 42.3° (SD 14.58°). Patients attended 4 weeks (120-hours) of group-based PSSE between February 2021 and October 2025 in London, United-Kingdom. AD, CI, and SA from ST scans and ATR via Bunnell's Scoliometer were recorded pre- and immediately post-treatment. Correlation (Pearson/Spearman), simple and multiple linear regression, paired t-tests and Wilcoxon signed-rank tests were used with significance set at p<0.05.

In the whole cohort, PSSE significantly improved ATR by 2.52° (n=53, 95%CI[1.24, 3.79], p<.001, d=0.544), AD by 5.48mm (n=54, 95%CI[1.20, 9.76], p=0.013, d=0.35) and SA by 3.59° (n=54, 95%CI[0.64, 6.55], p=0.018, d=0.33). CI did not significantly change. Pre-treatment, ATR significantly correlated with AD (r=0.526, p=<0.001), CI (r=0.272, p=0.049), and SA (r=0.366, p=0.007). A multiple linear regression model containing AD, CI and SA (Adj. R ² =0.287, F(3,49)=7.98, p<0.001) showed AD as the only significant predictor of ATR (β=0.432, p=0.015). AD and SA were strongly intercorrelated (r=0.715, R ² =0.64, F(1,52)=92.02, p<0.001). Post-treatment, ATR remained correlated with AD (r=0.341, p=0.012) and SA (r=0.314, p=0.022), and the AD-SA relationship persisted (r=0.642, R ² =0.41, F(1,52)=36.37, p<0.001).

Subgroup analyses revealed differential patterns. Adults (n=16) showed large improvements in ATR 4.10° (Z=–3.35, p<0.001), AD 10.81mm (Z=–2.36, p=0.018) and SA 7.38° (Z=–2.33, p=0.020). Pre-treatment AD (β=1.012, p=0.004) strongly predicted ATR (Adj. R ² =0.530, F(3,11)=6.267, p=0.010), but post-treatment significant ST-ATR relationships were absent. Paediatric patients overall (n=38) improved ATR by 1.89° (95%CI[1.40, 2.39], p<0.001, d=1.264). Pre-treatment, CI was the significant predictor (β=0.313, p=0.030) in a multiple linear regression model (Adj. R ² =0.335, F(3,34)=7.222, p<0.001), whereas post-treatment AD (β=0.504, p=0.009) became the only significant predictor (Adj. R ² =0.33, F(3,34)=7.046, p<0.001). Among non-braced paediatric patients (n=23), ATR improved by 1.96° (95%CI[1.24, 2.67], p<0.001, d=1.658). Pre-treatment, ATR correlated with AD (r=0.634, p=<0.001) and SA (r=0.628, p=<0.001) but no significant model emerged. Post-treatment, a significant model was observed (Adj. R ² =0.474, F(3,19)=7.596, p=0.002) with AD as the significant predictor (β=0.539, p=0.026). Braced paediatric patients (n=12) improved ATR 1.84° (Z=−2.98, p=0.003) but exhibited no significant ST-ATR correlations or regression models pre- or post-treatment. CI change remained unaffected in all subgroups.

Pre-treatment, AD was the strongest ST predictor of ATR across the whole thoracic AIS cohort, though collinearity with SA limits the independent predictive contribution of each. PSSE improved ATR across all subgroups, however the ST-ATR relationship diverged markedly by age and brace status. In adults, the predictive relationship was removed, suggesting PSSE may decouple coronal surface-morphology from ATR in skeletally mature patients. In non-braced paediatric patients, a significant relationship emerged post-treatment that was absent pre-treatment, with AD as the significant predictor. This divergence indicates that age and brace status substantially moderate how PSSE affects the ST-ATR relationship. The absence of any relationship in braced paediatric patients likely reflects brace-mediated manipulation of ST. CI was unaffected by PSSE. These limited findings support stratified use of ST and ATR in PSSE outcome assessment, with caution when interpreting braced patients' measurements.