Inter-observer and intra-observer agreement of cervical vertebral maturation staging: A systematic review and meta-analysis

Abstract

The accurate assessment of active growth is pivotal for the correction of skeletal malocclusion in growing patients. Cervical vertebral maturation (CVM) staging is easy and devoid of unnecessary radiation, but its inter- and intra-observer agreement is still debatable. This study aims to assess inter-observer (reliability) and intra-observer agreement (reproducibility) for CVM staging. A comprehensive literature search across five databases up to October 2023 was conducted. Inclusion criteria comprised observational studies that specifically reported intra-rater, inter-rater, or both agreements as their primary objectives for cervical vertebral maturation (CVM) staging. Studies with a sample size exceeding 15 participants were considered for inclusion. After duplicate study selection, data extraction, and risk-of-bias assessment, random-effects meta-analyses of kappa (k)/correlation coefficient (r) and their 95% confidence intervals (CIs) were performed, followed by meta-regressions, sensitivity analyses, and subgroup analyses. Seventeen observational studies (comprising 1437 lateral cephalograms and 110 assessors) were included. The reliability (8 studies; k=0.62 [95% CI: 0.44, 0.78]) and reproducibility (9 studies; k=0.708 [95% CI: 0.59, 0.82]) were substantial. The inter-observer correlation was almost perfect (in 9 studies; r=0.86 [95% CI: 0.82, 0.89]) while intra-observer correlation was substantial (in 2 studies; r=0.75 [95% CI: 0.62, 0.84]). Tracing of lateral cephalograms significantly increased inter-observer reliability (β=0.29 [0.57, 0.0031]) but cropping and time of assessment (initial vs. follow-up) had no significant impact. The reliability (P<0.001) and reproducibility (P=0.049) of high-quality studies were superior to those observed in low-quality studies. Assessment of the CVM staging is accurate and reproducible to a satisfactory level. The accuracy and reproducibility of CVM are higher in studies utilizing traced cephalograms and those with a low-risk of bias. PROSPERO registration (CRD42023468521). Data is openly available at https://doi.org/10.5281/zenodo.10599129.