Objective Few authors propose in vivo tools for 3D geometrical description of the rib cage for clinical or biomechanical applications. The purpose of this study is to present a method providing a personalized 3D geometry of the rib cage from biplanar X-rays, and to evaluate its intra- and interobserver reproducibility. A morphometric study of asymptomatic rib cages is also presented. Materials and methods Both frontal and lateral radiographs were taken on 15 asymptomatic volunteers. For each set of stereoradiographs the rib cage was reconstructed six times. Then, shape and morphometric comparisons were performed between the reconstructions realised by the same observer and by different observers. Results The extensive reproducibility evaluation performed demonstrated a good intra-observer reproducibility for local and global morphometric parameters, and a good interobserver reproducibility for global parameters (twice the standard deviation [2S.D.] < 5%). Acceptable interobserver variations for angles (2S.D. < 5°) and local shape parameters (2S.D. < 10%) were also found. The interobserver variability in term of point-to-surface distances was equal to 5.1 mm (mean: 1.9 mm). Discussion The present study evaluated the intra- and interobserver reproducibility of a 3D reconstruction method of the rib cage whose accuracy was previously evaluated. Reproducibility results could not be compared to previous studies because the evaluation procedures were different. However, the morphometric data provided by the present stereoradiographic method could be compared and were found globally consistent with the previous published data. Conclusion The present study completes the evaluation of a stereoradiographic 3D reconstruction method of the rib cage. In addition, the morphometric data acquired will have an important clinical significance for researchers evaluating chest cage morphometry and development.