This study aims to evaluate the repeatability and reproducibility of two different methods of 3D reconstruction of the spine sterEOS^® and BIOMODTM3S.

Repeatability and reproducibility study. Three observers performed the reconstructions: a radiologist, a X-ray technologist and a rehabilitation specialist, inexperienced in X-ray reading. The observers made these reconstructions with each modality: sterEOS^® and BIOMODTM3S. The parameters investigated were Cobb angle, sagittal parameters (cyphosis, lordosis), determination of apical and junctional vertebrae, axial rotation of the apical vertebra, pelvic parameters and time of reconstruction. Statistical analyses were done using Intraclass Correlation Coefficient (ICC) for reproducibility and Student's t test for time of reconstruction.

We analyzed X-rays of 44 women (71%) and 18 men (29%) with a mean age of 44±20.8. The repeatability was correct, good or excellent depending on observer. The reproducibility inter-observer was correct to excellent (ICC 0.73–0.96) for every parameter except the axial rotation of the apical vertebrae and the determination of levels of junctional and apical vertebrae. The reproducibility of the axial rotation of apical vertebrae was low to good with BIOMODTM3S (ICC 0.15–0.81; ESM=7.5°). The reproducibility of the determination of levels of junctional and apical vertebrae was low to excellent with sterEOS^® (ICC 0.36–0.90). With sterEOS^®, the reproducibility was impaired by the inexperienced observator for some parameters. The 3D reconstructions with sterEOS^® was significantly faster than with BIOMODTM3S (10.8min vs 14.2min, p<0.05).

Parameters’ reproducibility is different depending on the system. The 3D reconstruction with sterEOS^® is faster than with BIOMODTM3S. The reproducibility of BIOMODTM3S is less influenced by observator's experience