Title: VerSe: A Vertebrae Labelling and Segmentation Benchmark for Multi-detector CT Images

Abstract: Reliable automated processing of spinal images is expected to benefit decision-support systems for diagnosis, surgery planning, and population-based analysis on spine and bone health. Vertebral labelling and segmentation are two fundamental tasks in such an automated pipeline. Centred around these tasks, the Large Scale Vertebrae Segmentation Challenge (VerSe) was organised in conjunction with the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) 2019. This work is a technical report summarising the challenge's findings. A total of 160 multi-detector CT scans closely resembling a typical spine-centred clinical setting were prepared and annotated at voxel-level by a human-machine hybrid algorithm. Both the annotation protocol and the algorithm that aided the medical experts in this annotation process are presented. Eleven fully automated algorithms of the participating teams were benchmarked on the VerSe data. A detailed performance comparison of these algorithms along with insights into their design are presented. The best-performing algorithm achieved a vertebrae identification rate of 95\% and a Dice coefficient of 90% on a hidden test set. As an open-call challenge, VerSe'19's annotated image data and its evaluation tools will continue to be publicly accessible through its online portal.