Automatic Plane Adjustment of Orthopedic Intra-operative Flat Panel Detector CT-Volumes

TL;DR

This paper investigates multi-task learning regression networks to automatically adjust multiplanar reconstructed planes in intra-operative CT volumes, improving accuracy and efficiency in orthopedic trauma surgery imaging.

Contribution

It compares different mathematical descriptions and network architectures, demonstrating that a multi-head approach offers slight accuracy improvements and a unified solution for various body regions.

Findings

Matrix description improves normal regression accuracy.

Multi-head approach enhances plane position regression.

Single network reduces stored parameters for all body regions.

Abstract

Purpose 3D acquisitions are often acquired to assess the result in orthopedic trauma surgery. With a mobile C-Arm system, these acquisitions can be performed intra-operatively. That reduces the number of required revision surgeries. However, due to the operation room setup, the acquisitions typically cannot be performed such that the acquired volumes are aligned to the anatomical regions. Thus, the multiplanar reconstructed (MPR) planes need to be adjusted manually during the review of the volume. In this paper, we present a detailed study of multi-task learning (MTL) regression networks to estimate the parameters of the MPR planes. Approach First, various mathematical descriptions for rotation, including Euler angle, quaternion, and matrix representation, are revised. Then, three different MTL network architectures based on the PoseNet are compared with a single task learning network. Results Using a matrix description rather than the Euler angle description, the accuracy of the regressed normals improves from $7.7^{\circ}$ to $7.3^{\circ}$ in the mean value for single anatomies. The multi-head approach improves the regression of the plane position from $7.4mm$ to $6.1mm$, while the orientation does not benefit from this approach. Conclusions The results show that a multi-head approach can lead to slightly better results than the individual tasks networks. The most important benefit of the MTL approach is that it is a single network for standard plane regression for all body regions with a reduced number of stored parameters.