Optimizing the quality of emergency head CT imaging: An automated pipeline for correcting head image position

Abstract

Introduction

This study aims to evaluate the quality of head CT images in emergency radiology at a public hospital in China and to investigate whether the implementation of an automatic head CT image position correction pipeline can improve radiologists’ reading efficiency and reduce the rate of missed skull base fractures.

Methods

A total of 15,560 distinct emergency head CT examinations performed between January 2019 and December 2020 at Nanjing Drum Tower Hospital were included in this study. All head CT scans were normalized to Montreal Neurological Institute (MNI) space and the orientation matrices were obtained. Objective image quality analysis was conducted using signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) on both native and standard space CT images. Three rotation angles-yaw, roll and pitch-were calculated from the orientation matrices to evaluate the head position displacement relative to the standard position.

Results

The roll angle was significantly greater than yaw and pitch angles. After normalization, SNR and CNR values improved significantly, and the rate of missed skull base fractures decreased substantially (from 16.63 % to 5.54 %).

Conclusion

The automatic head CT image position correction pipeline significantly enhances the emergency head CT image quality and improves the radiologists’ diagnosis efficiency and accuracy.

Implications for practice

The automatic head image position correction pipeline offers significant improvements in emergency head CT image quality, enabling radiologists to interpret images more efficiently and accurately, saving valuable time for emergency patients ultimately.