ICPPNet: A semantic segmentation network model based on inter-class positional prior for scoliosis reconstruction in ultrasound images.

in Journal of biomedical informatics by Changlong Wang, You Zhou, Yuanshu Li, Wei Pang, Liupu Wang, Wei Du, Hui Yang, Ying Jin

TLDR

  • A novel neural network, ICPPNet, is developed for accurate segmentation of spine regions in ultrasound images for scoliosis diagnosis.
  • ICPPNet achieves high accuracy and efficiency, outperforming traditional methods and showing strong correlation with X-ray images.
  • ICPPNet has the potential to transform scoliosis diagnosis and treatment, providing a safer and more cost-effective alternative to X-ray imaging.

Abstract

Considering the radiation hazard of X-ray, safer, more convenient and cost-effective ultrasound methods are gradually becoming new diagnostic approaches for scoliosis. For ultrasound images of spine regions, it is challenging to accurately identify spine regions in images due to relatively small target areas and the presence of a lot of interfering information. Therefore, we developed a novel neural network that incorporates prior knowledge to precisely segment spine regions in ultrasound images. We constructed a dataset of ultrasound images of spine regions for semantic segmentation. The dataset contains 3,136 images of 30 patients with scoliosis. And we propose a network model (ICPPNet), which fully utilizes inter-class positional prior knowledge by combining an inter-class positional probability heatmap, to achieve accurate segmentation of target areas. ICPPNet achieved an average Dice similarity coefficient of 70.83% and an average 95% Hausdorff distance of 11.28 mm on the dataset, demonstrating its excellent performance. The average error between the Cobb angle measured by our method and the Cobb angle measured by X-ray images is 1.41 degrees, and the coefficient of determination is 0.9879 with a strong correlation. ICPPNet provides a new solution for the medical image segmentation task with positional prior knowledge between target classes. And ICPPNet strongly supports the subsequent reconstruction of spine models using ultrasound images.

Overview

  • The study aims to develop a novel neural network, ICPPNet, for accurate segmentation of spine regions in ultrasound images for scoliosis diagnosis.
  • The network incorporates prior knowledge and is designed to tackle the challenges of small target areas and interfering information in ultrasound images.
  • The study assesses the performance of ICPPNet on a dataset of 3,136 ultrasound images from 30 patients with scoliosis and compares it to X-ray images.

Comparative Analysis & Findings

  • ICPPNet achieved an average Dice similarity coefficient of 70.83% and an average 95% Hausdorff distance of 11.28 mm, demonstrating its excellent performance.
  • The average error between the Cobb angle measured by ICPPNet and X-ray images is 1.41 degrees, and the coefficient of determination is 0.9879 with a strong correlation.
  • ICPPNet outperforms traditional methods in terms of accuracy and efficiency, making it a promising solution for medical image segmentation tasks with positional prior knowledge.

Implications and Future Directions

  • ICPPNet has the potential to revolutionize the diagnosis and treatment of scoliosis, providing a safer, more convenient, and cost-effective alternative to X-ray imaging.
  • Future studies can explore the application of ICPPNet in other medical image segmentation tasks, such as brain and cardiovascular imaging.
  • The integration of ICPPNet with other machine learning algorithms and deep learning models can further improve its accuracy and robustness.
Read Full Article