Sagittal balance parameters measurement on cervical spine MR images based on superpixel segmentation

• DOI: 10.3389/fbioe.2024.1337808
• Publication Date: 2024-04-12T14:36:20Z
• Authors: Yi-Fan Zhong; Yu-Xiang Dai; Shi-Pian Li; Ke-Jia Zhu; Yong-Peng Lin; Yu Ran; Lin Chen; Ye Ruan; Peng-Fei Yu; Lin Li; Wen-Xiong Li; Chuang-Long Xu; Zhi-Tao Sun; Kenneth A. Weber; De-Wei Kong; Feng Yang; Wen-Ping Lin; Jiang Chen; Bo-Lai Chen; Hong Jiang; Ying-Jie Zhou; Bo Sheng; Yong-Jun Wang; Ying-Zhong Tian; Yue-Li Sun
• Keywords: 03 medical and health sciences; superpixel segmentation; 0302 clinical medicine; sagittal balance parameters; magnetic resonance imaging; Bioengineering and Biotechnology; cervical spine; artificial intelligence; TP248.13-248.65; Biotechnology

Introduction: Magnetic Resonance Imaging (MRI) is essential in diagnosing cervical spondylosis, providing detailed visualization of osseous and soft tissue structures the spine. However, manual measurements hinder assessment spine sagittal balance, leading to time-consuming error-prone processes. This study presents Pyramid DBSCAN Simple Linear Iterative Cluster (PDB-SLIC), an automated segmentation algorithm for vertebral bodies T2-weighted MR images, aiming streamline balance spinal surgeons. Method: PDB-SLIC combines SLIC superpixel with clustering underwent rigorous testing using extensive dataset mid-sagittal images from 4,258 patients across ten hospitals China. The efficacy was compared against other algorithms networks terms quality body accuracy. Validation included a comparative analysis parameters scrutiny PDB-SLIC’s measurement stability diverse hospital settings scanning machines. Result: outperforms quality, high accuracy, recall, Jaccard index. Minimal error deviation observed measurements, correlation coefficients exceeding 95%. demonstrated commendable performance processing various settings, MRI machines, patient demographics. Discussion: emerges as accurate, objective, efficient tool evaluating valuable assistance surgeons preoperative assessment, surgical strategy formulation, prognostic inference. Additionally, it facilitates comprehensive cohorts, contributing establishment normative standards imaging.