An end-to-end pipeline for automated scoliosis diagnosis with standardized clinical reporting using SNOMED CT

Abstract Recent advancements in automated scoliosis diagnosis using X-ray imaging have focused on detecting spinal curvature through angle measurements, but they often lack systematically standardized clinical terminology schemes. To address these limitations, we propose an end-to-end pipeline that...

Full description

Saved in:
Bibliographic Details
Main Authors: Abdullah Shahid, Jeonghoon Kim, Shi Sub Byon, SungHyuk Hong, Inyong Lee, Byoung-Dai Lee
Format: Article
Language:English
Published: Nature Portfolio 2025-05-01
Series:Scientific Reports
Online Access:https://doi.org/10.1038/s41598-025-01952-w
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract Recent advancements in automated scoliosis diagnosis using X-ray imaging have focused on detecting spinal curvature through angle measurements, but they often lack systematically standardized clinical terminology schemes. To address these limitations, we propose an end-to-end pipeline that detects individual vertebrae and Cobb angles including reports of structured medical terms. The performance of the proposed pipeline was evaluated on the SpineWeb public dataset, achieving state-of-the-art results with a Circular Mean Absolute Error of 3.50 and Symmetric Mean Absolute Percentage Error of 7.35, outperforming existing methods in both accuracy and efficiency. The results are defined and reported using SNOMED CT-based templates to improve data quality and facilitate efficient secondary use. We have designed our solution to perform large-scale data analysis and inter-hospital comparisons, which can positively impact the healthcare of spinal scoliosis patients.
ISSN:2045-2322