Dataset Overview
Definition：Scoliosis X-Ray Imaging Datasetis provided byChangsha Langhui Information Technology Co., Ltd.It is built by Long Salang Langhui Information Technology Ltd., a vertebrae X-ray image annotated database. The main object is a standing whole ridge orthopedic/backward, containing as far as possible the side and pelvis, and the fusion image retains the original projection and fusion.
| Dataset Name | Scoliosis X-Ray Imaging Dataset |
| Total Data Volume | Total X-Ray Volume Included: 1,700,000 Cases |
| Imaging Modality | X-Ray (DR/CR) |
| Covered Body Parts | Standing All Ridge PA/AP+ Side |
| Raw Data | Dissensitization original DICOM, with exposure parameters, pixel spacing, bits/views, fusion relationships |
| Measurement label | Cobb angle and measurement vertebrae, vertebrae, main bend/subverte, vertebrae rotation, coronal/veal balance |
| Maturity Mark | Risser / Sanders Grade |
| Pre-Operative / Post-Operative | Post-operative data accounts for ≤10%, including internal fixation markers |
| Source Institution | Co-operative Oracle/Rival Surgery, Sankar Hospital |
| Use Cases | Spinal side bend AI screening, Cobb corner automatic measurements, Lenke spectrometry, surgical planning |
Delivery Specifications and Field Requirements
The data set strictly follows the XR-SCOLIOS uniform delivery index requirements to ensure data quality and traceability.
Delivery & Counting：One full vertebrae inspection/Study in one row; one multi-series/multi-view only
Coverage：Mainly, stand-by whole-spring PA/AP, containing as far as possible the side and pelvis; the fusion image must retain the original projection and the fusion relationship
Positives and Composition：Not 90% positive as a hard requirement; different degrees of severity of side bend, pre/post-operative and normal/light contrasts to be covered
Dedicated Fields：Quality, whole spinal integrity, main/subtract, corner Cobb and measurement vertebrae, vertebrae, vertebrae, coronal/vereal balance, Risser/Sanders original, post-operative fixed
Cross-Data Association：Priority for clinical styling, surgery, follow-up and MRI; automatic measurement separated from doctor ' s final values
2026 Frontier AI Research Progress
Domain Review：The vertebrae bend AI in 2025-2026 achieved expert-level Cobb-point measurement consistency in multi-centre validation (ICC 0.94-0.98), with multi-perspective semantics and SNOMED CT standardized reporting lines approaching clinical availability.
End-to-end spinal bend diagnostic pipe (SNOMED CT standardized)
2025The Circular MAE = 3.50 and SMART = 7.35 at Point Cobb on the SpineWeb Open Data Set is better than the method available and automatically generates standardized clinical reports.
Multi-centre in-depth learning on the spinal bend assessment validation
arXiv, 2025The multi-centre validation results showed that the in-depth learning model can be replicated at the multi-centre expert-level corner of Cobb to measure and rank consistency.
The spinal alignment and implant auto-detection
npj Digital Medicine, 2026CNN automatically measures the spinal alignment parameters and detects spinal implants (snail + hooks) to provide an automated programme for post-operative assessment.
Multi-perspective semantics dividing the youth-specific spinal bends
2025In-depth learning is semantically divided by multiple perspectives, which are used for automatic measurements at the corner of Cobb and for the Lenke classification, and is more consistent than manual measurements.
AI versus multicurve measurements by clinicians
2026AI reaches ICC 0.94-0.98 in the main chest bend area and TL/L 0.74-0.89 in the TL/L area, close to the senior expert level.
Annotation Workflow & Quality Control
Image Acquisition & De-identification
Standardized acquisition workflow: complete data is exported directly from the equipment, and patient identifying information (PHI) is removed before the data is stored。
Initial Annotation (Specialist Physician)
Attending physicians annotate case by case against the standards, including lesion localization, morphological description and disease term determination。
Review (Associate Chief Physician or Above)
Experts with the title of Associate Chief Physician or above review the initial annotation results item by item, correcting erroneous annotations and supplementing missing dimensions。
Consistency Assessment
10% of samples are randomly selected and independently annotated by 3 physicians to calculate Fleiss' Kappa; below 0.Dimensions scored 75 are flagged for rework。
License and Usage Agreement
Academic Research License
For universities and research institutions, supporting academic research related to medical AI。After signing the agreement, a de-identified data subset is provided; the source must be acknowledged：Langhui Technology DataAssetsAPI。
Commercial License
For medical device companies and AI diagnostics companies, supporting medical AI product R&D and medical device registration。Provides full datasets + custom annotation + incremental update services。
Data Compliance Statement
All images come from legally authorized sources; PHI fields are fully de-identified and contain no information that can directly identify an individual, in compliance with the Personal Information Protection Law and the Data Security Law。
Customized services
Supports extended requirements such as adding specific disease types, multimodal annotation expansion, and custom training of AI-assisted diagnostic models。
Quick Facts
AI Frontier Research
The vertebrae bend AI in 2025-2026 achieved expert-level Cobb-point measurement consistency in multi-centre validation (ICC 0.94-0.98), with multi-perspective semantics and SNOMED CT standardized reporting lines approaching clinical availability.