JMIR Publications

ABSTRACT

Background:

Artificial intelligence (AI) has shown promise for automating spinal alignment assessment in adolescent idiopathic scoliosis (AIS). However, AI models typically exhibit reduced accuracy and robustness when deployed across multiple medical centres due to variability in imaging protocols and data characteristics, potentially compromising clinical diagnosis and treatment decisions.

Objective:

This study aimed to develop a real-time, plug-and-play data transformation method to enhance the robustness of deep learning models against data heterogeneity in radiographs, thereby improving their performance in assessing AIS across multiple medical centres.

Methods:

In this retrospective multicentre study, 4,111 full-spine radiographs from seven hospitals (two from Hong Kong and five from mainland China), collected between January 2012 and August 2024, were included. Data from two hospitals in Hong Kong (n=3,034) were used for model training and internal validation, while radiographs from the five mainland hospitals (n=1,077) formed five independent external validation datasets. A novel pixel-intensity-based data transformation method was developed to standardize image contrast and brightness across datasets and integrated into the model training process to enhance our previously developed AI model, SpineHRNet+. The enhanced model's accuracy and robustness for Cobb angle prediction and severity classification were evaluated using both internal and external datasets. Data heterogeneity across centres was quantified by brightness and contrast differences. Cobb angle prediction accuracy was evaluated using residual analysis, linear regression (coefficient of determination [R²]), and Bland-Altman analyses. Model performance for disease severity classification was assessed using sensitivity, specificity, precision, negative predictive value (NPV), accuracy, and confusion matrix analysis.

Results:

In this retrospective multicentre study, 4,111 full-spine radiographs from seven hospitals (two from Hong Kong and five from mainland China), collected between January 2012 and August 2024, were included. Data from two hospitals in Hong Kong (n=3,034) were used for model training and internal validation, while radiographs from the five mainland hospitals (n=1,077) formed five independent external validation datasets. A novel pixel-intensity-based data transformation method was developed to standardize image contrast and brightness across datasets and integrated into the model training process to enhance our previously developed AI model, SpineHRNet+. The enhanced model's accuracy and robustness for Cobb angle prediction and severity classification were evaluated using both internal and external datasets. Data heterogeneity across centres was quantified by brightness and contrast differences. Cobb angle prediction accuracy was evaluated using residual analysis, linear regression (coefficient of determination [R²]), and Bland-Altman analyses. Model performance for disease severity classification was assessed using sensitivity, specificity, precision, negative predictive value (NPV), accuracy, and confusion matrix analysis.

Conclusions:

The proposed data transformation approach effectively addressed data heterogeneity, significantly improving the accuracy and robustness of SpineHRNet+ in multicentre AIS assessments. The real-time processing capability and preservation of anatomical integrity underscore the method’s clinical practicality, enabling scalable and reliable AI applications in diverse healthcare environments. Clinical Trial: Name: Artificial Intelligence-based Models for Spine Malalignment Auto-analysis ClinicalTrials.gov ID: NCT06711757 URL: https://clinicaltrials.gov/study/NCT06711757?cond=NCT06711757&rank=1&tab=results