Learning from multiple readings for axial spondyloarthritis classification of the sacroiliac joints

2026-02-19 · Source: Machine learning : nature.com subject feeds · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Medical Devices & Health Technology, Data Science & Analytics · Depth: Expert, medium

Summary

A fully automated machine learning system has been developed for the classification of axial spondyloarthritis (axSpA) lesions in sacroiliac joints (SIJs) using magnetic resonance imaging (MRI). This end-to-end pipeline automatically delineates SIJ contours via a vector-field-based open-contour model and classifies five lesion types: bone marrow oedema, ankylosis, sclerosis, erosions, and fatty lesions, utilizing both T1-weighted and STIR sequences. The system incorporates a multi-reader learning framework to account for inter- and intra-reader variability, leveraging multiple expert readings and consensus labels. Evaluated using patient-wise cross-validation on the MEASURE-1 clinical trial data and validated on PREVENT and SURPASS datasets, the approach achieved AUCs ranging from 0.85 to 0.99 for lesion classification, with contouring accuracy showing 95% of errors below 2.76mm. This performance is comparable to expert inter-reader agreement, indicating its potential to reduce variability and burden in large-scale axSpA MRI studies.

Key takeaway

For radiologists and rheumatologists evaluating axial spondyloarthritis, this automated MRI classification system offers a robust solution to reduce diagnostic variability and workload. Its expert-level performance in delineating sacroiliac joints and classifying lesions, validated across multiple clinical datasets, suggests it can significantly enhance consistency in both research and patient management. Consider integrating such automated tools to streamline large-scale axSpA MRI studies and improve diagnostic reliability.

Key insights

Automated MRI analysis of sacroiliac joints can achieve expert-level axSpA lesion classification and reduce reader variability.

Principles

• Multi-reader learning improves model robustness.
• Consensus labels enhance classification performance.

Method

The system uses a vector-field-based open-contour model for SIJ delineation, followed by classification of five lesion types using T1-weighted and STIR MRI sequences within a multi-reader learning framework.

In practice

• Automate SIJ contouring for consistency.
• Classify five key axSpA lesion types.
• Incorporate multiple expert readings for model training.

Topics

• Axial Spondyloarthritis
• MRI Analysis
• Multi-Reader Learning
• Sacroiliac Joint Delineation
• Lesion Classification

Best for: Computer Vision Engineer, AI Scientist, Research Scientist, AI Researcher, Machine Learning Engineer, Data Scientist

Related on AIssential

• 3D Classification of Paramagnetic Rim Lesions in Multiple Sclerosis via Asymmetric QSM-FLAIR Modeling — Asymmetric multimodal deep learning improves 3D classification of MS paramagnetic rim lesions using QSM and FLAIR.
• CAHAL: Clinically Applicable resolution enHAncement for Low-resolution MRI scans — CAHAL enhances clinical MRI resolution by using physics-informed, native-space processing and a bivariate Mixture of Experts to prevent anatomical distortions.
• Intracranial Aneurysm Classification and Segmentation via Tri-Axial ROI and Multi-Task Learning — A multi-task framework enhances intracranial aneurysm analysis by integrating classification and multi-class segmentation across diverse modalities.
• Cascade Classification of Dermoscopic Images of Skin Neoplasms with Controllable Sensitivity and External Clinical Validation — Cascade classification with a tunable triage threshold offers superior sensitivity control and better mimics clinical diagnosis logic.
• Efficient AI-Driven Multi-Section Whole Slide Image Analysis for Biochemical Recurrence Prediction in Prostate Cancer — A novel AI framework predicts prostate cancer biochemical recurrence using multi-section whole slide images, outperforming clinical benchmarks.
• Applying ensemble machine learning techniques to MRI scans to predict Alzheimer’s disease — An ensemble machine learning framework predicts preclinical Alzheimer's from structural MRI using VGG16 features and soft voting.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.