โก Quick Summary
This study evaluated a deep learning model for detecting radiographic sacroiliitis in patients with axial spondyloarthritis (axSpA) using data from the RAPID-axSpA and C-OPTIMISE trials. The model demonstrated impressive performance, achieving 82% sensitivity and 81% specificity in one cohort, indicating its potential to enhance diagnostic accuracy and efficiency.
๐ Key Details
- ๐ Dataset: Radiographs from RAPID-axSpA (n=277) and C-OPTIMISE (n=739) trials
- ๐งฉ Features used: Baseline X-ray images
- โ๏ธ Technology: Deep learning model utilizing transfer learning
- ๐ Performance: RAPID-axSpA: 82% sensitivity, 81% specificity; C-OPTIMISE: 90% sensitivity, 56% specificity
๐ Key Takeaways
- ๐ค Machine learning can significantly reduce interreader variability in radiographic assessments.
- ๐ The deep learning model achieved a Cohen’s ฮบ of 0.61 in RAPID-axSpA, indicating good agreement with expert readers.
- ๐ก In C-OPTIMISE, the model’s performance varied, highlighting the need for further refinement.
- ๐ฅ Potential benefits include expedited diagnosis and reduced healthcare resource usage.
- ๐ The study supports the integration of AI in clinical settings for improved patient care pathways.
- ๐๏ธ ClinicalTrials.gov Identifiers: NCT01087762 and NCT02505542.
๐ Background
The diagnosis of axial spondyloarthritis (axSpA) heavily relies on the identification of sacroiliitis through radiographic imaging. However, traditional methods such as X-rays often suffer from significant interreader variability, leading to inconsistent diagnoses. The advent of machine learning offers a promising solution to enhance diagnostic accuracy and efficiency in this domain.
๐๏ธ Study
This study utilized radiographs from two major clinical trials, RAPID-axSpA and C-OPTIMISE, to assess the performance of a deep learning model in detecting radiographic sacroiliitis. The model was trained using a transfer learning approach on non-medical data, which was then validated against expert readers’ assessments of baseline X-rays.
๐ Results
The results indicated that the deep learning model performed well, achieving 82% sensitivity and 81% specificity in the RAPID-axSpA cohort, with a Cohen’s ฮบ of 0.61. In the C-OPTIMISE cohort, the model demonstrated 90% sensitivity but lower specificity at 56%, with a Cohen’s ฮบ of 0.48. The agreement between the model and central readers was notable, at 82% for RAPID-axSpA and 75% for C-OPTIMISE.
๐ Impact and Implications
The findings from this study suggest that machine learning can play a crucial role in the diagnosis of axSpA by providing a reliable and efficient method for detecting sacroiliitis. By reducing the time taken for diagnosis and minimizing variability among readers, this technology could lead to improved patient care pathways and more effective use of healthcare resources. The integration of AI in clinical practice is becoming increasingly vital for enhancing diagnostic processes.
๐ฎ Conclusion
This study highlights the significant potential of deep learning in the field of radiographic analysis for axSpA. The ability to accurately detect sacroiliitis not only expedites diagnosis but also enhances the overall quality of patient care. As we continue to explore the integration of AI technologies in healthcare, further research and development in this area are essential to fully realize their benefits.
๐ฌ Your comments
What are your thoughts on the use of machine learning in diagnosing conditions like axSpA? We would love to hear your insights! ๐ฌ Leave your comments below or connect with us on social media:
Machine learning vs human experts: sacroiliitis analysis from the RAPID-axSpA and C-OPTIMISE phase 3 axSpA trials.
Abstract
OBJECTIVE: Diagnosis of axial spondyloarthritis (axSpA) is primarily established through the identification of the presence or absence of radiographic sacroiliitis. However, the reliability of conventional radiographs (X-rays) is undermined by significant interreader variability. A machine learning tool could reduce diagnosis time, thereby minimising interreader variability. The present study aimed to evaluate the performance of a deep learning model for detecting radiographic sacroiliitis in axSpA patients from the RAPID-axSpA (NCT01087762) and C-OPTIMISE (NCT02505542) trials.
METHODS: Radiographs from the RAPID-axSpA and C-OPTIMISE cohorts were retrospectively used. The deep learning model was previously trained by using a transfer learning approach on non-medical data. The model’s agreement with expert readers was tested on baseline X-rays using central reader data. Sensitivity, specificity, Cohen’s ฮบ, positive and negative predictive values and the area under the receiver operating characteristics curve were calculated.
RESULTS: The model’s performance was evaluated in the RAPID-axSpA (nโ=โ277) and C-OPTIMISE (nโ=โ739) cohorts. In RAPID-axSpA, the model achieved 82% sensitivity, 81% specificity and a Cohen’s ฮบ of 0.61, closely matching central reader performance. In C-OPTIMISE, the model demonstrated 90% sensitivity, 56% specificity and a Cohen’s ฮบ of 0.48. The agreement between the model and central readers was 82% (RAPID-axSpA) and 75% (C-OPTIMISE).
CONCLUSIONS: The tested deep learning model exhibited accurate radiographic sacroiliitis detection in axSpA patients from diverse clinical trials. The proposed deep learning model could expedite diagnosis, reduce healthcare resource usage and improve patient care pathways.
Author: [‘Proft F’, ‘Vahldiek JL’, ‘Nicolaes J’, ‘Tham R’, ‘Hoepken B’, ‘Ufuktepe B’, ‘Poddubnyy D’, ‘Bressem KK’]
Journal: Rheumatol Adv Pract
Citation: Proft F, et al. Machine learning vs human experts: sacroiliitis analysis from the RAPID-axSpA and C-OPTIMISE phase 3 axSpA trials. Machine learning vs human experts: sacroiliitis analysis from the RAPID-axSpA and C-OPTIMISE phase 3 axSpA trials. 2025; 9:rkae118. doi: 10.1093/rap/rkae118