From Bench-to-Bedside: How Artificial Intelligence is Changing Thyroid Nodule Diagnostics, a Systematic Review
Authors' conclusion
Models predominantly used US images to predict malignancy. Of 4 Food and Drug Administration-approved products, only S-Detect was extensively validated. Implementing an AI model locally requires data sanitization and revalidation to ensure appropriate clinical performance.
Study context
- Population
- Studies of artificial intelligence (AI)-based diagnostic tests on suspected thyroid nodules with prospective or external validation, English-language, PubMed and IEEE Xplore, January 1, 2015 to January 1, 202361 studies
- Outcome
- Characteristics of included studies
Methodological detailsQuality graded by Oxford level of evidence
- Quality graded by Oxford level of evidence
Pooled result
As reported in the abstract
- Studies identified
- 61
- Studies with external validation
- all
- Prospective studies
- 16
- Studies comparing a model with physician prediction of ground truth
- 33
- Papers reporting statistical validation
- 50
- High-level outcomes of the diagnostic pipeline: nodule localization, ultrasound (US) risk score, molecular status, malignancy, long-term prognosis
- 5
- Prospective studies validating a single commercial AI; strengths: automated nodule feature assessment from US, assisting physician malignancy risk prediction; weaknesses: automated margin prediction, interobserver variability
- 7
Original publication
From Bench-to-Bedside: How Artificial Intelligence is Changing Thyroid Nodule Diagnostics, a Systematic Review. J Clin Endocrinol Metab. Published 29 April 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →