Radiomics and artificial Intelligence for thyroid cancer diagnosis: Concepts, challenges, and solutions

AI and radiomics show promise for thyroid cancer diagnosis: Among the 42 studies reviewed, radiomics analysis using ultrasound images demonstrated effectiveness in diagnosing thyroid cancer, with some studies presenting new strategies that outperformed existing diagnostic methods.

Interpretability remains a critical barrier: The review identified interpretability issues as a major challenge facing AI models in this domain, meaning doctors and patients often cannot understand how these systems reach their diagnostic conclusions—a key concern for responsible deployment in clinical settings.

Data limitations and operator dependence undermine reliability: The studies highlighted dataset constraints and operator dependence as significant obstacles, suggesting that current AI systems may not perform consistently across different hospitals, equipment, or technicians without careful oversight.

Standardization and multicenter testing are essential before widespread adoption: The synthesized findings emphasized the need for standardization efforts and prospective multicenter studies, indicating that more rigorous, coordinated research is necessary before these AI tools can be reliably implemented across different healthcare settings.

Explainable AI and personalized medicine offer pathways forward: The review identified advances in explainable AI technology and personalized medicine techniques as promising solutions to address current challenges, suggesting that making AI decisions transparent and tailored to individual patients could help build trust and improve outcomes.