Objective:

To examine the current progress in AI-driven thyroid disease management through a systematic review, challenges to clinical implementation, and priorities for future development.

Approach:

Key Findings:

• AI has improved diagnostic accuracy for thyroid nodules, achieving over 90% accuracy in ultrasound assessments.
• AI-assisted systems reduced unnecessary fine-needle aspiration biopsies from 30-38% to about 5%.
• AI demonstrated superior predictive capability for preoperative cervical lymph node metastasis compared to senior radiologists.
• AI models supported personalized treatment decisions and improved monitoring for recurrence risks, including remote monitoring through smartphone and wearable data.

Interpretation:

Despite significant advancements in AI applications for thyroid disease, challenges such as limited generalizability, the 'black-box' nature of AI, integration with clinical workflows, and unresolved ethical and legal issues hinder widespread adoption.

Limitations:

• 93% of studies relied on single-center, hospital-based cohorts, raising concerns about generalizability.
• 90% focused on classical papillary thyroid carcinoma.
• 83% evaluated models trained on Asian data sets.

Conclusion:

Addressing identified priorities for future research, such as improving algorithmic integration and conducting prospective trials, could bridge the gap between AI innovation and equitable clinical practice.

Sources:

• Frontiers in Endocrinology
Attribution Notice

This content is an AI-generated, fully rewritten summary based on a published scholarly article. It does not reproduce the original text and is not a substitute for the original publication. Readers are encouraged to consult the source for full context, data, and methodology.