RESOURCES

Thyroid nodules are extremely common and can be detected by sensitive imaging in more than 60% of the general population. They are often identified in patients without symptoms who are undergoing evaluation for other medical complaints. Indiscriminate evaluation of thyroid nodules with thyroid biopsy could cause a harmful epidemic of diagnoses of thyroid cancer, but inadequate selection of thyroid nodules for biopsy can lead to missed diagnoses of clinically relevant thyroid cancer. Recent clinical guidelines advocate a more conservative approach in the evaluation of thyroid nodules based on risk assessment for thyroid cancer, as determined by clinical and ultrasound features to guide the need for biopsy. Moreover, newer evidence suggests that for patients with indeterminate thyroid biopsy results, a combined assessment including the initial ultrasound risk stratification or other ancillary testing (molecular markers, second opinion on thyroid cytology) can further clarify the risk of thyroid cancer and the management strategies. This review summarizes the clinical importance of adequate evaluation of thyroid nodules, focuses on the clinical evidence for diagnostic tests that can clarify the risk of thyroid cancer, and highlights the importance of considering the patient’s values and preferences when deciding on management strategies in the setting of uncertainty about the risk of thyroid cancer.

Artificial intelligence (AI) has illuminated a clear path towards an evolving health-care system replete with enhanced precision and computing capabilities. Medical imaging analysis can be strengthened by machine learning as the multidimensional data generated by imaging naturally lends itself to hierarchical classification. In this Review, we describe the role of machine intelligence in image-based endocrine cancer diagnostics. We first provide a brief overview of AI and consider its intuitive incorporation into the clinical workflow. We then discuss how AI can be applied for the characterization of adrenal, pancreatic, pituitary and thyroid masses in order to support clinicians in their diagnostic interpretations. This Review also puts forth a number of key evaluation criteria for machine learning in medicine that physicians can use in their appraisals of these algorithms. We identify mitigation strategies to address ongoing challenges around data availability and model interpretability in the context of endocrine cancer diagnosis. Finally, we delve into frontiers in systems integration for AI, discussing automated pipelines and evolving computing platforms that leverage distributed, decentralized and quantum techniques.

Population-based studies have demonstrated that an increasing number of incidental thyroid nodules are being identified. The corresponding increase in thyroid-based diagnostic procedures, such as fine-needle aspiration biopsy, has in part led to an increase in the diagnoses of thyroid cancers and to more thyroid surgeries being performed. Small papillary thyroid cancers account for most of this increase in diagnoses. These cancers are considered to be low risk because of the excellent patient outcomes, with a 5-year disease-specific survival of >98%. As a result, controversy remains regarding the optimal management of newly diagnosed differentiated thyroid cancer, as the complications related to thyroidectomy (primarily recurrent laryngeal nerve injury and hypoparathyroidism) have considerable effects on patient quality of life. This Review highlights current debates, including undertaking active surveillance versus thyroid surgery for papillary thyroid microcarcinoma, the extent of thyroid surgery and lymphadenectomy for low-risk differentiated thyroid cancer, and the use of molecular testing to guide decision-making about whether surgery is required and the extent of the initial operation. This Review includes a discussion of current consensus guideline recommendations regarding these topics in patients with differentiated thyroid cancer. Additionally, innovative thyroidectomy techniques (including robotic and transoral approaches) are discussed, with an emphasis on patient preferences around decision-making and outcomes following thyroidectomy.