The minefield of indeterminate thyroid nodules: could artificial intelligence be a suitable diagnostic tool?

Artificial intelligence (AI) is assuming a central role in anatomic pathology for ancillary diagnosis in histology and cytology. AI techniques can analyse large amounts of data and identify patterns that may not be visible to the human eye. Several studies have explored the potential of such techniques to improve the accuracy and efficiency of thyroid nodule diagnosis and to increase the sensitivity and specificity of thyroid cytopathology. Specifically, the indeterminate categories of ‘the Bethesda system for reporting thyroid cytopathology’ (TBSRTC) represent a major diagnostic challenge, and articles reported in this review highlight the potential of new AI technologies in improving the accuracy and standardisation of the cytological diagnosis of indeterminate thyroid nodules. Although a large amount of data supports AI's utility in thyroid cytopathology, further research is needed to integrate and standardise AI-based diagnostic systems in clinical workflows.