RESUMO
BACKGROUND: Thyroid cancer is the most common endocrine cancer, and its incidence has been increasing worldwide in the past decades. The increasing demand in medicine for rapid and accurate diagnosis enabled the application of digital imaging analysis in order to increase workflow efficiency and accurate analyses. The present study aimed to automatically differentiate papillary thyroid carcinoma from normal thyroid cells using high-throughput image analysis. METHODS: Images of cellular specimens were taken with a digital camera and were subsequently analyzed. Other software was used for machine-learning-based cellular diagnostics. RESULTS: The two different classes were correctly identified with high sensitivity and specificity. CONCLUSION: The data created offers great potential for an automated diagnosis. Diagn. Cytopathol. 2016;44:574-577. © 2016 Wiley Periodicals, Inc.