Objective To investigate the diagnostic value of ultrasound artificial intelligence(AI)in medullary thyroid carcinoma.Methods Retrospective analysis of 47 patients with thyroid medullary carcinoma diagnosed through surgery at the First Affiliated Hospital of Naval Medical University from October 2012 to October 2022(thyroid medullary carcinoma group),and 94 patients with thyroid papillary carcinoma confirmed by surgery during the same period(thyroid papillary carcinoma group)were selected as the control group to compare the diagnostic accuracy of AI for the two groups of lesions;At the same time,compare the diagnostic accuracy of AI and doctors in the group of medullary thyroid cancer.Results The diagnostic coincidence rate of AI for medullary thyroid carcinoma group was 51.1%,and that for papillary thyroid carcinoma group was 88.3%.There was a statistical difference in the diagnostic coincidence rate between the two groups(P＜0.001).In the medullary thyroid carcinoma group,the diagnostic coincidence rate of AI for medullary thyroid carcinoma was 51.1%,and the diagnostic coincidence rate of doctors for medullary thyroid carcinoma was 72.3%.There was a statistical difference in the diagnostic coincidence rate of medullary thyroid carcinoma between the two diagnostic methods(P＜0.05).Conclusion The diagnostic value of AI for medullary thyroid carcinoma remains to be improved.
维普
万方数据