Value of artificial intelligence assisted scoring system combined with ultrasound elastography in the diagnosis of benign and malignant thyroid nodules

  Objective  To analyze the value of artificial intelligence assisted scoring system combined with ultrasound elastography in the diagnosis of benign and malignant thyroid nodules.

  Methods  Totally 414 patients with 543 thyroid nodules were enrolled in the study. The thyroid nodules were analyzed by artificial intelligence scoring system and ultrasound elastography, and the artificial intelligence assisted score (AIAS) and ratio of strain rate (SR) were recorded. Binary Logistic regression analysis was performed on AIAS and SR of benign and malignant thyroid nodules to calculate the combined diagnostic predictors. The receiver operating characteristic (ROC) curve was drawn to analyze the value of AIAS, SR and their combination in the diagnosis of benign and malignant thyroid nodules.

  Results  Of the 543 nodules, 339 nodules (62.43%) were malignant and 204 nodules (37.57%) were benign. The AIAS, SR and combined predictors of benign nodules were significantly lower than those of malignant nodules (P < 0.05). Area under curve (AUC) of combined diagnosis was significantly higher than that of AIAS and SR (P < 0.05).

  Conclusion  Both artificial intelligence assisted scoring system and ultrasound elastography have certain limitations in the diagnosis of benign and malignant thyroid nodules, but the combination of the two methods is more valuable in the diagnosis of nature of thyroid nodules.