Evaluation of CT in Predicting Central Lymph Node Metastasis of Papillary Thyroid Microcarcinoma

Abstract

Objective: To determine the value of CT in predicting CLNM in PTMC. Methods: 157 patients with PTMC confirmed by pathology in our hospital were enrolled, and the CT and clinicopathological data of the patients were retrospectively analyzed. ROC curve was used to determine the optimal cutoff value of PTMC greatest diameter for CLNM. The binary logistic regression model of PTMC CLNM was established based on CT and clinical pathological data, and the diagnostic value of the model was evaluated by ROC curve. Results: According to the ROC curve, the optimal cutoff value for predicting PTMC CLNM was 6 mm. Univariate analysis: Cookie bite sign , microcalcification, multifocality, PTMC greastest diameter ≥6 mm, male, Age<45 were risk factors for PTMC CLNM. Binary Logistic regression analysis: Cookie bite sign with protruding (OR=5.159, 95% CI=1.137 ~ 23.400), multifocality (OR=2.734, 95% CI=1.215 ~ 6.154), PTMC greastest diameter ≥6 mm (OR=3.259, 95% CI=1.326 ~ 8.008), male (OR=3.776, 95% CI=1.339 ~ 10.653), age <45 (OR=3.222, 95% CI=1.419 ~ 7.777), were independent risk factors for PTMC CLNM. According to the ROC curve, when the Youden index=0.502, the sensitivity and specificity in predicting CLNM were 82.5% and 68.0%, respectively. Conclusion: The binary logistic regression model is helpful in predicting PTMC CLNM. Cookie bite sign with protruding, PTMC greastest diameter≥6 mm, male, and age<45 were independent risk factors for PTMC CLNM. For this type of patients, we suggest that surgeons should consider central lymph node dissection.

Keywords

• computed tomograph
• papillary thyroid microcarcinoma
• logistic regression analysis