A Nomogram Based on Clinicopathological and Ultrasound Imaging Characteristics for Predicting Cervical Lymph Node Metastasis in cN0 Unilateral Papillary Thyroid Microcarcinoma

Objective: The aim of this study was to establish a practical nomogram for preoperatively predicting the possibility of cervical lymph node metastasis (CLNM) based on clinicopathological and ultrasound (US) imaging characteristics in patients with clinically node-negative (cN0) unilateral papillary thyroid microcarcinoma (PTMC) in order to determine a personal surgical volume and therapeutic strategy. Methods: A total of 269 consecutive patients diagnosed with cN0 unilateral PTMC by postoperative pathological examination from January 2018 to December 2020 were retrospectively analyzed. All the patients underwent lobectomy or thyroidectomy with routine prophylactic central lymph node dissection (CLND) and were divided into a CLNM group and a non-CLNM group. Using logistic regression, the least absolute shrinkage and selection operator (LASSO) regression analysis was applied to determine the risk factors for CLNM in patients with unilateral cN0 PTMC. A nomogram including risk-factor screening using LASSO regression for predicting the CLNM in patients with cN0 unilateral PTMC was further developed and validated. Results: Risk factors identified by LASSO regression, including age, sex, tumor size, presence of extrathyroidal extension (ETE), tumor diameter/lobe thickness (D/T), tumor location, and coexistent benign lesions, were potential predictors for CLNM in patients with cN0 unilateral PTMC. Meanwhile, age (odds ratio [OR] = 0.261, 95% CI.104–0.605; P = 0.003), sex (men: OR = 3.866; 95% CI 1.758–8.880; P < 0.001), ETE (OR = 3.821; 95% CI 1.168–13.861; P = 0.032), D/T (OR = 72.411; 95% CI 5.483–1212.497; P < 0.001), and coexistent benign lesions (OR = 3.112 95% CI 1.407–7.303; P = 0.007) were shown to be significantly related to CLNM by multivariant logistic regression. A nomogram for predicting CLNM in patients with cN0 unilateral PTMC was established based on the risk factors identified by the LASSO regression analysis. The receiver operating characteristic (ROC) curve for predicting CLNM by nomogram showed that the area under the curve (AUC) was 0.777 and exhibited an excellent consistency. Conclusions: A nomogram based on clinical and US imaging characteristics for predicting the probability of CLNM in patients with cN0 unilateral PTMC was developed, which showed a favorable predictive value and consistency. Further prospective research to observe the oncological outcomes is necessary to determine whether the nomogram could potentially guide a personalized surgical volume and surgical approach.