Nomogram for Preoperative Estimation of Cervical Lymph Node Metastasis Risk in Papillary Thyroid Microcarcinoma.

Objective: Accurate preoperative identification of cervical lymph node metastasis (CLNM) is essential for clinical management and established of different surgical protocol for patients with papillary thyroid microcarcinoma (PTMC). Herein, we aimed to develop an ultrasound (US) features and clinical characteristics-based nomogram for preoperative diagnosis of CLNM for PTMC. Method: Our study included 552 patients who were pathologically diagnosed with PTMC between January 2015 and June 2019. All patients underwent total thyroidectomy or lobectomy and divided into two groups: CLNM and non-CLNM. Univariate and multivariate analysis were performed to examine risk factors associated with CLNM. A nomogram comprising the prognostic model to predict the CLNM was established, and internal validation in the cohort was performed. Results: CLNM and non-CLNM were observed in 216(39.1%) and 336(60.9%) cases, respectively. Seven variables of clinical and US features as potential predictors including male sex (odd ratio [OR] = 1.974, 95% confidence interval [CI], 1.243-2.774; P =0.004), age < 45 years (OR = 4.621, 95% CI, 2.160-9.347; P < 0.001), US-reported CLN status (OR = 1.894, 95% CI, 0.754-3.347; P =0.005), multifocality (OR = 1.793, 95% CI, 0.774-2.649; P =0.007), tumor size ≥ 0.6cm (OR = 1.731, 95% CI,0.793-3.852; P =0.018), ETE (OR = 3.772, 95% CI, 1.752-8.441;P< 0.001) and microcalcification (OR = 2.316, 95% CI, 1.099-4.964; P < 0.001) were taken into account. The predictive nomogram was established by involving all the factors above used for preoperative prediction of CLNM in patients with PTCM. The nomogram model showed an AUC of 0.839 and an accuracy of 77.9% in predicting CLNM. Furthermore, the calibration curve demonstrated a strong consistency between nomogram and clinical findings in prediction CLNM for PTMC. Conclusions: The nomogram achieved promising results for predicting preoperative CLNM in PTMC by combining clinical and US risk factor. Our proposed prediction model is able to help determine an individual's risk of CLNM in PTMC, thus facilitate reasonable therapy decision making.