Ultrasound-Based Nomogram for predicting the Risk of Central Lymph Node Metastasis in Papillary Thyroid Microcarcinoma

Abstract

Abstract Background Although Prophylactic central compartment lymph node dissection (CLND) can reduce thyroid cancer recurrence, it can also cause surgical complications. Previous studies examining this issue have focused on papillary thyroid carcinoma rather than papillary thyroid microcarcinoma (PTMC) and lacked external validation, thus limiting their clinical applications. In this research, we retrospectively assessed independent predictors to produce a nomogram that can quantify the risk of central compartment lymph node metastasis (CLNM) in patients with PTMC to determine which patients should undergo CLND. Objective In this study, we intend to develop and validate a machine learning-based nomogram to predict the risk of central lymph node metastasis in papillary thyroid microcarcinoma and provide surgical recommendations to clinicians. Methods Patients with PTMC who received cervical lymph node biopsy at the Tangdu Hospital were included in the study sets. Demographic characteristics, ultrasonography results, and biochemical indicators were assessed. Multiple logistic regression was adopted as the basis for the nomogram. Concordance index (C-index), receiver operating characteristic (ROC) curve analysis, and decision curve analysis (DCA) were employed to evaluate the nomogram’s distinguishability, accuracy, and clinical availability. Results In our univariate logistic regression analysis, young age, large tumor size, calcification, aspect ratio ≥ 1, multifocality, indistinct lymphatic hilus, high free thyroxine (FT4), and low thyroid peroxidase antibody (TPOAb) were independent risk predictors for CLNM. Combining these predictors, the nomogram shows strong predictive capacity with C-index and accuracy of 0.784 and 0.713 in the training set and 0.779 and 0.709 in the validation set. DCA indicated that the nomogram had a well clinical application value. Conclusions We established a reliable, inexpensive, reproducible, and non-invasive preoperative prediction model that provides a potential tool for reducing the overtreatment of patients with PTMC.