Hu Hei1, Yongping Song2, Jianwu Qin3. 1. Department of Thyroid and Neck, The Affiliated Cancer Hospital of Zhengzhou University, Henan Cancer Hospital, Zhengzhou, 450003, China. 2. Department of Hematology, The Affiliated Cancer Hospital of Zhengzhou University, Henan Cancer Hospital, Zhengzhou, 450003, China. 3. Department of Thyroid and Neck, The Affiliated Cancer Hospital of Zhengzhou University, Henan Cancer Hospital, Zhengzhou, 450003, China. Electronic address: qinjianwu1962@163.com.
Abstract
INTRODUCTION: Much controversy exists over whether to perform lateral neck dissection (LND) on patients with papillary thyroid carcinoma (PTC). This study aimed to build predictive nomograms that could individually estimate lateral neck metastasis (LNM) risk and help determine follow up intensity. PATIENTS AND METHODS: Unifocal PTC patients who underwent LND between April 2012 and August 2014 were identified. Clinical and pathological variables were retrospectively evaluated using univariate and stepwise multivariate logistic regression analysis. Variables that had statistical significance in final multivariate logistic models were chosen to build nomograms, which were further corrected using the bootstrap resampling method. RESULTS: In all, 505 PTC patients were eligible for analysis. Among these, 178 patients (35.2%) had lateral neck metastasis. Two nomograms were generated: nomogram (c) and nomogram (c + p). Nomogram (c) incorporated four clinical variables: age, tumor size, tumor site, and extrathyroidal extension (ETE). It had a good discriminative ability, with a C-index of 0.79 (bootstrap-corrected, 0.78). Nomogram (c + p) incorporated two clinical variables and two pathological variables: tumor size, tumor site, extranodal extension (ENE), and number of positive nodes in the central compartment. Nomogram (c + p) showed an excellent discriminative ability, with a C-index of 0.86 (bootstrap-corrected, 0.85). CONCLUSION: Two predictive nomograms were generated. Nomogram (c) is a clinical model, whereas nomogram (c + p) is a clinicopathological model. Each nomogram incorporates only four variables and can give an accurate estimate of LNM risk in unifocal PTC patients, which may assist clinicians in patient counseling and decision making regarding LND.
INTRODUCTION: Much controversy exists over whether to perform lateral neck dissection (LND) on patients with papillary thyroid carcinoma (PTC). This study aimed to build predictive nomograms that could individually estimate lateral neck metastasis (LNM) risk and help determine follow up intensity. PATIENTS AND METHODS: Unifocal PTC patients who underwent LND between April 2012 and August 2014 were identified. Clinical and pathological variables were retrospectively evaluated using univariate and stepwise multivariate logistic regression analysis. Variables that had statistical significance in final multivariate logistic models were chosen to build nomograms, which were further corrected using the bootstrap resampling method. RESULTS: In all, 505 PTC patients were eligible for analysis. Among these, 178 patients (35.2%) had lateral neck metastasis. Two nomograms were generated: nomogram (c) and nomogram (c + p). Nomogram (c) incorporated four clinical variables: age, tumor size, tumor site, and extrathyroidal extension (ETE). It had a good discriminative ability, with a C-index of 0.79 (bootstrap-corrected, 0.78). Nomogram (c + p) incorporated two clinical variables and two pathological variables: tumor size, tumor site, extranodal extension (ENE), and number of positive nodes in the central compartment. Nomogram (c + p) showed an excellent discriminative ability, with a C-index of 0.86 (bootstrap-corrected, 0.85). CONCLUSION: Two predictive nomograms were generated. Nomogram (c) is a clinical model, whereas nomogram (c + p) is a clinicopathological model. Each nomogram incorporates only four variables and can give an accurate estimate of LNM risk in unifocal PTC patients, which may assist clinicians in patient counseling and decision making regarding LND.