Research Papers:

CUNR scoring system for the prediction of lateral lymph node metastasis in papillary thyroid carcinoma

Jianyong Lei, Gengpeng Li, Zhihui Li, RX Rong and Jingqiang Zhu _

PDF  |  HTML  |  How to cite

Oncotarget. 2018; 9:167-177. https://doi.org/10.18632/oncotarget.22772

Metrics: PDF 1430 views  |   HTML 1964 views  |   ?  


Jianyong Lei1, Gengpeng Li1, Zhihui Li1, RX Rong1 and Jingqiang Zhu1

1Thyroid and Parathyroid Surgery Center, West China Hospital of Sichuan University, Chengdu 610041, China

Correspondence to:

Jingqiang Zhu, email: [email protected]

Keywords: papillary; thyroid carcinoma; lateral; lymph node; metastasis

Received: July 19, 2017     Accepted: November 14, 2017     Published: November 30, 2017


Objective: Our present study aimed to evaluate and compare the number and rate of central lymph node metastases (LNMs) for the prediction of lateral LNM (LLNM) in papillary thyroid carcinoma (PTC) and to develop a scoring system.

Results: Capsule invasion, tumor location in the upper portion of the thyroid, an ipsilateral central compartment LNM number ≥3, and an ipsilateral central compartment LNM rate of ≥56% were identified as significant independent predictors of ipsilateral lateral LNM in PTC. The predictive ability of an ipsilateral central compartment LNM rate ≥56% (area under the curve (AUC) = 0.802) was better than that of an ipsilateral central compartment LNM number ≥3 (AUC = 0.755). The ROC curves identified the best index point (CUNR) to distinguish the presence or absence of ipsilateral LLNM as 11, which has a high sensitivity (0.860) and a low false-negative rate (0.100, 1-Specificity). These findings were supported by the validation cohort.

Conclusions: Patients with a CUNR index point equal to or greater than 11 and ipsilateral lateral lymph node dissection should be considered for a diagnosis of LLNM.

Patients and Methods: A total of 1,281 PTC patients were included and divided into two groups: those with a presence of LLNM (n = 222) and those with an absence of LLNM (n = 1059). Univariate and multivariate analyses were performed to detect the risk factors for LLNM, and receiver operating characteristic (ROC) curves were used to detect the best cutoff values of these predictors. Additionally, a scoring system for the odds ratio (OR) of independent factors was developed and validated in an independent cohort of PTC patients (n = 560).

Creative Commons License All site content, except where otherwise noted, is licensed under a Creative Commons Attribution 4.0 License.
PII: 22772