AUTHOR=Zhang Heng , Hu Shudong , Wang Xian , Liu Wenhua , He Junlin , Sun Zongqiong , Ge Yuxi , Dou Weiqiang 

TITLE=Using Diffusion-Weighted MRI to Predict Central Lymph Node Metastasis in Papillary Thyroid Carcinoma: A Feasibility Study

JOURNAL=Frontiers in Endocrinology

VOLUME=Volume 11 - 2020

YEAR=2020

URL=https://www.frontiersin.org/journals/endocrinology/articles/10.3389/fendo.2020.00326

DOI=10.3389/fendo.2020.00326

ISSN=1664-2392

ABSTRACT=Objective: To investigate whether diffusion-weighted MRI (DWI) with multi b values can be used as a quantitative assessment tool to predict central lymph node metastasis (CLNM) in papillary thyroid carcinoma (PTC). Method: PTCs patients were enrolled from January 2015 to April 2018. Each patient was measured with multi b value DWI (300, 500, and 800 s/mm2) preoperative and then underwent clinical treatment of central lymph node dissection at the thyroid surgery department. These patients with and without CLNM were divided as two groups. The apparent diffusion coefficients (ADCs) were evaluated for three different b values (b=300 s/mm2, b=500 s/mm2 and b=800 s/mm2). Clinicopathological variables and ADC values were retrospectively analysis using univariate and binary logistic regression analysis. Variables that had statistical significance in final multivariate logistic models were chosen to build nomogram, which were then further corrected using the bootstrap resampling method. Results: PTCs with CLNM had significantly lower ADC300, ADC500 and ADC800 values than without CLNM. The area under the curve (AUC) of the ADC500 value (0.817) was higher than those of the ADC300 and ADC800 values (0.610 and 0.641, respectively) in differentiating CLNM from without CLNM group. The cut-off value of ADC500 to discriminate PTCs with and without CLNM was determined at 1.444 × 10-3 mm2/s, with sensitivity of 88.6% and specificity of 66%. Nomogram was generated by Binary logistic regression results, incorporating four variables: Gender, Primary tumor size, extrathyroidal extension (ETE), and ADC500 value. The AUC of the nomogram was 0.894 in predicting CLNM from without CLNM group. Moreover, a calibration curve showed a good agreement in CLNM prediction between nomogram and actual clinical observation. Conclusions: In conclusion, the ADC value is a noninvasive and valuable imaging biomarker for evaluating CLNM in PTCs. The nomogram as a clinical predictive model can give an accurate preoperative evaluation of CLNM risk in PTCs patients. This may assist clinicians in patient diagnosis and decision making regarding CLNM.