[Radiomics for prediction of central lymph node metastasis in the neck in patients with thyroid papillary carcinoma]

Nan Fang Yi Ke Da Xue Xue Bao. 2019 Sep 30;39(9):1094-1098. doi: 10.12122/j.issn.1673-4254.2019.09.15.
[Article in Chinese]

Abstract

Objective: To explore the feasibility of radiomics for predicting lymph node metastasis in the central region of the neck in patients with thyroid papillary carcinoma (PTC).

Methods: A total of 189 patients with PTC confirmed by thyroid fine needle aspiration biopsy were prospectively enrolled in this study. The cross-sectional and longitudinal ultrasound images and the images of both sections were analyzed for predicting central lymph node metastasis using a radiomics approach with pathological results as the gold standard.

Results: In the 189 patients, the accuracy, sensitivity and specificity of preoperative thyroid ultrasonography for diagnosis of central lymph node metastasis was 69.39%, 64% and 73%, respectively. Based on the ultrasound images of the cross-sections, longitudinal sections and both sections, the accuracy, sensitivity and specificity of radiomics for predicting central lymph node metastasis was 66.06%/68.12%/77.69%, 53%/46%/40%, and 52%/53%/51%, respectively.

Conclusions: Radiomics with combined analysis of the ultrasound images on the cross-section and longitudinal section images achieves a higher accuracy for predicting central lymph node metastasis than analysis a single section, and its diagnostic accuracy is much higher than that of conventional ultrasound examination.

目的: 探讨影像组学方法提取甲状腺乳头状癌(PTC)的纹理特征以预判颈部中央区淋巴结转移的可能性。

方法: 前瞻性收集本院接受甲状腺细针穿刺活检证实为PTC的189例患者,常规超声检查有无中央区淋巴结转移;然后基于其PTC的超声横切面、纵切面及双切面的纹理特征,利用随机森林方法对纹理特征进行筛选并建立预测模型,预测颈部中央区淋巴结是否发生转移,并以最终手术病理结果为金标准进行比对。

结果: 189例患者中,术前甲状腺超声检查诊断颈部中央区淋巴结转移的准确率为35%,灵敏度为28.6%,特异度为55%。影像组学分析PTC的超声横切面/纵切面/双切面预测颈部中央区淋巴结转移的准确率为66.06%/68.12%/77.69%,灵敏度为53%/46%/40%,特异度为52%/53%/51%。

结论: 结合PTC的超声横及纵双切面纹理特征,影像组学方法预测颈部中央区淋巴结转移的准确性较单切面高,且远高于常规超声检查的准确率。

Keywords: central lymph node metastasis; papillary thyroid carcinoma; radiomics.

MeSH terms

  • Carcinoma, Papillary / diagnostic imaging*
  • Carcinoma, Papillary / pathology
  • Humans
  • Image Interpretation, Computer-Assisted
  • Lymph Nodes
  • Lymphatic Metastasis / diagnostic imaging*
  • Neck
  • Prospective Studies
  • Thyroid Neoplasms / diagnostic imaging*
  • Thyroid Neoplasms / pathology
  • Ultrasonography

Grants and funding

国家自然科学基金(81671709)