Comparison of ultrasound radiomics with conventional imaging models: diagnosis of central cervical lymph node metastasis in papillary thyroid carcinoma
10.3760/cma.j.issn.1004-4477.2019.10.011
- VernacularTitle: 超声影像组学与传统影像模式对甲状腺乳头状癌颈部中央区淋巴结转移的诊断价值比较
- Author:
Yunxia HUANG
1
;
Jin ZHOU
1
;
Tongtong LIU
2
,
3
;
Yi GUO
2
,
3
;
Yuyang TONG
1
;
Jinhua YU
2
,
3
;
Yuanyuan WANG
2
,
3
;
Min CHEN
1
;
Shichong ZHOU
1
;
Cai CHANG
1
Author Information
1. Department of Ultrasound, Fudan University Shanghai Cancer Center & Department of Oncology, Shanghai Medical College, Fudan University, Shanghai 200032, China
2. Department of Electronic Engineering, Fudan University, Shanghai 200433, China
3. Key Laboratory of Medical Imaging, Computing and Computer-Assisted Intervention, Shanghai 200433, China
- Publication Type:Clinical Trail
- Keywords:
Ultrasonography;
Radiomics;
Papillary thyroid carcinoma;
Lymph node metastasis;
Central neck;
Computer tomography
- From:
Chinese Journal of Ultrasonography
2019;28(10):882-887
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To compare the difference of diagnostic ability between ultrasound radiomics (USR) and different conventional imaging models of central neck (Ⅵ) lymph node metastasis in papillary thyroid carcinoma (PTC).
Methods:A training set of 609 cases was set up. USR features were extracted and screened by USR method. A weighted formula was established to calculate the USR score of each patient by ultrasound image. The USR score of the best diagnostic ability was obtained by statistical method and set as the diagnostic criterion. A test set of 326 cases was established to compare the diagnostic ability of USR score with ultrasound (US), computed tomography (CT) and US combined CT.
Results:The accuracy, sensitivity, specificity, area under ROC curve and Youden index of USR score in test set were 0.804, 0.867, 0.770, 0.766, 0.533, respectively, which were significantly higher than the corresponding values of US, CT and US combined CT(all P=0.000).
Conclusions:USR score obtained with USR method can effectively predict lymph node metastasis in Ⅵ region of PTC. The diagnostic efficiency and clinical value of USR score were significantly higher than those of conventional medical imaging models.