Affordable Access

deepdyve-link
Publisher Website

Preoperative prediction of papillary thyroid microcarcinoma via multiparameter ultrasound.

Authors
  • Ma, Hui Juan1
  • Yang, Jing Chun1
  • Leng, Zhen Peng1
  • Chang, Ying1
  • Kang, Hua2
  • Teng, Liang Hong3
  • 1 1 Department of Ultrasound, Xuanwu Hospital, Capital Medical University, Beijing, PR China. , (China)
  • 2 2 Department of Surgery, Xuanwu Hospital, Capital Medical University, Beijing, PR China. , (China)
  • 3 3 Department of Pathology, Xuanwu Hospital, Capital Medical University, Beijing, PR China. , (China)
Type
Published Article
Journal
Acta radiologica (Stockholm, Sweden : 1987)
Publication Date
Nov 01, 2017
Volume
58
Issue
11
Pages
1303–1311
Identifiers
DOI: 10.1177/0284185117692167
PMID: 28194993
Source
Medline
Keywords
License
Unknown

Abstract

Background Accurate diagnosis of papillary thyroid microcarcinoma (PTMC) is important for further management. Ultrasound (US) is the most frequently used imaging modality for PTMC. Purpose To evaluate the diagnostic value of conventional US, contrast-enhanced ultrasound (CEUS) and real-time elastography (RTE) for patients with PTMC. Material and Methods In total, 135 patients with subcentimeter thyroid nodules who underwent conventional US, CEUS, and RTE before surgery were enrolled. A multivariate logistic regression analysis was performed to assess the independent predictors of PTMC. The diagnostic performances of conventional US, CEUS, and RTE were evaluated with a receiver operating characteristic (ROC) curve analysis. Results A taller-than-wide shape was identified as the strongest predictor of PTMC (odds ratio [OR], 25.21), followed by heterogeneous enhancement (OR, 24.03), marked hypoechogenicity (OR, 21.71), poorly defined margin (OR, 5.51), strain ratio (OR, 2.59), and age (OR, 0.92; all P values < 0.05). Heterogeneous enhancement on CEUS showed the highest positive predictive value (PPV; 88.0%) and an accuracy of 83.7%. A logistic regression model was created to predict PTMC using conventional US, CEUS, and RTE. The area under the ROC curve was 0.97, with a sensitivity of 88.6% and a specificity of 94.6%. Conclusion Conventional US combined with CEUS and RTE can improve the diagnostic accuracy of PTMC.

Report this publication

Statistics

Seen <100 times