YE Fengying, LI Shangqing, SU Qichen, et al. Comparative analysis of computer-aided diagnosis system in the diagnosis of benign and malignant thyroid nodules based on different guidelines[J]. Journal of Clinical Ultrasound in Medicine, 2020,22(9):694-696.
YE Fengying, LI Shangqing, SU Qichen, et al. Comparative analysis of computer-aided diagnosis system in the diagnosis of benign and malignant thyroid nodules based on different guidelines[J]. Journal of Clinical Ultrasound in Medicine, 2020,22(9):694-696. DOI: 10.16245/j.cnki.issn1008-6978.2020.09.019.
Objective To compare the diagnostic efficacy of computer-aided diagnosis(CAD)system based on different guidelines for the diagnosis of benign and malignant thyroid nodules.Methods A multicentric prospective study was conducted in 453 patients with 521 thyroid nodules
which were confirmed by pathology of surgery or ultrasound guided fine needle aspiration.Taken pathology as the gold standard
the receiver operating characteristic(ROC)curve of CAD system was obtained based on different guidelines(ACR guideline
ATA guideline
AACE guideline
KSThR consensus
BTA guideline
Kwak guideline
Seo guideline
Russ guideline).The area under the curves(AUC)was calculated and the diagnostic efficiency was analyzed.Results When the cut-off value was 2 scores
the sensitivity
negative predictive value and AUC of KSThR consensus were 90.60%
84.1% and 0.759
respectively
which were higher than those of the other seven guidelines(all P<0.01).When the cut-off value was grade 4
the specificity and positive predictive value of BTA guidelines in the diagnosis of thyroid malignant nodules were 79.95% and 70.9%
respectively
which were higher than the other seven guidelines(all P<0.01).Conclusion The diagnostic efficay of CAD system in the diagnosis of benign and malignant thyroid nodules is satisfying.The KSThR consensus has the highest AUC and sensitivity
and the BTA guidelines has the highest specificity.
关键词
超声检查计算机辅助诊断甲状腺结节良恶性诊断效能
Keywords
UltrasonographyComputer-aided systemThyroid nodulebenign and malignantDiagnostic efficacy