Dataset Information

Differential diagnosis of thyroid nodules through a combination of multiple ultrasonography techniques: A decision-tree model.

ABSTRACT: The present study aimed to establish a decision tree (DT) model by combining the parameters of conventional gray-scale ultrasonography (US), elastosonography (ES), color Doppler US (CDUS) and contrast-enhanced US (CEUS) for the differential diagnosis of thyroid nodules. A single-center, retrospective study of 321 thyroid nodules was conducted. For 222 nodules, parameters of conventional gray-scale US, CDUS, ES and CEUS were evaluated using univariate logistic regression. Factors for with P<0.10 were further assessed using multivariate logistic regression. Significant factors (P<0.05) were used to establish a DT. The diagnostic accuracy of this DT was then evaluated by its application to the other 99 nodules. After univariate logistic analysis, factors including gender, number of nodules and diffuse disease were excluded, due to P>0.10. The results of multivariate logistic analysis determined that the following factors were required for the DT: Extent of blood flow determined by CDUS (P=0.002), area ratio determined by ES (P=0.033), peak phase patterns determined by CEUS (P<0.001) and micro-calcification determined by conventional gray-scale US (P=0.015). When compared to the pathological or cytological results of 99 nodules, the resulting DT had a sensitivity of 98.6%, specificity of 80.1%, positive predictive value of 93.5% and negative predictive value of 95.5%. These results suggested that a DT combining conventional gray-scale US, ES, CDUS and CEUS may be helpful for differentiating between types of thyroid nodules.

SUBMITTER: Luo W

PROVIDER: S-EPMC7185151 | biostudies-literature | 2020 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Differential diagnosis of thyroid nodules through a combination of multiple ultrasonography techniques: A decision-tree model.

Luo Wen W Zhang Yunfei Y Yuan Jiani J Yang Xiao X Pang Lina L Ding Lei L Zhang Peidi P Liu Liwen L Zhou Xiaodong X

Experimental and therapeutic medicine 20200323 6

The present study aimed to establish a decision tree (DT) model by combining the parameters of conventional gray-scale ultrasonography (US), elastosonography (ES), color Doppler US (CDUS) and contrast-enhanced US (CEUS) for the differential diagnosis of thyroid nodules. A single-center, retrospective study of 321 thyroid nodules was conducted. For 222 nodules, parameters of conventional gray-scale US, CDUS, ES and CEUS were evaluated using univariate logistic regression. Factors for with P<0.10 ...[more]

PMID: 32346431

Similar Datasets

Project description:ImportanceApproximately 20% of thyroid nodules display indeterminate cytology. Molecular testing can refine the risk of malignancy and reduce the need for diagnostic hemithyroidectomy.ObjectiveTo compare the diagnostic performance between an RNA test (Afirma genomic sequencing classifier) and DNA-RNA test (ThyroSeq v3 multigene genomic classifier).Design, setting, and participantsThis parallel randomized clinical trial of monthly block randomization included patients in the UCLA Health system who underwent thyroid biopsy from August 2017 to January 2020 with indeterminate cytology (Bethesda System for Reporting Thyroid Cytopathology category III or IV).InterventionsMolecular testing with the RNA test or DNA-RNA test.Main outcomes and measuresDiagnostic test performance of the RNA test compared with the DNA-RNA test. The secondary outcome was comparison of test performance with prior versions of the molecular tests.ResultsOf 2368 patients, 397 were eligible for inclusion based on indeterminate cytology, and 346 (median [interquartile range] age, 55 [44-67] years; 266 [76.9%] women) were randomized to 1 of the 2 tests. In the total cohort assessed for eligibility, 3140 thyroid nodules were assessed, and 427 (13.6%) nodules were cytologically indeterminate. The prevalence of malignancy was 20% among indeterminate nodules. The benign call rate was 53% (95% CI, 47%-61%) for the RNA test and 61% (95% CI, 53%-68%) for the DNA-RNA test. The specificities of the RNA test and DNA-RNA test were 80% (95% CI, 72%-86%) and 85% (95% CI, 77%-91%), respectively (P = .33); the positive predictive values (PPV) of the RNA test and DNA-RNA test were 53% (95% CI, 40%-67%) and 63% (95% CI, 48%-77%), respectively (P = .33). The RNA test exhibited a higher PPV compared with the prior test version (Afirma gene expression classifier) (54% [95% CI, 40%-67%] vs 38% [95% CI, 27%-48%]; P = .01). The DNA-RNA test had no statistically significant difference in PPV compared with its prior version (ThyroSeq v2 next-generation sequencing) (63% [95% CI, 48%-77%] vs 58% [95% CI, 43%-73%]; P = .75). Diagnostic thyroidectomy was avoided in 87 (51%) patients tested with the RNA test and 83 (49%) patients tested with the DNA-RNA test. Surveillance ultrasonography was available for 90 nodules, of which 85 (94%) remained stable over a median of 12 months follow-up.Conclusions and relevanceBoth the RNA test and DNA-RNA test displayed high specificity and allowed 49% of patients with indeterminate nodules to avoid diagnostic surgery. Although previous trials demonstrated that the prior version of the DNA-RNA test was more specific than the prior version of the RNA test, the current molecular test techniques have no statistically significant difference in performance.Trial registrationClinicalTrials.gov Identifier: NCT02681328.

Project description:Importance:Thyroid nodules are common incidental findings. Ultrasonography and molecular testing can be used to assess risk of malignant neoplasm. Objective:To examine whether a model developed through automated machine learning can stratify thyroid nodules as high or low genetic risk by ultrasonography imaging alone compared with stratification by molecular testing for high- and low-risk mutations. Design, Setting, and Participants:This diagnostic study was conducted at a single tertiary care urban academic institution and included patients (n?=?121) who underwent ultrasonography and molecular testing for thyroid nodules from January 1, 2017, through August 1, 2018. Nodules were classified as high risk or low risk on the basis of results of an institutional molecular testing panel for thyroid risk genes. All thyroid nodules that underwent genetic sequencing for cytological results with Bethesda System categories III and IV were reviewed. Patients without diagnostic ultrasonographic images within 6 months of fine-needle aspiration or who received definitive treatment at an outside medical center were excluded. Main Outcomes and Measures:Thyroid nodules were categorized by the model as high risk or low risk using ultrasonographic images. Results were compared using genetic testing. Results:Among the 134 lesions identified in 121 patients (mean [SD] age, 55.7 [14.2] years; 102 women [84.3%]), 683 diagnostic ultrasonographic images were selected. Of the 683 images, 556 (81.4%) were used for training the model, 74 (10.8%) for validation, and 53 (7.8%) for testing. Most nodules had no mutation (75 [56.0%]), whereas 43 nodules (32.1%) had a high-risk mutation and 16 (11.9%) had an unknown or a low-risk mutation (?2?=?39.060; P?<?.001). In total, 228 images (33.4%) were of nodules classified as genetically high risk (n?=?43), and 455 (66.6%) were of low-risk nodules (n?=?91). The model performed with a sensitivity of 45% (95% CI,?23.1%-68.5%), a specificity of 97% (95% CI,?84.2%-99.9%), a positive predictive value of 90% (95% CI,?55.2%-98.5%), a negative predictive value of 74.4% (95% CI,?66.1%-81.3%), and an overall accuracy of 77.4% (95% CI,?63.8%-97.7%). Conclusions and Relevance:The study found that the model developed through automated machine learning could produce high specificity for identifying nodules with high-risk mutations on molecular testing. This finding shows promise for the diagnostic applications of machine learning interpretation of sonographic imaging of indeterminate thyroid nodules.

Project description:BACKGROUND:Currently the cytological examination of fine needle aspiration (FNA) biopsies is the standard technique for the pre-operative differential diagnosis of thyroid nodules. However, the results may be non-informative in ~20% of cases due to an inadequate sampling and the lack of highly specific, measurable cytological criteria, therefore ancillary biomarkers that could aid in these cases are clearly needed. The aim of our study was to evaluate the mRNA expression levels of 8 candidate marker genes as the diagnostic biomarkers for the discrimination of benign and malignant thyroid nodules and to find a combination of biomarkers with the highest diagnostic value. MATERIALS AND METHODS:mRNA expression levels of eight candidate marker genes - BIRC5, CCND1, CDH1, CITED1, DPP4, LGALS3, MET and TFF3 was measured by real-time RT-PCR in paired nodular and surrounding normal thyroid tissue specimens of 105 consecutive patients undergoing thyroid surgery and compared between different types of thyroid lesions. RESULTS:Significant differences in the mRNA expression levels between the normal and malignant thyroid tissues and between benign and malignant nodules were found for BIRC5, CCND1, CITED1, DPP4, LGALS3, MET and TFF3, but not CDH1. On a single gene basis, relative quantity (RQ) of LGALS3 had the highest diagnostic value for the discrimination of malignant and benign thyroid nodules (AUC = 0.832, P < 0.0001 and 90.9% sensitivity and 65.6% specificity at the optimal cut-off on ROC curve). The only two-marker set that outperformed LGALS3 was RQ sum of LGALS3 and BIRC5 (AUC = 0.841, P < 0.0001). An application of multivariate logistic regression analysis resulted in the generation of a multiplex biomarker model based on LGALS3, BIRC5, TFF3, CCND1, MET and CITED1 that had considerably higher specificity than a single marker or two marker gene-based models (AUC = 0.895, P < 0.0001, 70.5% sensitivity and 93.4% specificity). CONCLUSIONS:This study confirmed that mRNA expression levels of 7 out of 8 candidate genes analysed have a diagnostic value for the distinction of benign and malignant thyroid nodules. The multiplex biomarker model based on 6 genes outperformed a single marker or two marker-based models and warrants feasibility studies on FNA biopsies and the validation in a larger cohort of patients.

Dataset Information

Differential diagnosis of thyroid nodules through a combination of multiple ultrasonography techniques: A decision-tree model.

Publications

Differential diagnosis of thyroid nodules through a combination of multiple ultrasonography techniques: A decision-tree model.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets