ABSTRACT
Purpose: To develop a deep learning-based risk stratification system for thyroid nodules using US cine images. Materials and Methods: In this retrospective study, 192 biopsy-confirmed thyroid nodules (175 benign, 17 malignant) in 167 unique patients (mean age, 56 years ± 16 [SD], 137 women) undergoing cine US between April 2017 and May 2018 with American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS)-structured radiology reports were evaluated. A deep learning-based system that exploits the cine images obtained during three-dimensional volumetric thyroid scans and outputs malignancy risk was developed and compared, using fivefold cross-validation, against a two-dimensional (2D) deep learning-based model (Static-2DCNN), a radiomics-based model using cine images (Cine-Radiomics), and the ACR TI-RADS level, with histopathologic diagnosis as ground truth. The system was used to revise the ACR TI-RADS recommendation, and its diagnostic performance was compared against the original ACR TI-RADS. Results: The system achieved higher average area under the receiver operating characteristic curve (AUC, 0.88) than Static-2DCNN (0.72, P = .03) and tended toward higher average AUC than Cine-Radiomics (0.78, P = .16) and ACR TI-RADS level (0.80, P = .21). The system downgraded recommendations for 92 benign and two malignant nodules and upgraded none. The revised recommendation achieved higher specificity (139 of 175, 79.4%) than the original ACR TI-RADS (47 of 175, 26.9%; P < .001), with no difference in sensitivity (12 of 17, 71% and 14 of 17, 82%, respectively; P = .63). Conclusion: The risk stratification system using US cine images had higher diagnostic performance than prior models and improved specificity of ACR TI-RADS when used to revise ACR TI-RADS recommendation.Keywords: Neural Networks, US, Abdomen/GI, Head/Neck, Thyroid, Computer Applications-3D, Oncology, Diagnosis, Supervised Learning, Transfer Learning, Convolutional Neural Network (CNN) Supplemental material is available for this article. © RSNA, 2022.
ABSTRACT
INTRODUCTION: The adult congenital heart disease population with repaired tetralogy of Fallot (TOF) is subject to chronic volume and pressure loading leading to a 40% probability of right ventricular (RV) failure by the 3rd decade of life. We sought to identify a non-invasive signature of adverse RV remodeling using peripheral blood microRNA (miRNA) profiling to better understand the mechanisms of RV failure. METHODS: Demographic, clinical data, and blood samples were collected from adults with repaired TOF (N = 20). RNA was isolated from the buffy coat of peripheral blood and whole genome miRNA expression was profiled using Agilent's global miRNA microarray platform. Fold change, pathway analysis, and unbiased hierarchical clustering of miRNA expression was performed and correlated to RV size and function assessed by echocardiography performed at or near the time of blood collection. RESULTS: MiRNA expression was profiled in the following groups: 1. normal RV size (N = 4), 2. mild/moderate RV enlargement (N = 11) and 3. severe RV enlargement (N = 5). 267 miRNAs were downregulated, and 66 were upregulated across the three groups (fold change >2.0, FDR corrected p<0.05) as RV enlargement increased and systolic function decreased. qPCR validation of a subset of these miRNAs identified increasing expression of miRNA 28-3p, 433-3p, and 371b-3p to be associated with increasing RV size and decreasing RV systolic function. Unbiased hierarchical clustering of all patients based on miRNA expression demonstrates three distinct patient clusters that largely coincide with progressive RV enlargement. Pathway analysis of dysregulated miRNAs demonstrates up and downregulation of cell cycle pathways, extracellular matrix proteins and fatty acid synthesis. HIF 1α signaling was downregulated while p53 signaling was predicted to be upregulated. CONCLUSION: Adults with TOF have a distinct miRNA profile with progressive RV enlargement and dysfunction implicating cell cycle dysregulation and upregulation in extracellular matrix and fatty acid metabolism. These data suggest peripheral blood miRNA can provide insight into the mechanisms of RV failure and can potentially be used for monitoring disease progression and to develop RV specific therapeutics to prevent RV failure in TOF.
