ObjectiveTo investigate the value of ultrasound gray-scale ratio (UGSR) for the differential diagnosis of papillary thyroid microcarcinoma (PTMC) and micronodular goiter (MNG) in two medical centers.MethodsUltrasound images of 881 PTMCs from 785 patients and 744 MNGs from 687 patients in center A were retrospectively analyzed and compared with 243 PTMCs from 203 patients and 251 MNGs from 198 patients in center B. All cases were confirmed by surgery and histology. The grayscale values of thyroid lesions and surrounding normal tissues were measured, and the UGSR was calculated. The optimal UGSR threshold for identifying PTMCs and MNGs in two medical centers was determined by receiver operating characteristic (ROC) curve, and the area under the curve (AUC), optimal UGSR threshold, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were compared between the two medical centers.ResultsThe UGSR values of PTMCs and MNGs in medical center A were 0.5537 (0.4699, 0.6515) and 0.8708 (0.7616, 1.0123) (Z = -27.691, P = 0), respectively, whereas those in medical center B were 0.5517 (0.4698, 0.6377) and 0.8539 (0.7366, 0.9929) (Z = -16.057, P = 0), respectively. The UGSR of PTMCs and MNGs did not differ significantly between the two medical centers (Z = -0.609, P = 0.543 and Z = -1.394, P = 0.163, respectively). The AUC, optimal UGSR threshold, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of the two medical centers were 0.898 vs. 0.918, 0.7214 vs. 0.6911, 0.881 vs. 0.868, 0.817 vs. 0.833, 0.851 vs. 0.834, 0.853 vs. 0.867, and 0.852 vs. 0.850, respectively.ConclusionsUGSR can quantify the echo intensity of PTMCs and MNGs and is therefore valuable for the differential diagnosis of the two diseases. The diagnostic efficacy was consistent between the two medical centers. This method should be widely promoted and applied.

This study was to investigate the diagnostic value of the computed tomography (CT) histogram in thyroid benign solitary coarse calcification nodules (BSCNs). A total of 89 thyroid solitary coarse calcification nodules (coarse calcification ≥5 mm, no definite soft tissue around calcification) confirmed either by surgery or histopathological examination in 86 cases enrolled from January 2009 to December 2015 were evaluated. These included 56 BSCNs and 33 malignant solitary coarse calcification nodules (MSCNs). Overall, 27 cut-off values were calculated by N (4≤N≤30) times of 50 Hounsfield units (HU) in the range of 200 to 1500 HU, and each cut-off value and the differences in the corresponding area percentages in the CT histogram were recorded for BSCN and MSCN. The optimal cut-off value and the corresponding area percentage were established by receiver operating characteristic (ROC) curve analysis. In the 19 groups with an ROC area under curve (AUC) of more than 0.7, at a cut-off value of 800 HU and at an area percentage of no more than 93.8%, the ROC AUC reached the maximum of 0.79, and the accuracy, sensitivity, and specificity were 75.3%, 80.4%, and 66.7%, respectively. At a cut-off value of 1050 HU and at an area percentage of no more than 93.6%, the accuracy, sensitivity, and specificity were 71.9%, 60.7%, and 90.9%, respectively. At a cut-off of 1150 HU and area of no more than 98.4%, the accuracy, sensitivity, and specificity were 70.8%, 57.1%, and 93.9%, respectively. At a cut-off of 600 HU and area of no more than 12.1%, the accuracy, sensitivity, and specificity were 61.8%, 39.3%, and 100.0%, respectively. Compared with the cut-off value of 800 HU and an area percentage of no more than 93.8%, the sensitivity of cut-off values and minimum areas of 1050 HU and 93.6%, of 1150 HU and 98.4%, and of 600 HU and 12.1%, was gradually decreasing; however, their specificity was gradually increasing. This can provide an important basis for reducing the misdiagnosis and unnecessary surgical trauma.

Papillary thyroid cancer (PTC) accounts for more than 80% of thyroid cancers, and ultrasound (US) imaging is the preferred method for the diagnosis of PTC. However, accurate prediction of different patterns of cervical lymph node metastasis (CLNM) in PTC continues to be a challenge. Herein, US images and clinical factors of PTC patients from three hospitals for more than 11 years are collected, and a multimodal deep learning model called DeepThy‐Net is then developed to predict different CLNM patterns. The proposed model not only uses the convolutional features extracted by deep learning but also integrates traditional clinical factors that are highly related to lymph node metastasis. Finally, the model is tested in two independent test sets, and the experimental results show that the area under curve (AUC) is between 0.870 and 0.905, indicating clinical applicability. The proposed method provides an important reference for the treatment and management of PTC. Moreover, for PTC cases involving an active surveillance strategy, the proposed method can serve as an important CLNM early warning tool.

No abstract

Objectives: Coarse calcifications are prone to cause echo attenuation during ultrasonography (US) and hence affect the classification of benign and malignant nodules. This study aimed to investigate the diagnostic role of computed tomography (CT) for differentiating the American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS) 4-5 nodules with coarse calcifications. Methods: CT data of 216 ACR TI-RADS 4-5 nodules with coarse calcifications confirmed by surgery and pathology in 207 patients were analyzed retrospectively. Halo sign, artifacts, and CT values (i.e., Hounsfield unit) of the nodules were determined by two radiologists. Univariate analysis and binary logistic regression were used to determine the relationship of halo sign, artifact, and CT value with benign nodules. A predictive model for benign nodules with coarse calcifications was then constructed. The receiver operating characteristic (ROC) curve was used to analyze the predictive value of halo sign, artifact, CT value, and logistic regression model. Results: Of the 216 ACR TI-RADS 4-5 nodules with coarse calcifications, 170 were benign and 46 were malignant. There were 92 benign and 7 malignant nodules with halo sign (χ 2 = 22.067, P < 0.001), and 79 benign and 10 malignant nodules with artifacts (χ 2 = 9.140, P < 0.001). The CT values of benign and malignant nodules were 791 (543-1,025) Hu and 486 (406-670) Hu, respectively (Z = −5.394, P < 0.001). Binary logistic regression demonstrated that the halo sign, artifact, and CT value were independent predictors for benign nodules with coarse calcifications. The area under the ROC curve (AUC) of halo sign, artifact, CT value and regression model for predicting benign nodules with coarse calcifications were 0.776, 0.711, 0.784, and 0.850, respectively, and the optimal threshold of CT value was 627.5 Hu. Conclusion: Halo sign, artifact, and CT value > 627.5 Hu were helpful for identifying ACR TI-RADS 4-5 thyroid benign nodules with coarse calcifications. Wei et al. CT Diagnose Thyroid Calcified Nodules The diagnostic performance of the logistic regression model was higher than that of any single indicator. Accurate identification of these indicators could identify benign nodules and reduce unnecessary surgical trauma.

Background Cervical lymph node (LN) status is a critical factor related to the treatment and prognosis of papillary thyroid carcinoma (PTC). The aim of this study was to investigate the preoperative predictions of cervical LN metastasis in PTC using computed tomography (CT) radiomics.Methods A total of 134 PTC patients who underwent CT examinations were enrolled in the study at two institutes between January 2018 and January 2020. Of these patients, 289 cervical LNs (institute 1: 206 LNs from 88 patients; institute 2: 83 LNs from 46 patents) were selected. All the cases had been confirmed by surgery and pathology. Each LN was segmented and 1408 radiomic features were calculated radiomic features in noncontrast and contrast-enhanced CT images. Features were selected using the Boruta algorithm followed by an iterative culling-out algorithm. We compared four machine learning classifiers, including random forest (RF), support vector machine (SVM), neural network (NN), and naïve bayes (NB) for the classification of LN metastasis. The models were first trained and validated by 10-fold cross-validation using data from institute 1 and then tested using independent data from institute 2. The performance of the models was compared using the area under the receiver operating characteristic curves (AUC).Results Seven radiomic features were selected for building the models − 3 histogram statistical textures, 1 gray level co-occurrence matrix texture, and 3 gray level zone size matrix textures. The AUCs of the radiomic models with 10-fold cross-validation were 0.941 (95% confidence interval [CI]: 0.93–0.95), 0.943 (95% CI: 0.93–0.95), 0.914 (95% CI: 0.90–0.95), and 0.905 (95% CI: 0.88–0.91) for RF, SVM, NN, and NB, respectively. The AUCs for the testing data were 0.926 (95% CI: 0.86–0.98), 0.932 (95% CI: 0.88–0.98), 0.925 (95% CI: 0.86–0.97), and 0.912 (95% CI: 0.83–0.98) for RF, SVM, NN, and NB, respectively.Conclusions CT radiomic model demonstrated robustness in preoperative classification of LN metastases for patients with PTC, which may provide significant support for clinical decision making and prognosis evaluation.

Background The present study aimed to quantify and differentiate the echo levels of papillary thyroid microcarcinomas (PTMCs) and micronodular goiters (MNGs) using the ultrasound grayscale ratio (UGSR) and to investigate the repeatability of UGSR. Methods The ultrasound (US) data of 241 patients with 265 PTMCs and 141 patients with 168 MNGs confirmed by surgery and pathology were retrospectively analyzed. All patients had received outpatient ultrasonic examination and preoperative ultrasonic positioning. The RADinfo radiograph reading system was used to measure the grayscales of PTMC, MNG, and thyroid tissues at the same gain level, and the UGSR values of the PTMC, MNG, and thyroid tissue were calculated. The patients were divided into outpatient examination, preoperative positioning, and mean value groups, and the receiver operating characteristic (ROC) curves were calculated to obtain the optimal UGSR threshold to distinguish PTMC from MNG. The interclass correlation coefficient (ICC) was used to assess the consistency of UGSR measured in three groups. Results The UGSR values of the PTMC and MNG were 0.56 ± 0.14 and 0.80 ± 0.19 (t = 5.84, P < 0.001) in the outpatient examination group, 0.55 ± 0.14 and 0.80 ± 0.19 (t = 18.74, P < 0.001) in the preoperative positioning group, and 0.56 ± 0.12 and 0.80 ± 0.18 (t = 16.49, P < 0.001) in the mean value group. The areas under the ROC curves in the three groups were 0.860, 0.856, and 0.875, respectively. When the UGSR values for the outpatient examination, preoperative positioning, and mean value groups were 0.649, 0.646, and 0.657, respectively, each group obtained its largest Youden index. A reliable UGSR value was obtained between the outpatient examination and preoperative positioning groups (ICC = 0.79, P = 0.68). Conclusion UGSR is a simple and repeatable method to distinguish PTMC from MNG, and hence, can be widely applicable.

To investigate the value of the signal intensity on T2-weighted magnetic resonance (MR) imaging using quantitative analysis in the differentiation of parotid tumors. Materials and Methods: MR data of 80 pleomorphic adenomas (PAs), 68 Warthin tumors (WTs), and 34 malignant tumors (MTs) confirmed by surgery and histology were retrospectively analyzed. The signal intensities of tumor, normal parotid gland, spinal cord, and buccal subcutaneous fat were measured, and the signal intensity ratios (SIRs) between the tumor and the three references were calculated. Receiver operating characteristic curve was used to determine the optimal threshold and diagnostic efficiency of SIR for differentiating PAs, WTs, and MTs. Results: The area under the curve (AUC) of tumor to parotid gland SIR (SIR P ), tumor to spinal cord SIR (SIR C ), and tumor to buccal subcutaneous fat SIR (SIR F ) for differentiating PAs and WTs was 0.922, 0.918, and 0.934, respectively. The sensitivity and specificity at an optimal SIR threshold were 86.3% and 91.2%, 80.0% and 97.1%, and 85.0% and 94.1%, respectively. The AUC of SIR P , SIR C , and SIR F for distinguishing PAs from MTs was 0.793, 0.802, and 0.774, respectively. The sensitivity and specificity at an optimal SIR threshold was 86.3% and 61.8%, 80.0% and 73.5%, and 82.5% and 73.5%, respectively. The AUC of SIR P , SIR C , and SIR F for distinguishing WTs from MTs was 0.716, 0.709, and 0.759, respectively. The sensitivity and specificity at an optimal SIR threshold were 61.8% and 82.4%, 55.9% and 82.4%, and 64.7% and 86.8%, respectively. Conclusion: SIR P , SIR C , and SIR F on T2-weighted MR images had high diagnostic efficiency for differentiating between PAs and WTs, while SIR P and SIR C for differentiating between PAs and MTs, and SIR F for differentiating between WTs and MTs had relatively high diagnostic efficiency.