Haolin Shen

Background: Early detection of breast cancer and accurate assessment of lesions are key goals of imaging evaluation. Ultrasound is widely used, but its diagnostic performance is influenced by complex image features, noise, and operator experience. Reducing operator dependence and improving accuracy are critical clinical issues. Methods: In this retrospective study, 7,025 breast ultrasound images from our center were annotated based on pathology and split into training, validation, and internal test sets (8:1:1). The Dataset of Breast Ultrasound Images was used as the external test set. YOLO-v7 and YOLO-v8 models were trained through transfer learning after data augmentation and balancing the classes. Performance was compared on internal and external test sets and was evaluated against a reader study. Results: YOLO-v7 and YOLO-v8 reached optimal performance at epochs 294 and 135, respectively. YOLO-v7 slightly outperformed YOLO-v8 on the internal test set, while YOLO-v8 achieved higher accuracy, recall, specificity, precision, and F1 score on the external test set. Both models showed significantly higher accuracy, specificity, and precision than the senior radiologist, with YOLO-v8 achieving a significantly higher F1 score. Discussion: YOLO-v8 demonstrated better generalization due to its anchor-free mechanism and deeper architecture, while YOLO-v7 showed signs of overfitting. Both models outperformed the junior radiologist and approached or exceeded the diagnostic performance of the senior radiologist, indicating potential to assist less experienced readers. Conclusion: YOLO-v7 and YOLO-v8 effectively classified breast lesions. YOLO-v8 showed faster convergence and higher diagnostic efficiency, suggesting strong potential for clinical application.

This study aimed to develop an explainable machine learning framework integrating dual-modality ultrasonography and thyroid function parameters for preoperative prediction of central lymph node metastasis (CLNM) in capsular-invasive papillary thyroid carcinoma. A retrospective cohort of 382 pathologically confirmed capsular-invasive papillary thyroid carcinoma patients was stratified into CLNM-negative and CLNM-positive cohorts. After comprehensive univariate and multivariate logistic regression analyses, predictive models were developed using 8 machine learning algorithms (including Logistic Regression, Support Vector Machine, Gradient Boosting Machine, eXtreme Gradient Boosting, K-Nearest Neighbors, Adaptive Boosting, Neural Network, and Categorical Boosting [CatBoost]) and rigorously validated through receiver operating characteristic analysis. Multivariate analysis showed irregular margins, tumor location in lower/mid poles, maximum diameter > 10 mm, rich blood supply, heterogeneous enhancement, and elevated thyroid-stimulating hormone were independent CLNM risk factors. Receiver operating characteristic curves demonstrated the CatBoost model achieved optimal performance (training area under the curve: 0.791; test area under the curve: 0.804). SHapley Additive exPlanations analysis revealed maximum diameter > 10 mm, tumor location in lower/mid poles, and irregular margins were the top 3 contributing features. Tumor size > 10 mm is the most important predictor of CLNM. The CatBoost model demonstrated superior performance and, combined with SHapley Additive exPlanations analysis, provides a clinically applicable tool for personalized surgical planning by identifying high-risk patients who may benefit from prophylactic central lymph node dissection.

Amid rising diabetes prevalence, the burden of diabetes-related chronic kidney disease (DKD) is escalating, particularly in China. This study aimed to describe the epidemiological trends of DKD in China and develop a Bayesian age–period–cohort (BAPC) model to predict its incidence from 2022 to 2036. Data from the Global Burden of Disease Study 2021 (GBD 2021) were used to assess DKD’s burden in China between 1990 and 2021, including age-standardized incidence (ASIR), prevalence (ASPR), mortality (ASMR), and disability-adjusted life years (DALYs). BAPC models were applied to evaluate age, period, and cohort effects and to project future trends. Between 1990 and 2021, the ASIR for chronic kidney disease due to diabetes mellitus type 1 (DKD1) decreased (annual percent change (APC) = -0.57, P < 0.001), while ASPR increased (APC = 1.25, P < 0.001). ASMR and DALYs both declined (APC = -1.98, P < 0.001; APC = -1.99, P < 0.001). For chronic kidney disease due to diabetes mellitus type 2 (DKD2), ASIR slightly increased (APC = 0.44, P < 0.001), but both ASPR and ASMR decreased (APC = -0.23, P = 0.004; APC = -0.56, P < 0.001). Projections suggest an increase in DKD2 incidence and declining incidence for DKD1 by 2036. This study provides insights into the trends of DKD burden in China over the past three decades. The BAPC offers a reliable forecast for the next 15 years, aiding future research to reduce DKD’s impact in China.

To investigate the ability of an auxiliary diagnostic model based on the YOLO-v7-based model in the classification of cervical lymphadenopathy images and compare its performance against qualitative visual evaluation by experienced radiologists. Three types of lymph nodes were sampled randomly but not uniformly. The dataset was randomly divided into for training, validation, and testing. The model was constructed with PyTorch. It was trained and weighting parameters were tuned on the validation set. Diagnostic performance was compared with that of the radiologists on the testing set. The mAP of the model was 96.4% at the 50% intersection-over-union threshold. The accuracy values of it were 0.962 for benign lymph nodes, 0.982 for lymphomas, and 0.960 for metastatic lymph nodes. The precision values of it were 0.928 for benign lymph nodes, 0.975 for lymphomas, and 0.927 for metastatic lymph nodes. The accuracy values of radiologists were 0.659 for benign lymph nodes, 0.836 for lymphomas, and 0.580 for metastatic lymph nodes. The precision values of radiologists were 0.478 for benign lymph nodes, 0.329 for lymphomas, and 0.596 for metastatic lymph nodes. The model effectively classifies lymphadenopathies from ultrasound images and outperforms qualitative visual evaluation by experienced radiologists in differential diagnosis.

The aim of this study was to develop a prognostic model for radioactive iodine (RAI) therapy outcome in patients with Graves‘ disease. We enrolled 127 patients. Information on RAI therapy, ultrasound indexes of thyroid, and other lifestyle factors was collected. The competing risk model was used to estimate the multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for nonhealing or recurrence of hyperthyroidism (NHRH). The performance of the model was assessed by receiver operator characteristic analysis and the Brier score and internally validated by bootstrap resampling. Then, a nomogram was developed. Forty-one cases (32.2%) of NHRH were documented. Positive Ki-67 expression, a higher dose of per-unit thyroid volume, and females showed lower risks of NHRH (all P < 0.05). The HR values (95% CI) were 0.42 (0.23, 0.79), 0.01 (0.00, 0.02), and 0.47 (0.25, 0.89), respectively. The bootstrap validation showed that the model had the highest accuracy and good calibration for predicting cumulative risk of NHRH at 180 days after RAI therapy (AUC = 0.772; 95% CI: 0.640–0.889, Brier score = 0.153). By decision curve analysis, the nomogram was shown to have a satisfactory net benefit between thresholds of 0.20 and 0.40. Ki-67, ultrasound volumetry, and scintigraphy techniques can play important roles in evaluating RAI therapy outcome in Graves‘ disease patients. The prediction nomogram shows reasonable accuracy in predicting NHRH.

Objective: To construct risk prediction models for cervical lymph node metastasis (CLNM) of papillary thyroid carcinoma (PTC) under different thyroid disease backgrounds and to analyze and compare risk factors among different groups.Methods: This retrospective study included 518 patients with PTC that was pathologically confirmed post-operatively from January 2021 to November 2021. Demographic, ultrasound and pathological data were recorded. Univariate and multivariate logistic regression analyses were performed to identify factors associated with CLNM in the whole patient cohort and in patients grouped according to diagnoses of Hashimoto’s thyroiditis (HT), nodular goiter (NG), and no background disease. Prediction models were constructed for each group, and their performances were compared.Results: Analysis of the whole PTC patient cohort identified NG as independently associated with CLNM. The independent risk factors for patients with no background disease were the maximum thyroid nodule diameter and American College of Radiology Thyroid Imaging Reporting & Data System score; those for patients with HT were the maximum thyroid nodule diameter, ACR TI-RADS score, and multifocality; and those for patients with NG were the maximum thyroid nodule diameter, ACR TI-RADS score, multifocality and gender.Conclusion: Background thyroid disease impacts CLNM in PTC patients, and risk factors for CLNM vary among PTC patients with different background diseases. Ultrasound is useful for diagnosing background thyroid disease, which can inform treatment planning. Different prediction models are recommended for PTC cases with different thyroid diseases.

Objective: Compared thyroid volumes measured by 2-D and 3-D US with those of resected specimens and proposed new models to improve measurement accuracy. Methods: This study included 80 patients who underwent total thyroidectomy. One 2D_model and one 3D_model were developed using piecewise linear regression analysis. The accuracy of these models was compared using an ellipsoid model (2-D_US value × 0.5), 3-D_US value, and Ying’s model [1.76 + (2-D_US value × 0.38)]. Results: The new 2D_model was: V=2.66 + (0.71 * X1) - (1.51 * X2). In this model, if 2-D_US value <= 228.39, X1 = 2-D_US value and X2 = 0; otherwise, X1 = 2-D_US value and X2 = 2-D_US value - 228.39. The 3D_model was: V= 2.90 + (1.08 * X1) + (2.43 * X2). In this model, if 3-D_US value <= 102.06, X1 = 3-D_US value and X2 = 0; otherwise, X1 = 3-D_US value and X2 = 3-D_US value - 102.06. The accuracy of the new models was higher than that of the 3-D_US value, the ellipsoid model, and Ying’s model (P<0.05). Conclusion: The models established are more accurate than the traditional ones and can accurately measure thyroid volume.

The objective of this study is to develop and validate the performance of 2 ultrasound (US) feature–guided machine learning models in distinguishing cervical lymphadenopathy. We enrolled 705 patients whose US characteristics of lymph nodes were collected at our hospital. B-mode US and color Doppler US features of cervical lymph nodes in both cohorts were analyzed by 2 radiologists. The decision tree and back propagation (BP) neural network were developed by combining clinical data (age, sex, and history of tumor) and US features. The performance of the 2 models was evaluated by calculating the area under the receiver operating characteristics curve (AUC), accuracy value, precision value, recall value, and balanced F score (F1 score). The AUC of the decision tree and BP model in the modeling cohort were 0.796 (0.757, 0.835) and 0.854 (0.756, 0.952), respectively. The AUC, accuracy value, precision value, recall value, and F1 score of the decision tree in the validation cohort were all higher than those of the BP model: 0.817 (0.786, 0.848) vs 0.674 (0.601, 0.747), 0.774 (0.737, 0.811) vs 0.702 (0.629, 0.775), 0.786 (0.739, 0.833) vs 0.644 (0.568, 0.720), 0.733 (0.694, 0.772) vs 0.630 (0.542, 0.718), and 0.750 (0.705, 0.795) vs 0.627 (0.541, 0.713), respectively. The US feature–guided decision tree model was more efficient in the diagnosis of cervical lymphadenopathy than the BP model.