Search

Background: Differentiating seminomas from nonseminomas is crucial for formulating optimal treatment strategies for testicular germ cell tumors (TGCTs). Therefore, our study aimed to develop and validate a clinical-radiomics model for this purpose., Methods: In this study, 221 patients with TGCTs confirmed by pathology from four hospitals were enrolled and classified into training (n = 126), internal validation (n = 55) and external test (n = 40) cohorts. Radiomics features were extracted from the CT images. After feature selection, we constructed a clinical model, radiomics models and clinical-radiomics model with different machine learning algorithms. The top-performing model was chosen utilizing receiver operating characteristic (ROC) curve analysis. Decision curve analysis (DCA) was also conducted to assess its practical utility., Results: Compared with those of the clinical and radiomics models, the clinical-radiomics model demonstrated the highest discriminatory ability, with AUCs of 0.918 (95 % CI: 0.870 - 0.966), 0.909 (95 % CI: 0.829 - 0.988) and 0.839 (95 % CI: 0.709 - 0.968) in the training, validation and test cohorts, respectively. Moreover, DCA confirmed that the combined model had a greater net benefit in predicting seminomas and nonseminomas., Conclusion: The clinical-radiomics model serves as a potential tool for noninvasive differentiation between testicular seminomas and nonseminomas, offering valuable guidance for clinical treatment., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier B.V. All rights reserved.)

Soft sensors have become reliable tools for estimating difficult-to-measure target variables in modern industrial processes. In order to make full use of labeled and unlabeled samples, an active semi-supervised soft sensor modeling method is proposed, which combines active learning and semi-supervised learning to maximize model performance and minimize the laboratory analysis cost of expanding the labeled sample data set. First, manifold regularization is introduced into the deep extreme learning machine (DELM) algorithm to form a semi-supervised DELM that improves the performance of a model trained with unlabeled samples. Then, considering non-Gaussian processes and the error information between the predicted and true values, an active sample selection strategy based on error Gaussian mixture model is developed. Using this strategy, the most uncertain and representative unlabeled samples are selected for labeling, and thereby expanding the labeled sample data set. Finally, the effectiveness of the proposed method is verified using industrial debutanizer process data. [ABSTRACT FROM AUTHOR]