Search citation statements
Paper Sections
Citation Types
Year Published
Publication Types
Relationship
Authors
Journals
Objectives. Analysis of possibilities of radiomics as a source of additional diagnostic information about the structural maturity of the lungsMaterials and methods. A retrospective study included 72 pregnant women: 35 with congenital fetal diaphragmatic hernia (group 1) and 37 without fetal lung pathology (group 2). Frontal or co-frontal T2 images (T2 FSE) were obtained. Segmentation of regions of interest at the fetal lung level was performed manually with ITK-Snap. A total of 107 radiomic features were extracted using pyradiomics. The statistical analysis was performed using the STATISTICA 10 statistical analysis package (USA) to detect correlation between trait values and the target variable (presence of lung pathology in CDH), and to show differences in the comparison groups according to the detected parameters.Results. Statistically significant features were identified for 2D and 3D segmentations (p < 0.05). For 2D and 3D segmentations, the number of significant features was 14 and 73, respectively. After exclusion of features with cross-correlations, their number decreased to 6 and 8 for single slices and 3D images, respectively. Correlation coefficients between the features and the presence of lung pathology were also calculated. In the case of 3D images, the number of features with significant correlation coefficients (r > 0.4, p < 0.05) equaled 20, while for single-slice images this number was 3.Conclusion. The data obtained allow to conclude that it is reasonable to use texture analysis of the 3D MRI images as a source of additional diagnostic information concerning the structural maturity of the lungs.
Objectives. Analysis of possibilities of radiomics as a source of additional diagnostic information about the structural maturity of the lungsMaterials and methods. A retrospective study included 72 pregnant women: 35 with congenital fetal diaphragmatic hernia (group 1) and 37 without fetal lung pathology (group 2). Frontal or co-frontal T2 images (T2 FSE) were obtained. Segmentation of regions of interest at the fetal lung level was performed manually with ITK-Snap. A total of 107 radiomic features were extracted using pyradiomics. The statistical analysis was performed using the STATISTICA 10 statistical analysis package (USA) to detect correlation between trait values and the target variable (presence of lung pathology in CDH), and to show differences in the comparison groups according to the detected parameters.Results. Statistically significant features were identified for 2D and 3D segmentations (p < 0.05). For 2D and 3D segmentations, the number of significant features was 14 and 73, respectively. After exclusion of features with cross-correlations, their number decreased to 6 and 8 for single slices and 3D images, respectively. Correlation coefficients between the features and the presence of lung pathology were also calculated. In the case of 3D images, the number of features with significant correlation coefficients (r > 0.4, p < 0.05) equaled 20, while for single-slice images this number was 3.Conclusion. The data obtained allow to conclude that it is reasonable to use texture analysis of the 3D MRI images as a source of additional diagnostic information concerning the structural maturity of the lungs.
Background: Breast cancer (BC) occupies a leading position among my oncological diseases detected in women. Identification and search for predictors of malignant neoplasms using radiation and molecular genetic methods of research allows timely diagnosis and treatment, which improves the prognosis for breast cancer. Purpose: To identify a correlation between the molecular subtype of a breast cancer tumor at an early clinical stage and the patterns of the mammographic method. Methods: A prospective, single-center study of 363 patients diagnosed with breast cancer followed up during 2021. X-ray mammography in two projections, ultrasound-guided trephine biopsy for histological verification, and immunohistochemical (IHC) analysis to determine molecular subtypes were performed. Results: There were statistically significant differences in age between subtypes luminal A, luminal BHER2+ (p < 0.001) and triple negative (p = 0.037), luminal B, luminal BHER2+ (p = 0.001) and triple negative (p = 0.046), luminal BHER2+ and nonluminal HER2+ (p = 0.002), between nonluminal HER2+ and triple negative subtype (p = 0.034). When comparing the structure of radiological density, statistically significant differences were revealed between the subgroups luminal B, luminal BHER2+ (p = 0.010) and triple negative (p = 0.010), between luminal A and triple negative subtypes (p = 0.010). When comparing the leading mammographic symptom (p < 0.001), radiological contours of the formation (p < 0.001), the density of pathological changes (p < 0.001), the size, the newly detected pathological process (p < 0.001) statistically significant differences were also found in the subgroups. A division into groups according to the size of pathological changes within the biotypes was noted, where the aggressive phenotypes of the triple negative subtype (p = 0.001), non-luminal HER2+ (p = 0.02) and luminal B (p = 0.02), in contrast to luminal A, were manifested by a greater extent. the maximum linear size of the tumor. A symptom of nipple retraction (p = 0.048) was described, which was not characteristic of triple negative and non-luminal HER2 cancer. Conclusions: Visualization features of differences in the radiological manifestation of breast cancer of different biological subtypes up to 20 mm can be predictors of molecular subtypes. Pathological verification and IHC study remain a mandatory study, but it may be necessary to conduct an X-ray histological correlation before starting treatment and, if obvious discrepancies are detected, repeat the IHC analysis from the surgical material.
Artificial intellect (AI) is a complex of technological solutions that allows simulating human cognitive functions (including self-learning and finding solutions without a predetermined algorithm) and obtaining, when performing specific tasks, results comparable, at least, to the results of human intellectual activity. The most promising area of AI in medicine, in which technologies have achieved obvious success and are already being actively applied, is the analysis of diagnostic images (X-ray, MRI, CT, PET, SPECT): detection, recognition and identification of various pathologies on them. The purpose of the review is to guide the reader in the potential and problem of AI technologies in radiation diagnosis of human pathologies. The review covers articles that meet the following criteria: the publication is an original research article; the publication is devoted to radiation diagnostics; the publications analyze the use of AI technology in solving the clinical problems of diagnosis, prognosis of a particular pathology; radiology is a potential or actual field of the AI technology application analyzed in the publication. The problem of data verification and marking, radiomics and radiogenomics (as a basis for machine and deep learning of AI), the use of AI in hepatology, limitations and risks of AI application in medicine are considered.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.