AB011. Quantification of thymic epithelial tumor volumes through deep learning
Ontology highlight
ABSTRACT:
SUBMITTER: Lay N
PROVIDER: S-EPMC10774927 | biostudies-literature | 2023 Jan
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Background Rare tumors are diagnostic challenges for pathologists. Thymic epithelial tumors (TETs) are heterogenous and their treatment strategies vary according to histological subgroup. Previous work has shown that a second pathological opinion may result in a change in diagnosis for more than half of cases, with a potential treatment shift in 44%. The aim of this study is to assess the feasibility of using artificial intelligence and deep learning to classify TETs, which could be used to help improve pathologist diagnostic consistency for these challenging tumors. Methods Digital diagnostic hematoxylin and eosin (H&E) stained slides of tumors for 103 patients with thymoma type A, AB, B1, B2, and B3 were downloaded from The Cancer Genome Atlas (TCGA). An Xception-based deep convolutional neural network model was trained on slide images at 10× magnification to predict histologic subtype as an ordinal variable in three-fold cross-validation. Hyperparameters were taken from previously published experiments, and no additional hyperparameter tuning was performed to reduce the risk of overfitting. Validation predictions from each cross-fold were aggregated and compared between groups using analysis of variance (ANOVA) and one-sided t-tests with Bonferroni correction for multiple comparisons. Model activations at the post-convolutional layer for validation images in the first cross-fold were visualized with uniform manifold approximation and projection (UMAP) dimensionality reduction to better understand the spatial relationship between learned image features. Results Deep learning predictions among the TET subtypes were significantly different by ANOVA (P<0.0001) and correlated with the ordinal labels (R-squared =0.39). Thymoma A and AB subtypes were distinguished from both B1 and B2/B3 (P=0.023 and <0.001, respectively), and B1 tumors were distinguished from B2/B3 (P=0.011). Analysis of post-convolutional layer activations revealed an axis of transition through the ordinal variables, providing evidence that the deep learning model learned image features on a morphologic spectrum. Conclusions This is the first example in TETs that deep learning can discriminate between TET histologic subtypes using digital H&E slides. We aim to further validate the algorithm with a multi-institution dataset from centers of expertise to improve the ability to distinguish thymoma subtypes.
| S-EPMC9792820 | biostudies-literature
Project description:Background: Because obesity is associated with the risk of posttransplant diabetes mellitus (PTDM), the precise estimation of visceral fat mass before transplantation may be helpful. Herein, we addressed whether a deep-learning based volumetric fat quantification on pretransplant computed tomographic images predicted the risk of PTDM more precisely than body mass index (BMI). Methods: We retrospectively included a total of 718 nondiabetic kidney recipients who underwent pretransplant abdominal computed tomography. The 2D (waist) and 3D (waist or abdominal) volumes of visceral, subcutaneous, and total fat masses were automatically quantified using the deep neural network. The predictability of the PTDM risk was estimated using a multivariate Cox model and compared among the fat parameters using the areas under the receiver operating characteristic curves (AUROCs). Results: PTDM occurred in 179 patients (24.9%) during the median follow-up period of 5 years (interquartile range, 2.5-8.6 years). All the fat parameters predicted the risk of PTDM, but the visceral and total fat volumes from 2D and 3D evaluations had higher AUROC values than BMI did, and the best predictor of PTDM was the 3D abdominal visceral fat volumes [AUROC, 0.688 (0.636-0.741)]. The addition of the 3D abdominal VF volume to the model with clinical risk factors increased the predictability of PTDM, but BMI did not. Conclusions: A deep-learning based quantification of visceral fat volumes on computed tomographic images better predicts the risk of PTDM after kidney transplantation than BMI.
| S-EPMC8185023 | biostudies-literature
Project description:BackgroundDynamic contrast-enhanced (DCE) MRI is widely used to assess vascular perfusion and permeability in cancer. In small animal applications, conventional modeling of pharmacokinetic (PK) parameters from DCE MRI images is complex and time consuming. This study is aimed at developing a deep learning approach to fully automate the generation of kinetic parameter maps, Ktrans (volume transfer coefficient) and Vp (blood plasma volume ratio), as a potential surrogate to conventional PK modeling in mouse brain tumor models based on DCE MRI.MethodsUsing a 7T MRI, DCE MRI was conducted in U87 glioma xenografts growing orthotopically in nude mice. Vascular permeability Ktrans and Vp maps were generated using the classical Tofts model as well as the extended-Tofts model. These vascular permeability maps were then processed as target images to a twenty-four layer convolutional neural network (CNN). The CNN was trained on T1-weighted DCE images as source images and designed with parallel dual pathways to capture multiscale features. Furthermore, we performed a transfer study of this glioma trained CNN on a breast cancer brain metastasis (BCBM) mouse model to assess the potential of the network for alternative brain tumors.ResultsOur data showed a good match for both Ktrans and Vp maps generated between the target PK parameter maps and the respective CNN maps for gliomas. Pixel-by-pixel analysis revealed intratumoral heterogeneous permeability, which was consistent between the CNN and PK models. The utility of the deep learning approach was further demonstrated in the transfer study of BCBM.ConclusionsBecause of its rapid and accurate estimation of vascular PK parameters directly from the DCE dynamic images without complex mathematical modeling, the deep learning approach can serve as an efficient tool to assess tumor vascular permeability to facilitate small animal brain tumor research.
| S-EPMC9048985 | biostudies-literature
Project description:BackgroundEnlarged deep medullary veins (EDMVs) in patients with Sturge-Weber syndrome (SWS) may channel venous blood from the surface to the deep vein system in brain regions affected by the leptomeningeal venous malformation. Thus, the quantification of EDMV volume may provide an objective imaging marker for this vascular compensatory process. The present study proposes a novel analytical method to quantify enlarged EDMV volumes in the affected hemisphere of patients with unilateral SWS.MethodsTwenty young subjects, including 10 patients with unilateral SWS and 10 healthy siblings (age 14.5±6.7 and 16.0±7.0 years, respectively) underwent 3T brain MRI scanning using susceptibility-weighted imaging (SWI) and volumetric T1-weighted sequences. The proposed image analytic steps segmented EDMVs in white matter regions, defined on the volumetric T1-weighted images, by statistically associating the likelihood of intensity, location, and tubular shape on SWI. The volumes of the segmented EDMVs, calculated in each hemisphere, were compared between affected and unaffected hemispheres. EDMV volumes were also correlated with visually assessed EDMV scores, hemispheric white matter volumes, and cortical surface areas. Parametric tests including Pearson's correlation, unpaired and paired t-tests, were used. A P value <0.05 was considered statistically significant.ResultsIt was found that EDMVs were identified well in SWS-affected hemispheres while calcified regions were excluded. Mean EDMV volumes in the SWS-affected hemispheres were 10-12-fold greater than in the unaffected or healthy control hemispheres; while white matter volumes and cortical surface areas were lower. EDMV volumes in the SWS-affected hemispheres showed a strong positive correlation with the visual EDMV scores (r=0.88, P=0.001) and an inverse correlation with cortical surface area ratios (r=-0.65, P=0.04) but no correlation with white matter volume ratios.ConclusionsEDMVs were detected in the SWS-affected atrophic hemispheres reliably while avoiding calcified regions. The approach can be used to quantify enlarged deep cerebral veins in the human brain, which may provide a potential marker of cerebral venous remodeling.
| S-EPMC10895099 | biostudies-literature
Project description:To investigate the performance of a deep learning-based algorithm for fully automated quantification of left ventricular (LV) volumes and function in cardiac MRI. We retrospectively analysed MR examinations of 50 patients (74% men, median age 57 years). The most common indications were known or suspected ischemic heart disease, cardiomyopathies or myocarditis. Fully automated analysis of LV volumes and function was performed using a deep learning-based algorithm. The analysis was subsequently corrected by a senior cardiovascular radiologist. Manual volumetric analysis was performed by two radiology trainees. Volumetric results were compared using Bland-Altman statistics and intra-class correlation coefficient. The frequency of clinically relevant differences was analysed using re-classification rates. The fully automated volumetric analysis was completed in a median of 8 s. With expert review and corrections, the analysis required a median of 110 s. Median time required for manual analysis was 3.5 min for a cardiovascular imaging fellow and 9 min for a radiology resident (p < 0.0001 for all comparisons). The correlation between fully automated results and expert-corrected results was very strong with intra-class correlation coefficients of 0.998 for end-diastolic volume, 0.997 for end-systolic volume, 0.899 for stroke volume, 0.972 for ejection fraction and 0.991 for myocardial mass (all p < 0.001). Clinically meaningful differences between fully automated and expert corrected results occurred in 18% of cases, comparable to the rate between the two manual readers (20%). Deep learning-based fully automated analysis of LV volumes and function is feasible, time-efficient and highly accurate. Clinically relevant corrections are required in a minority of cases.
| S-EPMC7568707 | biostudies-literature
Project description:More than 50 human diseases are characterized by the deposition of specific protein aggregates in the form of insoluble amyloid fibrils. However, only a very small number of proteins are known to form amyloids with high propensity, limiting our ability to understand, predict and engineer amyloid aggregation from sequence. Here we use a massively parallel assay to quantify the amyloid nucleation propensity of >100,000 random 20 amino acid sequences. Approximately 5% of assayed random sequences nucleate the formation of aggregates, generating a very large and diverse training dataset from which to train models to predict amyloid nucleation. We use this dataset to train CANYA, a convolution-attention hybrid neural network that predicts the propensity of any primary sequence to form amyloids. CANYA outperforms previous predictors of protein aggregation on additional random sequences and out-of-sample datasets including human disease-causing amyloids, with very stable performance across diverse prediction tasks. We adapt and extend recent advances in interpretability of genomic neural networks to elucidate CANYA’s decision-making process and learned grammar and to provide mechanistic insights into amyloid formation. Our results demonstrate the power of massive experimental random sequence-space exploration and provide an interpretable and robust neural network model for understanding, predicting and designing amyloid-forming proteins.
2024-07-17 | GSE268261 | GEO
Project description:ObjectivesThis study aimed to distinguish preoperatively anterior mediastinal thymic cysts from thymic epithelial tumors via a computed tomography (CT)-based radiomics nomogram.MethodsThis study analyzed 74 samples of thymic cysts and 116 samples of thymic epithelial tumors as confirmed by pathology examination that were collected from January 2014 to December 2020. Among the patients, 151 cases (scanned at CT 1) were selected as the training cohort, and 39 cases (scanned at CT 2 and 3) served as the validation cohort. Radiomics features were extracted from pre-contrast CT images. Key features were selected by SelectKBest and least absolute shrinkage and selection operator and then used to build a radiomics signature (Rad-score). The radiomics nomogram developed herein via multivariate logistic regression analysis incorporated clinical factors, conventional CT findings, and Rad-score. Its performance in distinguishing the samples of thymic cysts from those of thymic epithelial tumors was assessed via discrimination, calibration curve, and decision curve analysis (DCA).ResultsThe radiomics nomogram, which incorporated 16 radiomics features and 3 conventional CT findings, including lesion edge, lobulation, and CT value, performed better than Rad-score, conventional CT model, and the clinical judgment by radiologists in distinguishing thymic cysts from thymic epithelial tumors. The area under the receiver operating characteristic (ROC) curve of the nomogram was 0.980 [95% confidence interval (CI), 0.963-0.993] in the training cohort and 0.992 (95% CI, 0.969-1.000) in the validation cohort. The calibration curve and the results of DCA indicated that the nomogram has good consistency and valuable clinical utility.ConclusionThe CT-based radiomics nomogram presented herein may serve as an effective and convenient tool for differentiating thymic cysts from thymic epithelial tumors. Thus, it may aid in clinical decision-making.
| S-EPMC8702557 | biostudies-literature
Project description:Thymic epithelial tumors (TETs) comprise a rare group of thoracic cancers, classified as thymomas and thymic carcinomas (TC). To date, chemotherapy is still the standard treatment for advanced disease. Unfortunately, few therapeutic options are available for relapsed/refractory tumors. Unlike other solid cancers, the development of targeted biologic and/or immunologic therapies in TETs remains in its nascent stages. Moreover, since the thymus plays a key role in the development of immune tolerance, thymic tumors have a unique biology, which can confer susceptibility to autoimmune diseases and ultimately influence the risk-benefit balance of immunotherapy, especially for patients with thymoma. Indeed, early results from single-arm studies have shown interesting clinical activity, albeit at a cost of a higher incidence of immune-related side effects. The lack of knowledge of the immune mechanisms associated with TETs and the absence of biomarkers predictive of response or toxicity to immunotherapy risk limiting the evolution of immunotherapeutic strategies for managing these rare tumors. The aim of this review is to summarize the existing literature about the thymus's immune biology and its association with autoimmune paraneoplastic diseases, as well as the results of the available studies with immune checkpoint inhibitors and cancer vaccines.
| S-EPMC10705681 | biostudies-literature
Project description:BackgroundThymic epithelial tumors (TETs) are rare tumors originating from the thymic epithelial cells. SOX9, a member of the family of SOX (SRY-related high-mobility group box) genes, has been considered as an oncogene and therapeutic target in various cancers. However, its role in TETs remains uncertain.MethodsUsing the immunohistochemistry method, the expression of SOX9 was analyzed in TETs tissues, including 34 thymoma (8 cases with type A, 6 with type AB, 6 with type B1, 9 with type B2, and 5 with type B3 thymomas) and 20 thymic cancer tissues and the clinicopathologic and prognostic significances were evaluated. Further bioinformatics analysis of gene expression profiles of thymomas with high and low SOX9 expressions and the corresponding survival analyses were based on the thymoma cases identified in The Cancer Genome Atlas (TCGA) database, with the median expression level of SOX9 selected as cutoff.ResultsImmunohistochemistry staining showed that SOX9 was highly expressed in the nuclei of the epithelial cells of the Hassall's corpuscles and of the TET tumor cells. SOX9 expression was significantly associated with histological type and high expression indicated unfavorable clinical outcomes of thymomas. Bioinformatics analysis revealed that genes positively associated with SOX9 expression were mapped in proteoglycans in cancer, cell adhesion molecules, and molecules involved in extracellular matrix-receptor interaction and the TGF-β signaling pathway, and that genes negatively associated with SOX9 expression were mapped in molecules involved in primary immunodeficiency, the T cell receptor signaling pathway, Th17 cell differentiation, PD-L1 expression, and the PD-1 checkpoint pathway in cancer. In addition, SOX9 expression was positively associated with POU2F3 and TRPM5 expressions, the master regulators of tuft cells, suggesting that high SOX9 expression might be associated with the tuft cell phenotype of thymomas. Moreover, high SOX9 expression was associated with immune dysregulation of thymoma, and M2 macrophage significantly dominated in the high SOX9 expression group.ConclusionSOX9 may serve as a diagnostic and prognostic marker for TETs. Notably, high SOX9 expression in TETs may indicate a tuft cell phenotype and an immune suppressive microenvironment of thymomas.
| S-EPMC8580949 | biostudies-literature
Project description:The parameters of activatedg sludge are crucial for the daily operation of wastewater treatment plants (WWTPs). In particular, mixed liquor suspended solids (MLSS) and apparent viscosity provide metrics for the biomass and rheological properties of activated sludge. Traditional methods for determining these parameters are time-consuming, require separate measurements for each index, and fail to provide real-time data for future 'smart' WWTPs. Here we show a real-time online microscopic image data analysis system that quantitatively identifies MLSS and apparent viscosity. Microscopic videos of activated sludge are captured in lab-scale sequencing batch reactors under chemical oxygen demand shock, yielding 41482 high-quality images. The Xception convolutional neural network architecture is used to establish both qualitative and quantitative correlations between these microscopic images and MLSS/apparent viscosity. The accuracies of qualitative identification for MLSS and apparent viscosity are both higher than 97%, and the quantitative correlation coefficients are 0.95 and 0.96, respectively. This quantitative correlation between microscopic images of activated sludge and its physical parameters, specifically MLSS and apparent viscosity, provides a basis for real-time online measurements of activated sludge parameters in WWTPs.
| S-EPMC11905822 | biostudies-literature