scholarly journals Multimodal 3D DenseNet for IDH Genotype Prediction in Gliomas

Genes ◽  
2018 ◽  
Vol 9 (8) ◽  
pp. 382 ◽  
Author(s):  
Sen Liang ◽  
Rongguo Zhang ◽  
Dayang Liang ◽  
Tianci Song ◽  
Tao Ai ◽  
...  
Keyword(s):  
Deep Learning ◽  
Clinical Decision Making ◽  
Three Dimensional ◽  
Area Under The Curve ◽  
Image Data ◽  
Clinical Decision ◽  
The Cancer Genome Atlas ◽  
World Health ◽  
Validation Dataset ◽  
Who Grade

Non-invasive prediction of isocitrate dehydrogenase (IDH) genotype plays an important role in tumor glioma diagnosis and prognosis. Recently, research has shown that radiology images can be a potential tool for genotype prediction, and fusion of multi-modality data by deep learning methods can further provide complementary information to enhance prediction accuracy. However, it still does not have an effective deep learning architecture to predict IDH genotype with three-dimensional (3D) multimodal medical images. In this paper, we proposed a novel multimodal 3D DenseNet (M3D-DenseNet) model to predict IDH genotypes with multimodal magnetic resonance imaging (MRI) data. To evaluate its performance, we conducted experiments on the BRATS-2017 and The Cancer Genome Atlas breast invasive carcinoma (TCGA-BRCA) dataset to get image data as input and gene mutation information as the target, respectively. We achieved 84.6% accuracy (area under the curve (AUC) = 85.7%) on the validation dataset. To evaluate its generalizability, we applied transfer learning techniques to predict World Health Organization (WHO) grade status, which also achieved a high accuracy of 91.4% (AUC = 94.8%) on validation dataset. With the properties of automatic feature extraction, and effective and high generalizability, M3D-DenseNet can serve as a useful method for other multimodal radiogenomics problems and has the potential to be applied in clinical decision making.

scholarly journals SGPNet: A Three-Dimensional Multitask Residual Framework for Segmentation and IDH Genotype Prediction of Gliomas

2021 ◽  
Vol 2021 ◽  
pp. 1-9
Author(s):  
Yao Wang ◽  
Yan Wang ◽  
Chunjie Guo ◽  
Shuangquan Zhang ◽  
Lili Yang
Keyword(s):  
Brain Tumor ◽  
Clinical Decision Making ◽  
Three Dimensional ◽  
Clinical Decision ◽  
Error Rates ◽  
The Cancer Genome Atlas ◽  
Classification Error ◽  
Tumor Segmentation ◽  
Brain Tumor Segmentation ◽  
The Status

Glioma is the main type of malignant brain tumor in adults, and the status of isocitrate dehydrogenase (IDH) mutation highly affects the diagnosis, treatment, and prognosis of gliomas. Radiographic medical imaging provides a noninvasive platform for sampling both inter and intralesion heterogeneity of gliomas, and previous research has shown that the IDH genotype can be predicted from the fusion of multimodality radiology images. The features of medical images and IDH genotype are vital for medical treatment; however, it still lacks a multitask framework for the segmentation of the lesion areas of gliomas and the prediction of IDH genotype. In this paper, we propose a novel three-dimensional (3D) multitask deep learning model for segmentation and genotype prediction (SGPNet). The residual units are also introduced into the SGPNet that allows the output blocks to extract hierarchical features for different tasks and facilitate the information propagation. Our model reduces 26.6% classification error rates comparing with previous models on the datasets of Multimodal Brain Tumor Segmentation Challenge (BRATS) 2020 and The Cancer Genome Atlas (TCGA) gliomas’ databases. Furthermore, we first practically investigate the influence of lesion areas on the performance of IDH genotype prediction by setting different groups of learning targets. The experimental results indicate that the information of lesion areas is more important for the IDH genotype prediction. Our framework is effective and generalizable, which can serve as a highly automated tool to be applied in clinical decision making.

The impact of genotype on outcome in oligodendroglioma: validation of the loss of chromosome arm 1p as an important factor in clinical decision making

2006 ◽  
Vol 104 (4) ◽  
pp. 542-550 ◽  
Author(s):  
Andrew A. Kanner ◽  
Susan M. Staugaitis ◽  
Elias A. Castilla ◽  
Olga Chernova ◽  
Richard A. Prayson ◽  
...  
Keyword(s):  
Decision Making ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
World Health ◽  
Allelic Loss ◽  
Who Grade ◽  
Term Survival ◽  
True Value ◽  
Grade Ii ◽  
The Impact

Object Oligodendrogliomas are rare primary brain tumors. They comprise approximately 5 to 33% of all glial tumors but differ from astrocytomas by being associated with a more favorable prognosis, making their correct identification important. Allelic loss of chromosome arms 1p and 19q is found in a substantial subpopulation of tumors with an oligodendroglioma phenotype. Anaplastic oligodendrogliomas with allelic loss of 1p have been associated with chemosensitivity and a longer patient survival period. Methods Oligodendroglial neoplasms were studied using fluorescence in situ hybridization of formalin-fixed, paraffin-embedded tissue specimens; reference and target probe sets were used to map the telomeric regions of 1p and 19q. The results were correlated with the clinical characteristics of patients treated at our institution between 1993 and 2003. Data obtained in 96 patients were analyzed. This included 63 patients (65.6%) with World Health Organization (WHO) Grade II oligodendroglioma, 22 (23%) with Grade III oligodendroglioma, and 11 (11.4%) with mixed oligoastrocytoma. Analysis of 1p in patients with pure oligodendroglioma revealed a loss of 1p in 42 patients (49.4%). In 46 of these patients 19q was lost and in 70 (82.3%) there was concordance for combined loss or retention of both 1p and 19q (p < 0.0001). Patients with oligodendroglioma in whom a loss of 1p was present fared significantly better, and this outcome was unrelated to the treatment modality or WHO grade, compared with patients in whom 1p was intact (p < 0.05). Conclusions To the authors’ knowledge, this study includes the largest published series of WHO Grade II oligodendroglioma and 1p analysis. The results suggest that the association between long-term survival and 1p loss in oligodendroglioma is unrelated to treatment. The authors of further prospective studies may better determine the true value of the allelic loss of 1p and its implication for clinical decision making.

scholarly journals NIMG-09. PREDICTING OVERALL SURVIVAL OF GLIOBLASTOMA PATIENTS ON MULTI-INSTITUTIONAL HISTOPATHOLOGY STAINED SLIDES USING DEEP LEARNING AND POPULATION-BASED NORMALIZATION

2020 ◽  
Vol 22 (Supplement_2) ◽  
pp. ii148-ii148
Author(s):  
Jie Hao ◽  
Jose Agraz ◽  
Caleb Grenko ◽  
Ji Won Park ◽  
Angela Viaene ◽  
...  
Keyword(s):  
Overall Survival ◽  
Deep Learning ◽  
Clinical Data ◽  
Clinical Decision Making ◽  
Population Based ◽  
Adult Brain ◽  
The Cancer Genome Atlas ◽  
World Health ◽  
Feature Maps ◽  
Testing Data

Abstract MOTIVATION Glioblastoma is the most common and aggressive adult brain tumor. Clinical histopathologic evaluation is essential for tumor classification, which according to the World Health Organization is associated with prognostic information. Accurate prediction of patient overall survival (OS) from clinical routine baseline histopathology whole slide images (WSI) using advanced computational methods, while considering variations in the staining process, could contribute to clinical decision-making and patient management optimization. METHODS We utilize The Cancer Genome Atlas glioblastoma (TCGA-GBM) collection, comprising multi-institutional hematoxylin and eosin (H&E) stained frozen top-section WSI, genomic, and clinical data from 121 subjects. Data are randomly split into training (80%), validation (10%), and testing (10%) sets, while proportionally keeping the ratio of censored patients. We propose a novel deep learning algorithm to identify survival-discriminative histopathological patterns in a WSI, through feature maps, and quantitatively integrate them with gene expression and clinical data to predict patient OS. The concordance index (C-index) is used to quantify the predictive OS performance. Variations in slide staining are assessed through a novel population-based stain normalization approach, informed of glioblastoma distinct histologic sub-regions and their appearance from 509 H&E stained slides with corresponding anatomical annotations from the Ivy Glioblastoma Atlas Project (IvyGAP). RESULTS C-index was equal to 0.797, 0.713, and 0.703 for the training, validation, and testing data, respectively, prior to stain normalization. Following normalization, staining variations in H&E and ‘E’ gained significant improvements in IvyGAP (pWilcoxon&lt; 0.01) and TCGA-GBM (pWilcoxon&lt; 0.0001) data, respectively. These improvements contributed to further optimizing the C-index to 0.871, 0.777, and 0.780 for the training, validation, and testing data, respectively. CONCLUSIONS Appropriate normalization and integrative deep learning yield accurate OS prediction of glioblastoma patients through H&E slides, generalizable in multi-institutional data, potentially contributing to patient stratification in clinical trials. Our computationally-identified survival-discriminative histopathological patterns can contribute in further understanding glioblastoma.

Automated Detection and Classification of Pavement Distresses using 3D Pavement Surface Images and Deep Learning

2021 ◽  
pp. 036119812110074
Author(s):  
Rohit Ghosh ◽  
Omar Smadi
Keyword(s):  
Deep Learning ◽  
Data Augmentation ◽  
Three Dimensional ◽  
Area Under The Curve ◽  
Roc Curves ◽  
Validation Dataset ◽  
Accurate Identification ◽  
Pavement Distresses ◽  
High Level

Pavement distresses lead to pavement deterioration and failure. Accurate identification and classification of distresses helps agencies evaluate the condition of their pavement infrastructure and assists in decision-making processes on pavement maintenance and rehabilitation. The state of the art is automated pavement distress detection using vision-based methods. This study implements two deep learning techniques, Faster Region-based Convolutional Neural Networks (R-CNN) and You Only Look Once (YOLO) v3, for automated distress detection and classification of high resolution (1,800 × 1,200) three-dimensional (3D) asphalt and concrete pavement images. The training and validation dataset contained 625 images that included distresses manually annotated with bounding boxes representing the location and types of distresses and 798 no-distress images. Data augmentation was performed to enable more balanced representation of class labels and prevent overfitting. YOLO and Faster R-CNN achieved 89.8% and 89.6% accuracy respectively. Precision-recall curves were used to determine the average precision (AP), which is the area under the precision-recall curve. The AP values for YOLO and Faster R-CNN were 90.2% and 89.2% respectively, indicating strong performance for both models. Receiver operating characteristic (ROC) curves were also developed to determine the area under the curve, and the resulting area under the curve values of 0.96 for YOLO and 0.95 for Faster R-CNN also indicate robust performance. Finally, the models were evaluated by developing confusion matrices comparing our proposed model with manual quality assurance and quality control (QA/QC) results performed on automated pavement data. A very high level of match to manual QA/QC, namely 97.6% for YOLO and 96.9% for Faster R-CNN, suggest the proposed methodology has potential as a replacement for manual QA/QC.

scholarly journals Explaining deep neural networks for knowledge discovery in electrocardiogram analysis

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Steven A. Hicks ◽  
Jonas L. Isaksen ◽  
Vajira Thambawita ◽  
Jonas Ghouse ◽  
Gustav Ahlberg ◽  
...  
Keyword(s):  
Decision Making ◽  
Deep Learning ◽  
Deep Neural Networks ◽  
Clinical Decision Making ◽  
Medical Knowledge ◽  
Clinical Decision ◽  
Medical Data ◽  
Medical Doctors ◽  
Medical Tests ◽  
Deep Learning Model

AbstractDeep learning-based tools may annotate and interpret medical data more quickly, consistently, and accurately than medical doctors. However, as medical doctors are ultimately responsible for clinical decision-making, any deep learning-based prediction should be accompanied by an explanation that a human can understand. We present an approach called electrocardiogram gradient class activation map (ECGradCAM), which is used to generate attention maps and explain the reasoning behind deep learning-based decision-making in ECG analysis. Attention maps may be used in the clinic to aid diagnosis, discover new medical knowledge, and identify novel features and characteristics of medical tests. In this paper, we showcase how ECGradCAM attention maps can unmask how a novel deep learning model measures both amplitudes and intervals in 12-lead electrocardiograms, and we show an example of how attention maps may be used to develop novel ECG features.

scholarly journals Ethical use of off-label disease-modifying therapies for multiple sclerosis

2021 ◽  
pp. 135245852110302
Author(s):  
Joanna Laurson-Doube ◽  
Nick Rijke ◽  
Anne Helme ◽  
Peer Baneke ◽  
Brenda Banwell ◽  
...  
Keyword(s):  
Multiple Sclerosis ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
World Health ◽  
World Federation ◽  
Off Label ◽  
Disease Modifying Therapies ◽  
The World ◽  
Disease Modifying ◽  
Health Organization

Background: Off-label disease-modifying therapies (DMTs) for multiple sclerosis (MS) are used in at least 89 countries. There is a need for structured and transparent evidence-based guidelines to support clinical decision-making, pharmaceutical policies and reimbursement decisions for off-label DMTs. Objectives/Results: The authors put forward general principles for the ethical use of off-label DMTs for treating MS and a process to assess existing evidence and develop recommendations for their use. Conclusion: The principles and process are endorsed by the World Federation of Neurology (WFN), American Academy of Neurology (AAN), European Academy of Neurology (EAN), Americas Committee for Treatment and Research in Multiple Sclerosis (ACTRIMS), European Committee for Treatment and Research in Multiple Sclerosis (ECTRIMS), Middle-East North Africa Committee for Treatment and Research in Multiple Sclerosis (MENACTRIMS) and Pan-Asian Committee for Treatment and Research in Multiple Sclerosis (PACTRIMS), and we have regularly consulted with the Brain Health Unit, Mental Health and Substance Use Department at the World Health Organization (WHO).

scholarly journals Impact of Grade Groups on Prostate Cancer-Specific and Other-Cause Mortality: Competing Risk Analysis from a Large Single Institution Series

Cancers ◽  
2021 ◽  
Vol 13 (8) ◽  
pp. 1963
Author(s):  
Daimantas Milonas ◽  
Tomas Ruzgas ◽  
Zilvinas Venclovas ◽  
Mindaugas Jievaltas ◽  
Steven Joniau
Keyword(s):  
Clinical Decision Making ◽  
Area Under The Curve ◽  
Clinical Decision ◽  
Competing Risk ◽  
Grade Group ◽  
Tertiary Center ◽  
Stable Performance ◽  
Cancer Specific Mortality ◽  
Risk Of Cancer

Objective: To assess the risk of cancer-specific mortality (CSM) and other-cause mortality (OCM) using post-operative International Society of Urological Pathology Grade Group (GG) model in patients after radical prostatectomy (RP). Patients and Methods: Overall 1921 consecutive men who underwent RP during 2001 to 2017 in a single tertiary center were included in the study. Multivariate competing risk regression analysis was used to identify significant predictors and quantify cumulative incidence of CSM and OCM. Time-depending area under the curve (AUC) depicted the performance of GG model on prediction of CSM. Results: Over a median follow-up of 7.9-year (IQR 4.4-11.7) after RP, 235 (12.2%) deaths were registered, and 52 (2.7%) of them were related to PCa. GG model showed high and stable performance (time-dependent AUC 0.88) on prediction of CSM. Cumulative 10-year CSM in GGs 1 to 5 was 0.9%, 2.3%, 7.6%, 14.7%, and 48.6%, respectively; 10-year OCM in GGs was 15.5%, 16.1%, 12.6%, 17.7% and 6.5%, respectively. The ratio between 10-year CSM/OCM in GGs 1 to 5 was 1:17, 1:7, 1:2, 1:1, and 7:1, respectively. Conclusions: Cancer-specific and other-cause mortality differed widely between GGs. Presented findings could aid in personalized clinical decision making for active treatment.

scholarly journals Telomere Architecture Correlates with Aggressiveness in Multiple Myeloma

Cancers ◽  
2021 ◽  
Vol 13 (8) ◽  
pp. 1969
Author(s):  
Aline Rangel-Pozzo ◽  
Pak Yu ◽  
Sadhana LaL ◽  
Yasmin Asbaghi ◽  
Luiza Sisdelli ◽  
...  
Keyword(s):  
Multiple Myeloma ◽  
Disease Progression ◽  
Genomic Instability ◽  
Clinical Decision Making ◽  
Three Dimensional ◽  
Clinical Decision ◽  
Response To Treatment ◽  
Average Intensity ◽  
Patient Response ◽  
Smoldering Multiple Myeloma

The prognosis of multiple myeloma (MM), an incurable B-cell malignancy, has significantly improved through the introduction of novel therapeutic modalities. Myeloma prognosis is essentially determined by cytogenetics, both at diagnosis and at disease progression. However, for a large cohort of patients, cytogenetic analysis is not always available. In addition, myeloma patients with favorable cytogenetics can display an aggressive clinical course. Therefore, it is necessary to develop additional prognostic and predictive markers for this disease to allow for patient risk stratification and personalized clinical decision-making. Genomic instability is a prominent characteristic in MM, and we have previously shown that the three-dimensional (3D) nuclear organization of telomeres is a marker of both genomic instability and genetic heterogeneity in myeloma. In this study, we compared in a longitudinal prospective study blindly the 3D telomeric profiles from bone marrow samples of 214 initially treatment-naïve patients with either monoclonal gammopathy of undetermined significance (MGUS), smoldering multiple myeloma (SMM), or MM, with a minimum follow-up of 5 years. Here, we report distinctive 3D telomeric profiles correlating with disease aggressiveness and patient response to treatment in MM patients, and also distinctive 3D telomeric profiles for disease progression in smoldering multiple myeloma patients. In particular, lower average intensity (telomere length, below 13,500 arbitrary units) and increased number of telomere aggregates are associated with shorter survival and could be used as a prognostic factor to identify high-risk SMM and MM patients.

scholarly journals Diagnostic accuracy of symptoms as a diagnostic tool for SARS-CoV 2 infection: a cross-sectional study in a cohort of 2,173 patients

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Carlos Alfonso Romero-Gameros ◽  
Tania Colin-Martínez ◽  
Salomón Waizel-Haiat ◽  
Guadalupe Vargas-Ortega ◽  
Eduardo Ferat-Osorio ◽  
...  
Keyword(s):  
Diagnostic Accuracy ◽  
Diagnostic Tool ◽  
Clinical Decision Making ◽  
Multivariate Logistic Regression Analysis ◽  
Clinical Decision ◽  
Cross Sectional Study ◽  
World Health ◽  
Likelihood Ratios ◽  
Cross Sectional ◽  
Health Organization

Abstract Background The SARS-CoV-2 pandemic continues to be a priority health problem; According to the World Health Organization data from October 13, 2020, 37,704,153 confirmed COVID-19 cases have been reported, including 1,079,029 deaths, since the outbreak. The identification of potential symptoms has been reported to be a useful tool for clinical decision-making in emergency departments to avoid overload and improve the quality of care. The aim of this study was to evaluate the performances of symptoms as a diagnostic tool for SARS -CoV-2 infection. Methods An observational, cross-sectional, prospective and analytical study was carried out, during the period of time from April 14 to July 21, 2020. Data (demographic variables, medical history, respiratory and non-respiratory symptoms) were collected by emergency physicians. The diagnosis of COVID-19 was made using SARS-CoV-2 RT-PCR. The diagnostic accuracy of these characteristics for COVID-19 was evaluated by calculating the positive and negative likelihood ratios. A Mantel-Haenszel and multivariate logistic regression analysis was performed to assess the association of symptoms with COVID-19. Results A prevalence of 53.72% of SARS-CoV-2 infection was observed. The symptom with the highest sensitivity was cough 71%, and a specificity of 52.68%. The symptomatological scale, constructed from 6 symptoms, obtained a sensitivity of 83.45% and a specificity of 32.86%, taking ≥2 symptoms as a cut-off point. The symptoms with the greatest association with SARS-CoV-2 were: anosmia odds ratio (OR) 3.2 (95% CI; 2.52–4.17), fever OR 2.98 (95% CI; 2.47–3.58), dyspnea OR 2.9 (95% CI; 2.39–3.51]) and cough OR 2.73 (95% CI: 2.27–3.28). Conclusion The combination of ≥2 symptoms / signs (fever, cough, anosmia, dyspnea and oxygen saturation < 93%, and headache) results in a highly sensitivity model for a quick and accurate diagnosis of COVID-19, and should be used in the absence of ancillary diagnostic studies. Symptomatology, alone and in combination, may be an appropriate strategy to use in the emergency department to guide the behaviors to respond to the disease. Trial registration Institutional registration R-2020-3601-145, Federal Commission for the Protection against Sanitary Risks 17 CI-09-015-034, National Bioethics Commission: 09 CEI-023-2017082.

scholarly journals Deep learning-based estimation of Flory–Huggins parameter of A–B block copolymers from cross-sectional images of phase-separated structures

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Katsumi Hagita ◽  
Takeshi Aoyagi ◽  
Yuto Abe ◽  
Shinya Genda ◽  
Takashi Honda
Keyword(s):  
Deep Learning ◽  
Diblock Copolymers ◽  
Three Dimensional ◽  
Image Data ◽  
Real Space ◽  
Cross Sectional ◽  
Regression Problem ◽  
Self Consistent Field ◽  
Experimental Image ◽  
Chain Lengths

AbstractIn this study, deep learning (DL)-based estimation of the Flory–Huggins χ parameter of A-B diblock copolymers from two-dimensional cross-sectional images of three-dimensional (3D) phase-separated structures were investigated. 3D structures with random networks of phase-separated domains were generated from real-space self-consistent field simulations in the 25–40 χN range for chain lengths (N) of 20 and 40. To confirm that the prepared data can be discriminated using DL, image classification was performed using the VGG-16 network. We comprehensively investigated the performances of the learned networks in the regression problem. The generalization ability was evaluated from independent images with the unlearned χN. We found that, except for large χN values, the standard deviation values were approximately 0.1 and 0.5 for A-component fractions of 0.2 and 0.35, respectively. The images for larger χN values were more difficult to distinguish. In addition, the learning performances for the 4-class problem were comparable to those for the 8-class problem, except when the χN values were large. This information is useful for the analysis of real experimental image data, where the variation of samples is limited.

Sign in / Sign up

Export Citation Format

Share Document