| 1. |
- Ali, Muhaddisa Barat, 1986, et al.
(författare)
-
A novel federated deep learning scheme for glioma and its subtype classification
- 2023
-
Ingår i: Frontiers in Neuroscience. - 1662-4548 .- 1662-453X. ; 17
-
Tidskriftsartikel (refereegranskat)abstract
- Background: Deep learning (DL) has shown promising results in molecular-based classification of glioma subtypes from MR images. DL requires a large number of training data for achieving good generalization performance. Since brain tumor datasets are usually small in size, combination of such datasets from different hospitals are needed. Data privacy issue from hospitals often poses a constraint on such a practice. Federated learning (FL) has gained much attention lately as it trains a central DL model without requiring data sharing from different hospitals. Method: We propose a novel 3D FL scheme for glioma and its molecular subtype classification. In the scheme, a slice-based DL classifier, EtFedDyn, is exploited which is an extension of FedDyn, with the key differences on using focal loss cost function to tackle severe class imbalances in the datasets, and on multi-stream network to exploit MRIs in different modalities. By combining EtFedDyn with domain mapping as the pre-processing and 3D scan-based post-processing, the proposed scheme makes 3D brain scan-based classification on datasets from different dataset owners. To examine whether the FL scheme could replace the central learning (CL) one, we then compare the classification performance between the proposed FL and the corresponding CL schemes. Furthermore, detailed empirical-based analysis were also conducted to exam the effect of using domain mapping, 3D scan-based post-processing, different cost functions and different FL schemes. Results: Experiments were done on two case studies: classification of glioma subtypes (IDH mutation and wild-type on TCGA and US datasets in case A) and glioma grades (high/low grade glioma HGG and LGG on MICCAI dataset in case B). The proposed FL scheme has obtained good performance on the test sets (85.46%, 75.56%) for IDH subtypes and (89.28%, 90.72%) for glioma LGG/HGG all averaged on five runs. Comparing with the corresponding CL scheme, the drop in test accuracy from the proposed FL scheme is small (−1.17%, −0.83%), indicating its good potential to replace the CL scheme. Furthermore, the empirically tests have shown that an increased classification test accuracy by applying: domain mapping (0.4%, 1.85%) in case A; focal loss function (1.66%, 3.25%) in case A and (1.19%, 1.85%) in case B; 3D post-processing (2.11%, 2.23%) in case A and (1.81%, 2.39%) in case B and EtFedDyn over FedAvg classifier (1.05%, 1.55%) in case A and (1.23%, 1.81%) in case B with fast convergence, which all contributed to the improvement of overall performance in the proposed FL scheme. Conclusion: The proposed FL scheme is shown to be effective in predicting glioma and its subtypes by using MR images from test sets, with great potential of replacing the conventional CL approaches for training deep networks. This could help hospitals to maintain their data privacy, while using a federated trained classifier with nearly similar performance as that from a centrally trained one. Further detailed experiments have shown that different parts in the proposed 3D FL scheme, such as domain mapping (make datasets more uniform) and post-processing (scan-based classification), are essential.
|
|
| 2. |
- Henje, Catharina, 1960-, et al.
(författare)
-
Obstacles and risks in the traffic environment for users of powered wheelchairs in Sweden
- 2021
-
Ingår i: Accident Analysis and Prevention. - : Elsevier. - 0001-4575 .- 1879-2057. ; 159
-
Tidskriftsartikel (refereegranskat)abstract
- Objective: According to the European Union, fatal road accidents involving Vulnerable Road Users (VRUs) are equal in proportion to fatal car road accidents (46%). VRUs include individuals with mobility challenges such as the elderly and Powered Wheelchair (PWC) users. The aim of this interdisciplinary qualitative study was to identify obstacles and risks for PWC users by exploring their behaviour and experiences in traffic environments.Methods: Videos and in-depth interviews with 13 PWC users aged 20–66 were analysed for this study. The interviews and videos, which include real-life outdoor observations, originate from a qualitative study exploring experiences of PWC use on a daily basis in Sweden. Underlying causal factors to identified risks and obstacles were identified, based on human, vehicle (PWC) and environmental factors in accordance with the Haddon Matrix.Results: The results show significant potential for improvement within all three perspectives of the Haddon Matrix used in the analysis. Participants faced and dealt with various obstacles and risks in order to reach their destination. For example, this includes uneven surfaces, differences in ground levels, steep slopes, as well as interactions with other road users and the influence of weather conditions, resulting in PWC users constantly accommodating and coping with the shortcomings of the vehicle and the environment.Conclusions: There are still major challenges with regard to preventing obstacles and risks in the traffic environment for PWC users. To discern PWC users in traffic accident and injury data bases, a start would be to register type of aid used for persons involved in an accident. Furthermore, to emphasise PWC users’ role as VRUs, it may also be advantageous to describe them as drivers rather than users when navigating the traffic environment. Given the limited sample, further research covering more data from a broader perspective would be beneficial. By incorporating emerging knowledge of PWC users’ prerequisites and needs, and including them in research and traffic planning, the society will grow safer and more inclusive, and become better prepared for meeting future demands on accessibility from an aging population.
|
|
| 3. |
- Ge, Chenjie, 1991, et al.
(författare)
-
Enlarged Training Dataset by Pairwise GANs for Molecular-Based Brain Tumor Classification
- 2020
-
Ingår i: IEEE Access. - 2169-3536 .- 2169-3536. ; 8:1, s. 22560-22570
-
Tidskriftsartikel (refereegranskat)abstract
- This paper addresses issues of brain tumor subtype classification using Magnetic Resonance Images (MRIs) from different scanner modalities like T1 weighted, T1 weighted with contrast-enhanced, T2 weighted and FLAIR images. Currently most available glioma datasets are relatively moderate in size, and often accompanied with incomplete MRIs in different modalities. To tackle the commonly encountered problems of insufficiently large brain tumor datasets and incomplete modality of image for deep learning, we propose to add augmented brain MR images to enlarge the training dataset by employing a pairwise Generative Adversarial Network (GAN) model. The pairwise GAN is able to generate synthetic MRIs across different modalities. To achieve the patient-level diagnostic result, we propose a post-processing strategy to combine the slice-level glioma subtype classification results by majority voting. A two-stage course-to-fine training strategy is proposed to learn the glioma feature using GAN-augmented MRIs followed by real MRIs. To evaluate the effectiveness of the proposed scheme, experiments have been conducted on a brain tumor dataset for classifying glioma molecular subtypes: isocitrate dehydrogenase 1 (IDH1) mutation and IDH1 wild-type. Our results on the dataset have shown good performance (with test accuracy 88.82%). Comparisons with several state-of-the-art methods are also included.
|
|
| 4. |
- Bergh, Ingrid, 1956, et al.
(författare)
-
Descriptions of pain in elderly patients following orthopaedic surgery.
- 2005
-
Ingår i: Scandinavian journal of caring sciences. - : Wiley. - 0283-9318 .- 1471-6712. ; 19:2, s. 110-8
-
Tidskriftsartikel (refereegranskat)abstract
- The aims of this study were to investigate what words elderly patients, who had undergone hip surgery, used to describe their experience of pain in spoken language and to compare these words with those used in the Short-Form McGill Pain Questionnaire (SF-MPQ) and Pain-O-Meter (POM). The study was carried out at two orthopaedic and two geriatric clinical departments at a large university hospital in Sweden. Altogether, 60 patients (mean age =77) who had undergone orthopaedic surgery took part in the study. A face-to-face interview was conducted with each patient on the second day after the operation. This was divided into two parts, one tape-recorded and semi-structured in character and one structured interview. The results show that a majority of the elderly patients who participated in this study verbally stated pain and spontaneously used a majority of the words found in the SF-MPQ and in the POM. The patients also used a number of additional words not found in the SF-MPQ or the POM. Among those patients who did not use any of the words in the SF-MPQ and the POM, the use of the three additional words 'stel' (stiff), 'hemsk' (awful) and 'rad(d)(sla)' (afraid/fear) were especially marked. The patients also combined the words with a negation to describe what pain was not. To achieve a more balanced and nuanced description of the patient's pain and to make it easier for the patients to talk about their pain, there is a need for access to a set of predefined words that describe pain from a more multidimensional perspective than just intensity. If the elderly patient is allowed, and finds it necessary, to use his/her own words to describe what pain is but also to describe what pain is not, by combining the words with a negation, then the risk of the patient being forced to choose words that do not fully correspond to their pain can be reduced. If so, pain scales such as the SF-MPQ and the POM can create a communicative bridge between the elderly patient and health care professionals in the pain evaluation process.
|
|
| 5. |
- Altay, Özlem, et al.
(författare)
-
Combined Metabolic Activators Accelerates Recovery in Mild-to-Moderate COVID-19
- 2021
-
Ingår i: Advanced Science. - : Wiley. - 2198-3844. ; 8:17
-
Tidskriftsartikel (refereegranskat)abstract
- COVID-19 is associated with mitochondrial dysfunction and metabolic abnormalities, including the deficiencies in nicotinamide adenine dinucleotide (NAD+) and glutathione metabolism. Here it is investigated if administration of a mixture of combined metabolic activators (CMAs) consisting of glutathione and NAD+ precursors can restore metabolic function and thus aid the recovery of COVID-19 patients. CMAs include l-serine, N-acetyl-l-cysteine, nicotinamide riboside, and l-carnitine tartrate, salt form of l-carnitine. Placebo-controlled, open-label phase 2 study and double-blinded phase 3 clinical trials are conducted to investigate the time of symptom-free recovery on ambulatory patients using CMAs. The results of both studies show that the time to complete recovery is significantly shorter in the CMA group (6.6 vs 9.3 d) in phase 2 and (5.7 vs 9.2 d) in phase 3 trials compared to placebo group. A comprehensive analysis of the plasma metabolome and proteome reveals major metabolic changes. Plasma levels of proteins and metabolites associated with inflammation and antioxidant metabolism are significantly improved in patients treated with CMAs as compared to placebo. The results show that treating patients infected with COVID-19 with CMAs lead to a more rapid symptom-free recovery, suggesting a role for such a therapeutic regime in the treatment of infections leading to respiratory problems.
|
|
| 6. |
- Ali, Muhaddisa Barat, 1986, et al.
(författare)
-
Multi-stream Convolutional Autoencoder and 2D Generative Adversarial Network for Glioma Classification
- 2019
-
Ingår i: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). - Cham : Springer International Publishing. - 1611-3349 .- 0302-9743. ; 11678 LNCS, s. 234-245
-
Konferensbidrag (refereegranskat)abstract
- Diagnosis and timely treatment play an important role in preventing brain tumor growth. Deep learning methods have gained much attention lately. Obtaining a large amount of annotated medical data remains a challenging issue. Furthermore, high dimensional features of brain images could lead to over-fitting. In this paper, we address the above issues. Firstly, we propose an architecture for Generative Adversarial Networks to generate good quality synthetic 2D MRIs from multi-modality MRIs (T1 contrast-enhanced, T2, FLAIR). Secondly, we propose a deep learning scheme based on 3-streams of Convolutional Autoencoders (CAEs) followed by sensor information fusion. The rational behind using CAEs is that it may improve glioma classification performance (as comparing with conventional CNNs), since CAEs offer noise robustness and also efficient feature reduction hence possibly reduce the over-fitting. A two-round training strategy is also applied by pre-training on GAN augmented synthetic MRIs followed by refined-training on original MRIs. Experiments on BraTS 2017 dataset have demonstrated the effectiveness of the proposed scheme (test accuracy 92.04%). Comparison with several exiting schemes has provided further support to the proposed scheme.
|
|
| 7. |
- Van Olden, C. C., et al.
(författare)
-
A systems biology approach to understand gut microbiota and host metabolism in morbid obesity: design of the BARIA Longitudinal Cohort Study
- 2021
-
Ingår i: Journal of Internal Medicine. - : Wiley. - 0954-6820 .- 1365-2796. ; 289:3, s. 340-354
-
Tidskriftsartikel (refereegranskat)abstract
- Introduction Prevalence of obesity and associated diseases, including type 2 diabetes mellitus, dyslipidaemia and non-alcoholic fatty liver disease (NAFLD), are increasing. Underlying mechanisms, especially in humans, are unclear. Bariatric surgery provides the unique opportunity to obtain biopsies and portal vein blood-samples. Methods The BARIA Study aims to assess how microbiota and their metabolites affect transcription in key tissues and clinical outcome in obese subjects and how baseline anthropometric and metabolic characteristics determine weight loss and glucose homeostasis after bariatric surgery. We phenotype patients undergoing bariatric surgery (predominantly laparoscopic Roux-en-Y gastric bypass), before weight loss, with biometrics, dietary and psychological questionnaires, mixed meal test (MMT) and collect fecal-samples and intra-operative biopsies from liver, adipose tissues and jejunum. We aim to include 1500 patients. A subset (approximately 25%) will undergo intra-operative portal vein blood-sampling. Fecal-samples are analyzed with shotgun metagenomics and targeted metabolomics, fasted and postprandial plasma-samples are subjected to metabolomics, and RNA is extracted from the tissues for RNAseq-analyses. Data will be integrated using state-of-the-art neuronal networks and metabolic modeling. Patient follow-up will be ten years. Results Preoperative MMT of 170 patients were analysed and clear differences were observed in glucose homeostasis between individuals. Repeated MMT in 10 patients showed satisfactory intra-individual reproducibility, with differences in plasma glucose, insulin and triglycerides within 20% of the mean difference. Conclusion The BARIA study can add more understanding in how gut-microbiota affect metabolism, especially with regard to obesity, glucose metabolism and NAFLD. Identification of key factors may provide diagnostic and therapeutic leads to control the obesity-associated disease epidemic.
|
|
| 8. |
- Huvila, J., et al.
(författare)
-
Combined ASRGL1 and p53 immunohistochemistry as an independent predictor of survival in endometrioid endometrial carcinoma
- 2018
-
Ingår i: Gynecologic Oncology. - : Academic Press Inc.. - 0090-8258 .- 1095-6859. ; 149:1, s. 173-180
-
Tidskriftsartikel (refereegranskat)abstract
- Objective: In clinical practise, prognostication of endometrial cancer is based on clinicopathological risk factors. The use of immunohistochemistry-based markers as prognostic tools is generally not recommended and a systematic analysis of their utility as a panel is lacking. We evaluated whether an immunohistochemical marker panel could reliably assess endometrioid endometrial cancer (EEC) outcome independent of clinicopathological information. Methods: A cohort of 306 EEC specimens was profiled using tissue microarray (TMA). Cost- and time-efficient immunohistochemical analysis of well-established tissue biomarkers (ER, PR, HER2, Ki-67, MLH1 and p53) and two new biomarkers (L1CAM and ASRGL1) was carried out. Statistical modelling with embedded variable selection was applied on the staining results to identify minimal prognostic panels with maximal prognostic accuracy without compromising generalizability. Results: A panel including p53 and ASRGL1 immunohistochemistry was identified as the most accurate predictor of relapse-free and disease-specific survival. Within this panel, patients were allocated into high- (5.9%), intermediate- (29.5%) and low- (64.6%) risk groups where high-risk patients had a 30-fold risk (P < 0.001) of dying of EEC compared to the low-risk group. Conclusions: P53 and ASRGL1 immunoprofiling stratifies EEC patients into three risk groups with significantly different outcomes. This simple and easily applicable panel could provide a useful tool in EEC risk stratification and guiding the allocation of treatment modalities.
|
|
| 9. |
- Ge, Chenjie, 1991, et al.
(författare)
-
Cross-Modality Augmentation of Brain Mr Images Using a Novel Pairwise Generative Adversarial Network for Enhanced Glioma Classification
- 2019
-
Ingår i: Proceedings - International Conference on Image Processing, ICIP. - 1522-4880.
-
Konferensbidrag (refereegranskat)abstract
- © 2019 IEEE. Brain Magnetic Resonance Images (MRIs) are commonly used for tumor diagnosis. Machine learning for brain tumor characterization often uses MRIs from many modalities (e.g., T1-MRI, Enhanced-T1-MRI, T2-MRI and FLAIR). This paper tackles two issues that may impact brain tumor characterization performance from deep learning: insufficiently large training dataset, and incomplete collection of MRIs from different modalities. We propose a novel pairwise generative adversarial network (GAN) architecture for generating synthetic brain MRIs in missing modalities by using existing MRIs in other modalities. By improving the training dataset, we aim to mitigate the overfitting and improve the deep learning performance. Main contributions of the paper include: (a) propose a pairwise generative adversarial network (GAN) for brain image augmentation via cross-modality image generation; (b) propose a training strategy to enhance the glioma classification performance, where GAN-augmented images are used for pre-training, followed by refined-training using real brain MRIs; (c) demonstrate the proposed method through tests and comparisons of glioma classifiers that are trained from mixing real and GAN synthetic data, as well as from real data only. Experiments were conducted on an open TCGA dataset, containing 167 subjects for classifying IDH genotypes (mutation or wild-type). Test results from two experimental settings have both provided supports to the proposed method, where glioma classification performance has consistently improved by using mixed real and augmented data (test accuracy 81.03%, with 2.57% improvement).
|
|
| 10. |
- Ge, Chenjie, 1991, et al.
(författare)
-
Deep semi-supervised learning for brain tumor classification
- 2020
-
Ingår i: BMC Medical Imaging. - : Springer Science and Business Media LLC. - 1471-2342. ; 20:1
-
Tidskriftsartikel (refereegranskat)abstract
- Background: This paper addresses issues of brain tumor, glioma, classification from four modalities of Magnetic Resonance Image (MRI) scans (i.e., T1 weighted MRI, T1 weighted MRI with contrast-enhanced, T2 weighted MRI and FLAIR). Currently, many available glioma datasets often contain some unlabeled brain scans, and many datasets are moderate in size. Methods: We propose to exploit deep semi-supervised learning to make full use of the unlabeled data. Deep CNN features were incorporated into a new graph-based semi-supervised learning framework for learning the labels of the unlabeled data, where a new 3D-2D consistent constraint is added to make consistent classifications for the 2D slices from the same 3D brain scan. A deep-learning classifier is then trained to classify different glioma types using both labeled and unlabeled data with estimated labels. To alleviate the overfitting caused by moderate-size datasets, synthetic MRIs generated by Generative Adversarial Networks (GANs) are added in the training of CNNs. Results: The proposed scheme has been tested on two glioma datasets, TCGA dataset for IDH-mutation prediction (molecular-based glioma subtype classification) and MICCAI dataset for glioma grading. Our results have shown good performance (with test accuracies 86.53% on TCGA dataset and 90.70% on MICCAI dataset). Conclusions: The proposed scheme is effective for glioma IDH-mutation prediction and glioma grading, and its performance is comparable to the state-of-the-art.
|
|
