Andrew D. Kane

Bio: Andrew D. Kane is an academic researcher from University of Cambridge. The author has contributed to research in topics: Fetus & Hypoxia (medical). The author has an hindex of 19, co-authored 33 publications receiving 3760 citations. Previous affiliations of Andrew D. Kane include James Cook University Hospital & Royal College of Anaesthetists.

Unsupervised domain adaptation in brain lesion segmentation with adversarial networks

Abstract: Significant advances have been made towards building accurate automatic segmentation systems for a variety of biomedical applications using machine learning. However, the performance of these systems often degrades when they are applied on new data that differ from the training data, for example, due to variations in imaging protocols. Manually annotating new data for each test domain is not a feasible solution. In this work we investigate unsupervised domain adaptation using adversarial neural networks to train a segmentation method which is more robust to differences in the input data, and which does not require any annotations on the test domain. Specifically, we derive domain-invariant features by learning to counter an adversarial network, which attempts to classify the domain of the input data by observing the activations of the segmentation network. Furthermore, we propose a multi-connected domain discriminator for improved adversarial training. Our system is evaluated using two MR databases of subjects with traumatic brain injuries, acquired using different scanners and imaging protocols. Using our unsupervised approach, we obtain segmentation accuracies which are close to the upper bound of supervised domain adaptation.

Outcomes from intensive care in patients with COVID-19: a systematic review and meta-analysis of observational studies.

Abstract: The emergence of coronavirus disease 2019 (COVID-19) has led to high demand for intensive care services worldwide. However, the mortality of patients admitted to the intensive care unit (ICU) with COVID-19 is unclear. Here, we perform a systematic review and meta-analysis, in line with PRISMA guidelines, to assess the reported ICU mortality for patients with confirmed COVID-19. We searched MEDLINE, EMBASE, PubMed and Cochrane databases up to 31 May 2020 for studies reporting ICU mortality for adult patients admitted with COVID-19. The primary outcome measure was death in intensive care as a proportion of completed ICU admissions, either through discharge from the ICU or death. The definition thus did not include patients still alive on ICU. Twenty-four observational studies including 10,150 patients were identified from centres across Asia, Europe and North America. In-ICU mortality in reported studies ranged from 0 to 84.6%. Seven studies reported outcome data for all patients. In the remaining studies, the proportion of patients discharged from ICU at the point of reporting varied from 24.5 to 97.2%. In patients with completed ICU admissions with COVID-19 infection, combined ICU mortality (95%CI) was 41.6% (34.0-49.7%), I2 = 93.2%). Sub-group analysis by continent showed that mortality is broadly consistent across the globe. As the pandemic has progressed, the reported mortality rates have fallen from above 50% to close to 40%. The in-ICU mortality from COVID-19 is higher than usually seen in ICU admissions with other viral pneumonias. Importantly, the mortality from completed episodes of ICU differs considerably from the crude mortality rates in some early reports.

Unsupervised domain adaptation in brain lesion segmentation with adversarial networks

Abstract: Significant advances have been made towards building accurate automatic segmentation systems for a variety of biomedical applications using machine learning. However, the performance of these systems often degrades when they are applied on new data that differ from the training data, for example, due to variations in imaging protocols. Manually annotating new data for each test domain is not a feasible solution. In this work we investigate unsupervised domain adaptation using adversarial neural networks to train a segmentation method which is more invariant to differences in the input data, and which does not require any annotations on the test domain. Specifically, we learn domain-invariant features by learning to counter an adversarial network, which attempts to classify the domain of the input data by observing the activations of the segmentation network. Furthermore, we propose a multi-connected domain discriminator for improved adversarial training. Our system is evaluated using two MR databases of subjects with traumatic brain injuries, acquired using different scanners and imaging protocols. Using our unsupervised approach, we obtain segmentation accuracies which are close to the upper bound of supervised domain adaptation.

Developmental Programming of Cardiovascular Dysfunction by Prenatal Hypoxia and Oxidative Stress

Abstract: Fetal hypoxia is a common complication of pregnancy. It has been shown to programme cardiac and endothelial dysfunction in the offspring in adult life. However, the mechanisms via which this occurs remain elusive, precluding the identification of potential therapy. Using an integrative approach at the isolated organ, cellular and molecular levels, we tested the hypothesis that oxidative stress in the fetal heart and vasculature underlies the molecular basis via which prenatal hypoxia programmes cardiovascular dysfunction in later life. In a longitudinal study, the effects of maternal treatment of hypoxic (13% O(2)) pregnancy with an antioxidant on the cardiovascular system of the offspring at the end of gestation and at adulthood were studied. On day 6 of pregnancy, rats (n = 20 per group) were exposed to normoxia or hypoxia ± vitamin C. At gestational day 20, tissues were collected from 1 male fetus per litter per group (n = 10). The remaining 10 litters per group were allowed to deliver. At 4 months, tissues from 1 male adult offspring per litter per group were either perfusion fixed, frozen, or dissected for isolated organ preparations. In the fetus, hypoxic pregnancy promoted aortic thickening with enhanced nitrotyrosine staining and an increase in cardiac HSP70 expression. By adulthood, offspring of hypoxic pregnancy had markedly impaired NO-dependent relaxation in femoral resistance arteries, and increased myocardial contractility with sympathetic dominance. Maternal vitamin C prevented these effects in fetal and adult offspring of hypoxic pregnancy. The data offer insight to mechanism and thereby possible targets for intervention against developmental origins of cardiac and peripheral vascular dysfunction in offspring of risky pregnancy.

V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation

Abstract: Convolutional Neural Networks (CNNs) have been recently employed to solve problems from both the computer vision and medical image analysis fields. Despite their popularity, most approaches are only able to process 2D images while most medical data used in clinical practice consists of 3D volumes. In this work we propose an approach to 3D image segmentation based on a volumetric, fully convolutional, neural network. Our CNN is trained end-to-end on MRI volumes depicting prostate, and learns to predict segmentation for the whole volume at once. We introduce a novel objective function, that we optimise during training, based on Dice coefficient. In this way we can deal with situations where there is a strong imbalance between the number of foreground and background voxels. To cope with the limited number of annotated volumes available for training, we augment the data applying random non-linear transformations and histogram matching. We show in our experimental evaluation that our approach achieves good performances on challenging test data while requiring only a fraction of the processing time needed by other previous methods.

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Abstract: Summary Background Since December, 2019, Wuhan, China, has experienced an outbreak of coronavirus disease 2019 (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Epidemiological and clinical characteristics of patients with COVID-19 have been reported but risk factors for mortality and a detailed clinical course of illness, including viral shedding, have not been well described. Methods In this retrospective, multicentre cohort study, we included all adult inpatients (≥18 years old) with laboratory-confirmed COVID-19 from Jinyintan Hospital and Wuhan Pulmonary Hospital (Wuhan, China) who had been discharged or had died by Jan 31, 2020. Demographic, clinical, treatment, and laboratory data, including serial samples for viral RNA detection, were extracted from electronic medical records and compared between survivors and non-survivors. We used univariable and multivariable logistic regression methods to explore the risk factors associated with in-hospital death. Findings 191 patients (135 from Jinyintan Hospital and 56 from Wuhan Pulmonary Hospital) were included in this study, of whom 137 were discharged and 54 died in hospital. 91 (48%) patients had a comorbidity, with hypertension being the most common (58 [30%] patients), followed by diabetes (36 [19%] patients) and coronary heart disease (15 [8%] patients). Multivariable regression showed increasing odds of in-hospital death associated with older age (odds ratio 1·10, 95% CI 1·03–1·17, per year increase; p=0·0043), higher Sequential Organ Failure Assessment (SOFA) score (5·65, 2·61–12·23; p Interpretation The potential risk factors of older age, high SOFA score, and d-dimer greater than 1 μg/mL could help clinicians to identify patients with poor prognosis at an early stage. Prolonged viral shedding provides the rationale for a strategy of isolation of infected patients and optimal antiviral interventions in the future. Funding Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences; National Science Grant for Distinguished Young Scholars; National Key Research and Development Program of China; The Beijing Science and Technology Project; and Major Projects of National Science and Technology on New Drug Creation and Development.

Attention U-Net: Learning Where to Look for the Pancreas

Abstract: We propose a novel attention gate (AG) model for medical imaging that automatically learns to focus on target structures of varying shapes and sizes. Models trained with AGs implicitly learn to suppress irrelevant regions in an input image while highlighting salient features useful for a specific task. This enables us to eliminate the necessity of using explicit external tissue/organ localisation modules of cascaded convolutional neural networks (CNNs). AGs can be easily integrated into standard CNN architectures such as the U-Net model with minimal computational overhead while increasing the model sensitivity and prediction accuracy. The proposed Attention U-Net architecture is evaluated on two large CT abdominal datasets for multi-class image segmentation. Experimental results show that AGs consistently improve the prediction performance of U-Net across different datasets and training sizes while preserving computational efficiency. The code for the proposed architecture is publicly available.

V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation

Abstract: Convolutional Neural Networks (CNNs) have been recently employed to solve problems from both the computer vision and medical image analysis fields. Despite their popularity, most approaches are only able to process 2D images while most medical data used in clinical practice consists of 3D volumes. In this work we propose an approach to 3D image segmentation based on a volumetric, fully convolutional, neural network. Our CNN is trained end-to-end on MRI volumes depicting prostate, and learns to predict segmentation for the whole volume at once. We introduce a novel objective function, that we optimise during training, based on Dice coefficient. In this way we can deal with situations where there is a strong imbalance between the number of foreground and background voxels. To cope with the limited number of annotated volumes available for training, we augment the data applying random non-linear transformations and histogram matching. We show in our experimental evaluation that our approach achieves good performances on challenging test data while requiring only a fraction of the processing time needed by other previous methods.