Joshua Bridge

Joshua Bridge

Auto-generated profile, based on publicly available information

Joshua Bridge

Division of Surgical & Interventional Science, University College London, London, UK

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 July 2020 in IEEE Journal of Biomedical and Health Informatics
Verified
Journal Article
Open access
JBJoshua Bridge
Joshua Bridge
YMYanda Meng
Yanda Meng
Yifan ZhaoYifan Zhao
Yifan Zhao
YDYong Du
Yong Du
MZMingfeng Zhao
Mingfeng Zhao
Yalin ZhengYalin Zheng
Yalin Zheng
Fast and accurate diagnosis is essential for the efficient and effective control of the COVID-19 pandemic that is currently disrupting the whole world. Despite the prevalence of the COVID-19 outbreak, relatively few diagnostic images are openly available to develop automatic diagnosis algorithms. Traditional deep learning methods often struggle when data is highly unbalanced with many cases in one class and only a few cases in another; new methods must be developed to overcome this challenge. We propose a novel activation function based on the generalized extreme value (GEV) distribution from extreme value theory, which improves performance over the traditional sigmoid activation function when one class significantly outweighs the other. We demonstrate the proposed activation function on a publicly available dataset and externally validate on a dataset consisting of 1,909 healthy chest X-rays and 84 COVID-19 X-rays. The proposed method achieves an improved area under the receiver operating characteristic (DeLong's p-value < 0.05) compared to the sigmoid activation. Our method is also demonstrated on a dataset of healthy and pneumonia vs. COVID-19 X-rays and a set of computerized tomography images, achieving improved sensitivity. The proposed GEV activation function significantly improves upon the previously used sigmoid activation for binary classification. This new paradigm is expected to play a significant role in the fight against COVID-19 and other diseases, with relatively few training cases available.

5 February 2020 in BMC Medical Research Methodology
Verified
Journal Article
Open access
PDF file
JBJoshua Bridge
Joshua Bridge
JBJohn D. Blakey
John D. Blakey
Laura BonnettLaura Bonnett
Laura Bonnett
Background: Clinical prediction models are widely used to guide medical advice and therapeutic interventions. Asthma is one of the most common chronic diseases globally and is characterised by acute deteriorations. These exacerbations are largely preventable, so there is interest in using clinical prediction models in this area. The objective of this review was to identify studies which have developed such models, determine whether consistent and appropriate methodology was used and whether statistically reliable prognostic models exist. Methods: We searched online databases MEDLINE (1948 onwards), CINAHL Plus (1937 onwards), The Cochrane Library, Web of Science (1898 onwards) and ClinicalTrials.gov, using index terms relating to asthma and prognosis. Data was extracted and assessment of quality was based on GRADE and an early version of PROBAST (Prediction study Risk of Bias Assessment Tool). A meta-analysis of the discrimination and calibration measures was carried out to determine overall performance across models. Results: Ten unique prognostic models were identified. GRADE identified moderate risk of bias in two of the studies, but more detailed quality assessment via PROBAST highlighted that most models were developed using highly selected and small datasets, incompletely recorded predictors and outcomes, and incomplete methodology. None of the identified models modelled recurrent exacerbations, instead favouring either presence/absence of an event, or time to first or specified event. Preferred methodologies were logistic regression and Cox proportional hazards regression. The overall pooled c-statistic was 0.77 (95% confidence interval 0.73 to 0.80), though individually some models performed no better than chance. The meta-analysis had an I2 value of 99.75% indicating a high amount of heterogeneity between studies. The majority of studies were small and did not include internal or external validation, therefore the individual performance measures are likely to be optimistic. Conclusions: Current prognostic models for asthma exacerbations are heterogeneous in methodology, but reported c-statistics suggest a clinically useful model could be created. Studies were consistent in lacking robust validation and in not modelling serial events. Further research is required with respect to incorporating recurrent events, and to externally validate tools in large representative populations to demonstrate the generalizability of published results.

15 October 2020 in BMJ Open Ophthalmology
Verified
Journal Article
Open access
PDF file
JBJoshua Bridge
Joshua Bridge
SHSimon Harding
Simon Harding
Yalin ZhengYalin Zheng
Yalin Zheng
Objective To develop a prognostic tool to predict the progression of age-related eye disease progression using longitudinal colour fundus imaging. Methods and analysis Previous prognostic models using deep learning with imaging data require annotation during training or only use a single time point. We propose a novel deep learning method to predict the progression of diseases using longitudinal imaging data with uneven time intervals, which requires no prior feature extraction. Given previous images from a patient, our method aims to predict whether the patient will progress onto the next stage of the disease. The proposed method uses InceptionV3 to produce feature vectors for each image. In order to account for uneven intervals, a novel interval scaling is proposed. Finally, a recurrent neural network is used to prognosticate the disease. We demonstrate our method on a longitudinal dataset of colour fundus images from 4903 eyes with age-related macular degeneration (AMD), taken from the Age-Related Eye Disease Study, to predict progression to late AMD. Results Our method attains a testing sensitivity of 0.878, a specificity of 0.887 and an area under the receiver operating characteristic of 0.950. We compare our method to previous methods, displaying superior performance in our model. Class activation maps display how the network reaches the final decision. Conclusion The proposed method can be used to predict progression to advanced AMD at some future visit. Using multiple images at different time points improves predictive performance.

21 May 2024 in European Urology Open Science
Verified
Journal Article
Open access
JBJoshua Bridge
Joshua Bridge
MLMuhieddine Labban
Muhieddine Labban
ACAlexander P. Cole
Alexander P. Cole
BABusola Adebusoye
Busola Adebusoye
SSSarah C. Smith
Sarah C. Smith
CMCaroline M. Moore
Caroline M. Moore
Radical prostatectomy (RP) is an established treatment for localised prostate cancer that can have a significant impact on urinary and sexual function, with recovery over time. Our aim was to describe functional recovery in the first year after RP, reporting descriptive outcomes alongside validated patient-reported outcome measure scores (Expanded Prostate Cancer Index Composite, EPIC-26). Men undergoing RP between September 2015 and November 2019 completed EPIC-26 at baseline and 1, 3, 6, and 12 mo. Overall, 2030 men consented to participation, underwent RP, and completed EPIC-26. At baseline, 97% were pad-free (1928/1996; 95% confidence interval [CI] 96–97%) and 77% were leak-free and pad-free (1529/1996; 95% CI 75–78), with a median EPIC-26 incontinence domain score of 100 (interquartile range [IQR] 86–100). At 12 mo, 65% were pad-free (904/1388; 95% CI 63–68%) and 42% were leak-free and pad-free (583/1388; 95% CI 39–45%), with a median EPIC-26 score of 76 (IQR 61–100). While one in three men reported wearing a pad at 12 mo, fewer than one in ten men needed more than 1 pad/d. At baseline, 1.9% reported a “moderate or big problem” with urine leakage, which increased to 9.7% at 12 mo. At baseline, the median sexual domain score among 1880 men was 74 (IQR 43–92) and 52% had erections sufficient for intercourse without medication (975/1880; 95% CI 50–54%). Among these 975 men, 630 responded at 12 mo, of whom 17% reported sufficient erections for intercourse (105/630; 95% CI 14–20%), without medication in 6% (37/630; 95% CI 4–8%) and needing medication in 11% (68/630; 95% CI 9–13%); the median EPIC-26 domain score was 26 (IQR 13–57). Reporting of functional outcomes after RP in terms of easily understood concepts such as pad-free and leak-free status, and erections with and with medication, alongside the classical report using EPIC-26 domain scores, increases the understanding of RP recovery patterns over the first year. At 12 months after surgery for prostate cancer, one in ten men reported a moderate or big problem with urine leakage and one in five men reported sufficient erections.

29 March 2022 in Journal of Arrhythmia
Verified
Journal Article
Open access
JBJoshua Bridge
Joshua Bridge
LFLu Fu
Lu Fu
WLWeidong Lin
Weidong Lin
YXYumei Xue
Yumei Xue
Gregory LipGregory Lip
Gregory Lip
Yalin ZhengYalin Zheng
Yalin Zheng
Background: Electrocardiogram (ECG) interpretation is an integral part of the clinical ECG workflow; however, this process is often time‐consuming and labor‐intensive. We aim to develop a rapid, inexpensive means to detect abnormal ECGs using artificial intelligence (AI) from scanned ECG printouts.Methods: The study included 1172 12‐lead ECG scans performed in 1172 individuals from a community in Guangzhou, China; 878 (74.9%) were diagnosed with sinus rhythm, and the remaining 294 (25.1%) with abnormal rhythms. A deep learning model consisting of a convolutional neural network based on InceptionV3 and a fully connected layer followed by a GEV activation was trained to classify scanned tracings as either normal or abnormal.Results: In a hold‐out testing set, the model achieved a area under curve (AUC), sensitivity, specificity, PPV, and NPV of 0.932 (95% confidence interval [CI]: 0.890, 0.976), 0.816 (95% CI: 0.657, 0.923), 0.993 (95% CI: 0.959, 1.0), 0.969 (95% CI: 0.838, 0.999), and 0.950 (95% CI: 0.90, 0.980) respectively, when using a probability threshold of 0.5. When compared with a physiological expert, these results show comparable performance with a statistically significant increase in specificity and a non‐significant decrease in sensitivity at the 95% level.Conclusions: We have developed a rapid, inexpensive, accurate means to detect abnormal ECGs using AI. Easy and accurate identification of such “abnormal” ECGs could allow the mass automated review of ECGs in community settings where abnormal ones could be flagged using AI for detailed clinical review by healthcare professionals.

16 December 2022 in IEEE Transactions on Intelligent Transportation Systems
Verified
Journal Article
Closed access
YMYanda Meng
Yanda Meng
JBJoshua Bridge
Joshua Bridge
Yifan ZhaoYifan Zhao
Yifan Zhao
MJMartha Joddrell
Martha Joddrell
YQYihong Qiao
Yihong Qiao
Yalin ZhengYalin Zheng
Yalin Zheng
This paper proposes an adaptive auxiliary task learning-based approach for transport object counting problems such as humans and vehicles. These problems are essential in many real-world tasks such as video surveillance, traffic monitoring, public security, and urban planning, to aid intelligent transportation systems. Unlike existing auxiliary task learning-based methods, we develop an attention-enhanced adaptively shared backbone network to enable both task-shared and task-tailored features that are learned in an end-to-end manner. The network seamlessly combines a standard Convolution Neural Network ( CNN ) and a Graph Convolution Network ( GCN ) for feature extraction and feature reasoning among different domains of tasks. Our approach gains enriched contextual information by iteratively and hierarchically fusing features across different task branches of the adaptive CNN backbone. The whole framework pays special attention to objects’ spatial locations and varied density levels, informed by object (or crowd) segmentation and density level segmentation auxiliary tasks. In particular, thanks to the proposed dilated contrastive density loss function, our network benefits from individual and regional context supervision, along with strengthened robustness. Experiments on six challenging multi-domain datasets demonstrate that our method achieves superior performance compared with state-of-the-art auxiliary task learning-based counting methods. Our code is publicly available.