SSPT-bpMRI: A Self-supervised Pre-training Scheme for Improving Prostate Cancer Detection and Diagnosis in Bi-parametric MRI.

Prostate cancer (PCa) is one of the most prevalent cancers in men. Early diagnosis plays a pivotal role in reducing the mortality rate from clinically significant PCa (csPCa). In recent years, bi-parametric magnetic resonance imaging (bpMRI) has attracted great attention for the detection and diagnosis of csPCa. bpMRI is able to overcome some limitations of multi-parametric MRI (mpMRI) such as the use of contrast agents, the time-consuming for imaging and the costs, and achieve detection performance comparable to mpMRI. However, inter-reader agreements are currently low for prostate MRI. Advancements in artificial intelligence (AI) have propelled the development of deep learning (DL)-based computer-aided detection and diagnosis system (CAD). However, most of the existing DL models developed for csPCa identification are restricted by the scale of data and the scarcity in labels. In this paper, we propose a self-supervised pre-training scheme named SSPT-bpMRI with an image restoration pretext task integrating four different image transformations to improve the performance of DL algorithms. Specially, we explored the potential value of the self-supervised pre-training in fully supervised and weakly supervised situations. Experiments on the publicly available PI-CAI dataset demonstrate that our model outperforms the fully supervised or weakly supervised model alone.

Automated periodontitis bone loss diagnosis in panoramic radiographs using a bespoke two-stage detector.

Periodontitis is a serious oral disease that can lead to severe conditions such as bone loss and teeth falling out if left untreated. Diagnosis of radiographic bone loss (RBL) is critical for the staging and treatment of periodontitis. Unfortunately, the RBL diagnosis by examining the panoramic radiographs is time-consuming. The demand for automated image analysis is urgent. However, existing deep learning methods have limited performances in diagnosis accuracy and have certain difficulties in implementation. Hence, we propose a novel two-stage periodontitis detection convolutional neural network (PDCNN), where we optimize the detector with an anchor-free encoding that allows fast and accurate prediction. We also introduce a proposal-connection module in our detector that excludes less relevant regions of interests (ROIs), making the network focus on more relevant ROIs to improve detection accuracy. Furthermore, we introduced a large-scale, high-resolution panoramic radiograph dataset that captures various complex cases with professional periodontitis annotations. Experiments on our panoramic-image dataset show that the proposed approach achieved an RBL classification accuracy of 0.762. This result shows that our approach outperforms state-of-the-art detectors such as Faster R-CNN and YOLO-v4. We can conclude that the proposed method successfully improves the RBL detection performance. The dataset and our code have been released on GitHub. (https://github.com/PuckBlink/PDCNN).

A Deep-Learning Enabled Automatic Fetal Thalamus Diameter Measurement Algorithm.

The analysis of maternal factors that impact the normal development of the fetal thalamus is an emerging field of research and requires the retrospective measurement of fetal thalamus diameter (FTD). Unfortunately, FTD is not measured in routine 2D ultrasound (2D-US) screenings of fetuses. Manual measurement of FTD is a laborious, difficult, and error-prone process because the thalamus lacks well-defined boundaries in 2D-US images of the fetal brain as it has a similar echogenicity to the surrounding brain tissue. Traditional methods based on statistical shape models (SSMs) perform poorly in measuring FTD due to the noisy textures and fuzzy edges of the fetal thalamus in 2D-US images of the fetal brain. To overcome these difficulties, we propose a deep learning-based automatic FTD measurement algorithm, FTDNet. FTDNet measures FTD by learning to directly detect the measurement landmarks through supervised learning. The algorithm first detects the region of the brain that contains the thalamus structure, and then focuses on processing that region for FTD landmark detection. Our FTD dataset, developed through a consensus between two ultrasonographers, contains 1,111 pairs of landmark coordinates for measuring FTD and verified bounding boxes surrounding the fetal thalamus. To assess FTDNet's measurement consistency compared to the ground truth, we used the intraclass correlation coefficient (ICC). FTDNet achieved an ICC score of 0.734, significantly outperforming the prior SSM method and other baseline comparison methods. Our findings are an important step forward in understanding the maternal factors which influence fetal brain development.Clinical relevance- This work proposes an end-to-end thalamus detection and measurement algorithm for measuring fetal thalamus diameter. Our work represents a significant step in the research of how maternal factors can impact fetal thalamus development. The development of an automatic and accurate method for measuring FTD through deep learning has the potential to greatly advance this field of study.

The Checkpoint Program: Collaborative Care to Reduce the Reliance of Frequent Presenters on ED.

INTRODUCTION: Growing pressures upon Emergency Departments [ED] call for new ways of working with frequent presenters who, although small in number, place extensive demands on services, to say nothing of the costs and consequences for the patients themselves. EDs are often poorly equipped to address the multi-dimensional nature of patient need and the complex circumstances surrounding repeated presentation. Employing a model of intensive short-term community-based case management, the Checkpoint program sought to improve care coordination for this patient group, thereby reducing their reliance on ED. METHOD: This study employed a single group interrupted time series design, evaluating patient engagement with the program and year-on-year individual differences in the number of ED visits pre and post enrolment. Associated savings were also estimated. RESULTS: Prior to intervention, there were two dominant modes in the ED presentation trends of patients. One group had a steady pattern with ≥7 presentations in each of the last four years. The other group had an increasing trend in presentations, peaking in the 12 months immediately preceding enrolment. Following the intervention, both groups demonstrated two consecutive year-on-year reductions. By the second year, and from an overall peak of 22.5 presentations per patient per annum, there was a 53% reduction in presentations. This yielded approximate savings of $7100 per patient. DISCUSSION: Efforts to improve care coordination, when combined with proactive case management in the community, can impact positively on ED re-presentation rates, provided they are concerted, sufficiently intensive and embed the principles of integration. CONCLUSION: The Checkpoint program demonstrated sufficient promise to warrant further exploration of its sustainability. However, health services have yet to determine the ideal organisational structures and funding arrangements to support such initiatives.

A Mobile App and Dashboard for Early Detection of Infectious Disease Outbreaks: Development Study.

BACKGROUND: Outbreaks of infectious diseases pose great risks, including hospitalization and death, to public health. Therefore, improving the management of outbreaks is important for preventing widespread infection and mitigating associated risks. Mobile health technology provides new capabilities that can help better capture, monitor, and manage infectious diseases, including the ability to quickly identify potential outbreaks. OBJECTIVE: This study aims to develop a new infectious disease surveillance (IDS) system comprising a mobile app for accurate data capturing and dashboard for better health care planning and decision making. METHODS: We developed the IDS system using a 2-pronged approach: a literature review on available and similar disease surveillance systems to understand the fundamental requirements and face-to-face interviews to collect specific user requirements from the local public health unit team at the Nepean Hospital, Nepean Blue Mountains Local Health District, New South Wales, Australia. RESULTS: We identified 3 fundamental requirements when designing an electronic IDS system, which are the ability to capture and report outbreak data accurately, completely, and in a timely fashion. We then developed our IDS system based on the workflow, scope, and specific requirements of the public health unit team. We also produced detailed design and requirement guidelines. In our system, the outbreak data are captured and sent from anywhere using a mobile device or a desktop PC (web interface). The data are processed using a client-server architecture and, therefore, can be analyzed in real time. Our dashboard is designed to provide a daily, weekly, monthly, and historical summary of outbreak information, which can be potentially used to develop a future intervention plan. Specific information about certain outbreaks can also be visualized interactively to understand the unique characteristics of emerging infectious diseases. CONCLUSIONS: We demonstrated the design and development of our IDS system. We suggest that the use of a mobile app and dashboard will simplify the overall data collection, reporting, and analysis processes, thereby improving the public health responses and providing accurate registration of outbreak information. Accurate data reporting and collection are a major step forward in creating a better intervention plan for future outbreaks of infectious diseases.

Unsupervised Domain Adaptation to Classify Medical Images Using Zero-Bias Convolutional Auto-Encoders and Context-Based Feature Augmentation.

The accuracy and robustness of image classification with supervised deep learning are dependent on the availability of large-scale labelled training data. In medical imaging, these large labelled datasets are sparse, mainly related to the complexity in manual annotation. Deep convolutional neural networks (CNNs), with transferable knowledge, have been employed as a solution to limited annotated data through: 1) fine-tuning generic knowledge with a relatively smaller amount of labelled medical imaging data, and 2) learning image representation that is invariant to different domains. These approaches, however, are still reliant on labelled medical image data. Our aim is to use a new hierarchical unsupervised feature extractor to reduce reliance on annotated training data. Our unsupervised approach uses a multi-layer zero-bias convolutional auto-encoder that constrains the transformation of generic features from a pre-trained CNN (for natural images) to non-redundant and locally relevant features for the medical image data. We also propose a context-based feature augmentation scheme to improve the discriminative power of the feature representation. We evaluated our approach on 3 public medical image datasets and compared it to other state-of-the-art supervised CNNs. Our unsupervised approach achieved better accuracy when compared to other conventional unsupervised methods and baseline fine-tuned CNNs.

Convolutional sparse kernel network for unsupervised medical image analysis.

The availability of large-scale annotated image datasets and recent advances in supervised deep learning methods enable the end-to-end derivation of representative image features that can impact a variety of image analysis problems. Such supervised approaches, however, are difficult to implement in the medical domain where large volumes of labelled data are difficult to obtain due to the complexity of manual annotation and inter- and intra-observer variability in label assignment. We propose a new convolutional sparse kernel network (CSKN), which is a hierarchical unsupervised feature learning framework that addresses the challenge of learning representative visual features in medical image analysis domains where there is a lack of annotated training data. Our framework has three contributions: (i) we extend kernel learning to identify and represent invariant features across image sub-patches in an unsupervised manner. (ii) We initialise our kernel learning with a layer-wise pre-training scheme that leverages the sparsity inherent in medical images to extract initial discriminative features. (iii) We adapt a multi-scale spatial pyramid pooling (SPP) framework to capture subtle geometric differences between learned visual features. We evaluated our framework in medical image retrieval and classification on three public datasets. Our results show that our CSKN had better accuracy when compared to other conventional unsupervised methods and comparable accuracy to methods that used state-of-the-art supervised convolutional neural networks (CNNs). Our findings indicate that our unsupervised CSKN provides an opportunity to leverage unannotated big data in medical imaging repositories.

Development of a risk predictive scoring system to identify patients at risk of representation to emergency department: a retrospective population-based analysis in Australia.

OBJECTIVE: To examine the characteristics of frequent visitors (FVs) to emergency departments (EDs) and develop a predictive model to identify those with high risk of a future representations to ED among younger and general population (aged ≤70 years). DESIGN AND SETTING: A retrospective analysis of ED data targeting younger and general patients (aged ≤70 years) were collected between 1 January 2009 and 30 June 2016 from a public hospital in Australia. PARTICIPANTS: A total of 343 014 ED presentations were identified from 170 134 individual patients. MAIN OUTCOME MEASURES: Proportion of FVs (those attending four or more times annually), demographic characteristics (age, sex, indigenous and marital status), mode of separation (eg, admitted to ward), triage categories, time of arrival to ED, referral on departure and clinical conditions. Statistical estimates using a mixed-effects model to develop a risk predictive scoring system. RESULTS: The FVs were characterised by young adulthood (32.53%) to late-middle (26.07%) aged patients with a higher proportion of indigenous (5.7%) and mental health-related presentations (10.92%). They were also more likely to arrive by ambulance (36.95%) and leave at own risk without completing their treatments (9.8%). They were also highly associated with socially disadvantage groups such as people who have been divorced, widowed or separated (12.81%). These findings were then used for the development of a predictive model to identify potential FVs. The performance of our derived risk predictive model was favourable with an area under the receiver operating characteristic (ie, C-statistic) of 65.7%. CONCLUSION: The development of a demographic and clinical profile of FVs coupled with the use of predictive model can highlight the gaps in interventions and identify new opportunities for better health outcome and planning.

Dermoscopic Image Segmentation via Multistage Fully Convolutional Networks.

OBJECTIVE: Segmentation of skin lesions is an important step in the automated computer aided diagnosis of melanoma. However, existing segmentation methods have a tendency to over- or under-segment the lesions and perform poorly when the lesions have fuzzy boundaries, low contrast with the background, inhomogeneous textures, or contain artifacts. Furthermore, the performance of these methods are heavily reliant on the appropriate tuning of a large number of parameters as well as the use of effective preprocessing techniques, such as illumination correction and hair removal. METHODS: We propose to leverage fully convolutional networks (FCNs) to automatically segment the skin lesions. FCNs are a neural network architecture that achieves object detection by hierarchically combining low-level appearance information with high-level semantic information. We address the issue of FCN producing coarse segmentation boundaries for challenging skin lesions (e.g., those with fuzzy boundaries and/or low difference in the textures between the foreground and the background) through a multistage segmentation approach in which multiple FCNs learn complementary visual characteristics of different skin lesions; early stage FCNs learn coarse appearance and localization information while late-stage FCNs learn the subtle characteristics of the lesion boundaries. We also introduce a new parallel integration method to combine the complementary information derived from individual segmentation stages to achieve a final segmentation result that has accurate localization and well-defined lesion boundaries, even for the most challenging skin lesions. RESULTS: We achieved an average Dice coefficient of 91.18% on the ISBI 2016 Skin Lesion Challenge dataset and 90.66% on the PH2 dataset. CONCLUSION AND SIGNIFICANCE: Our extensive experimental results on two well-established public benchmark datasets demonstrate that our method is more effective than other state-of-the-art methods for skin lesion segmentation.

Saliency-Based Lesion Segmentation Via Background Detection in Dermoscopic Images.

The segmentation of skin lesions in dermoscopic images is a fundamental step in automated computer-aided diagnosis of melanoma. Conventional segmentation methods, however, have difficulties when the lesion borders are indistinct and when contrast between the lesion and the surrounding skin is low. They also perform poorly when there is a heterogeneous background or a lesion that touches the image boundaries; this then results in under- and oversegmentation of the skin lesion. We suggest that saliency detection using the reconstruction errors derived from a sparse representation model coupled with a novel background detection can more accurately discriminate the lesion from surrounding regions. We further propose a Bayesian framework that better delineates the shape and boundaries of the lesion. We also evaluated our approach on two public datasets comprising 1100 dermoscopic images and compared it to other conventional and state-of-the-art unsupervised (i.e., no training required) lesion segmentation methods, as well as the state-of-the-art unsupervised saliency detection methods. Our results show that our approach is more accurate and robust in segmenting lesions compared to other methods. We also discuss the general extension of our framework as a saliency optimization algorithm for lesion segmentation.