Using visual attention estimation on videos for automated prediction of autism spectrum disorder and symptom severity in preschool children.

Atypical visual attention in individuals with autism spectrum disorders (ASD) has been utilised as a unique diagnosis criterion in previous research. This paper presents a novel approach to the automatic and quantitative screening of ASD as well as symptom severity prediction in preschool children. We develop a novel computational pipeline that extracts learned features from a dynamic visual stimulus to classify ASD children and predict the level of ASD-related symptoms. Experimental results demonstrate promising performance that is superior to using handcrafted features and machine learning algorithms, in terms of evaluation metrics used in diagnostic tests. Using a leave-one-out cross-validation approach, we obtained an accuracy of 94.59%, a sensitivity of 100%, a specificity of 76.47% and an area under the receiver operating characteristic curve (AUC) of 96% for ASD classification. In addition, we obtained an accuracy of 94.74%, a sensitivity of 87.50%, a specificity of 100% and an AUC of 99% for ASD symptom severity prediction.

Evaluating a Human Detection Model in a Behaviour Analysis Pipeline for Suicide Prevention.

Suicides in public places, such as railways, can have a significant impact on bystanders, railway staff, first responders and the surrounding communities. Behaviours prior to a suicide attempt have been identified, that could potentially be detected automatically. As a first step, the algorithm is required to accurately identify individuals exhibiting these behaviours in different settings. Our study analyses a human detection model focussing on pedestrian detection at railway stations as one component of a broader project to detect pre-suicidal behaviours. Closed-circuit television footage from two stations collected for the same 24-hour period were manually analysed to obtain parameters (true positives, false positives, and false negatives) which were then used to compute performance measures (sensitivity, precision, and F1 score). The model performed differently in both stations with a sensitivity of 0.73 and F1 score of 0.84 in Station A and a sensitivity of 0.48 and F1 score of 0.65 in Station B. Root causes of false negatives identified include differing body postures and occlusion. Although the model was adequate, its performance is dependent on the view captured by the cameras in stations. Collectively, these findings can be used to improve the model's performance.Clinical Relevance-Detecting behaviours prior to a suicide attempt offers a critical period for intervention by bystanders or first responders, potentially interrupting the attempt. This offers the potential to directly reduce suicide attempts, as well as reduce third-party exposure to these traumatic events.

Artificial intelligence in the diagnosis of glaucoma and neurodegenerative diseases.

Artificial Intelligence is a rapidly expanding field within computer science that encompasses the emulation of human intelligence by machines. Machine learning and deep learning - two primary data-driven pattern analysis approaches under the umbrella of artificial intelligence - has created considerable interest in the last few decades. The evolution of technology has resulted in a substantial amount of artificial intelligence research on ophthalmic and neurodegenerative disease diagnosis using retinal images. Various artificial intelligence-based techniques have been used for diagnostic purposes, including traditional machine learning, deep learning, and their combinations. Presented here is a review of the literature covering the last 10 years on this topic, discussing the use of artificial intelligence in analysing data from different modalities and their combinations for the diagnosis of glaucoma and neurodegenerative diseases. The performance of published artificial intelligence methods varies due to several factors, yet the results suggest that such methods can potentially facilitate clinical diagnosis. Generally, the accuracy of artificial intelligence-assisted diagnosis ranges from 67-98%, and the area under the sensitivity-specificity curve (AUC) ranges from 0.71-0.98, which outperforms typical human performance of 71.5% accuracy and 0.86 area under the curve. This indicates that artificial intelligence-based tools can provide clinicians with useful information that would assist in providing improved diagnosis. The review suggests that there is room for improvement of existing artificial intelligence-based models using retinal imaging modalities before they are incorporated into clinical practice.

Radiological Diagnosis of Chronic Liver Disease and Hepatocellular Carcinoma: A Review.

Medical image analysis plays a pivotal role in the evaluation of diseases, including screening, surveillance, diagnosis, and prognosis. Liver is one of the major organs responsible for key functions of metabolism, protein and hormone synthesis, detoxification, and waste excretion. Patients with advanced liver disease and Hepatocellular Carcinoma (HCC) are often asymptomatic in the early stages; however delays in diagnosis and treatment can lead to increased rates of decompensated liver diseases, late-stage HCC, morbidity and mortality. Ultrasound (US) is commonly used imaging modality for diagnosis of chronic liver diseases that includes fibrosis, cirrhosis and portal hypertension. In this paper, we first provide an overview of various diagnostic methods for stages of liver diseases and discuss the role of Computer-Aided Diagnosis (CAD) systems in diagnosing liver diseases. Second, we review the utility of machine learning and deep learning approaches as diagnostic tools. Finally, we present the limitations of existing studies and outline future directions to further improve diagnostic accuracy, as well as reduce cost and subjectivity, while also improving workflow for the clinicians.

hist2RNA: An Efficient Deep Learning Architecture to Predict Gene Expression from Breast Cancer Histopathology Images.

Gene expression can be used to subtype breast cancer with improved prediction of risk of recurrence and treatment responsiveness over that obtained using routine immunohistochemistry (IHC). However, in the clinic, molecular profiling is primarily used for ER+ breast cancer, which is costly, tissue destructive, requires specialised platforms, and takes several weeks to obtain a result. Deep learning algorithms can effectively extract morphological patterns in digital histopathology images to predict molecular phenotypes quickly and cost-effectively. We propose a new, computationally efficient approach called hist2RNA inspired by bulk RNA sequencing techniques to predict the expression of 138 genes (incorporated from 6 commercially available molecular profiling tests), including luminal PAM50 subtype, from hematoxylin and eosin (H&E)-stained whole slide images (WSIs). The training phase involves the aggregation of extracted features for each patient from a pretrained model to predict gene expression at the patient level using annotated H&E images from The Cancer Genome Atlas (TCGA, n = 335). We demonstrate successful gene prediction on a held-out test set (n = 160, corr = 0.82 across patients, corr = 0.29 across genes) and perform exploratory analysis on an external tissue microarray (TMA) dataset (n = 498) with known IHC and survival information. Our model is able to predict gene expression and luminal PAM50 subtype (Luminal A versus Luminal B) on the TMA dataset with prognostic significance for overall survival in univariate analysis (c-index = 0.56, hazard ratio = 2.16 (95% CI 1.12-3.06), p < 5 × 10-3), and independent significance in multivariate analysis incorporating standard clinicopathological variables (c-index = 0.65, hazard ratio = 1.87 (95% CI 1.30-2.68), p < 5 × 10-3). The proposed strategy achieves superior performance while requiring less training time, resulting in less energy consumption and computational cost compared to patch-based models. Additionally, hist2RNA predicts gene expression that has potential to determine luminal molecular subtypes which correlates with overall survival, without the need for expensive molecular testing.

A Multi-Scale Context Aware Attention Model for Medical Image Segmentation.

Medical image segmentation is critical for efficient diagnosis of diseases and treatment planning. In recent years, convolutional neural networks (CNN)-based methods, particularly U-Net and its variants, have achieved remarkable results on medical image segmentation tasks. However, they do not always work consistently on images with complex structures and large variations in regions of interest (ROI). This could be due to the fixed geometric structure of the receptive fields used for feature extraction and repetitive down-sampling operations that lead to information loss. To overcome these problems, the standard U-Net architecture is modified in this work by replacing the convolution block with a dilated convolution block to extract multi-scale context features with varying sizes of receptive fields, and adding a dilated inception block between the encoder and decoder paths to alleviate the problem of information recession and the semantic gap between features. Furthermore, the input of each dilated convolution block is added to the output through a squeeze and excitation unit, which alleviates the vanishing gradient problem and improves overall feature representation by re-weighting the channel-wise feature responses. The original inception block is modified by reducing the size of the spatial filter and introducing dilated convolution to obtain a larger receptive field. The proposed network was validated on three challenging medical image segmentation tasks with varying size ROIs: lung segmentation on chest X-ray (CXR) images, skin lesion segmentation on dermoscopy images and nucleus segmentation on microscopy cell images. Improved performance compared to state-of-the-art techniques demonstrates the effectiveness and generalisability of the proposed Dilated Convolution and Inception blocks-based U-Net (DCI-UNet).

Cancer Survival Prediction From Whole Slide Images With Self-Supervised Learning and Slide Consistency.

Histopathological Whole Slide Images (WSIs) at giga-pixel resolution are the gold standard for cancer analysis and prognosis. Due to the scarcity of pixel- or patch-level annotations of WSIs, many existing methods attempt to predict survival outcomes based on a three-stage strategy that includes patch selection, patch-level feature extraction and aggregation. However, the patch features are usually extracted by using truncated models (e.g. ResNet) pretrained on ImageNet without fine-tuning on WSI tasks, and the aggregation stage does not consider the many-to-one relationship between multiple WSIs and the patient. In this paper, we propose a novel survival prediction framework that consists of patch sampling, feature extraction and patient-level survival prediction. Specifically, we employ two kinds of self-supervised learning methods, i.e. colorization and cross-channel, as pretext tasks to train convnet-based models that are tailored for extracting features from WSIs. Then, at the patient-level survival prediction we explicitly aggregate features from multiple WSIs, using consistency and contrastive losses to normalize slide-level features at the patient level. We conduct extensive experiments on three large-scale datasets: TCGA-GBM, TCGA-LUSC and NLST. Experimental results demonstrate the effectiveness of our proposed framework, as it achieves state-of-the-art performance in comparison with previous studies, with concordance index of 0.670, 0.679 and 0.711 on TCGA-GBM, TCGA-LUSC and NLST, respectively.

Eye-tracking correlates of response to joint attention in preschool children with autism spectrum disorder.

BACKGROUND: A number of differences in joint attention behaviour between children with autism spectrum disorder (ASD) and typically developing (TD) individuals have previously been documented. METHOD: We use eye-tracking technology to assess response to joint attention (RJA) behaviours in 77 children aged 31 to 73 months. We conducted a repeated-measures analysis of variance to identify differences between groups. In addition, we analysed correlations between eye-tracking and clinical measures using Spearman's correlation. RESULTS: The children diagnosed with ASD were less likely to follow gaze compared to TD children. Children with ASD were less accurate at gaze following when only eye gaze information was available, compared to when eye gaze with head movement was observed. Higher accuracy gaze-following profiles were associated with better early cognition and more adaptive behaviours in children with ASD. Less accurate gaze-following profiles were associated with more severe ASD symptomatology. CONCLUSION: There are differences in RJA behaviours between ASD and TD preschool children. Several eye-tracking measures of RJA behaviours in preschool children were found to be associated with clinical measures for ASD diagnosis. This study also highlights the construct validity of using eye-tracking measures as potential biomarkers in the assessment and diagnosis of ASD in preschool children.

Drusen-aware model for age-related macular degeneration recognition.

INTRODUCTION: The purpose of this study was to build an automated age-related macular degeneration (AMD) colour fundus photography (CFP) recognition method that incorporates confounders (other ocular diseases) and normal age-related changes by using drusen masks for spatial feature supervision. METHODS: A range of clinical sources were used to acquire 7588 CFPs. Contrast limited adaptive histogram equalisation was used for pre-processing. ResNet50 was used as the backbone network, and a spatial attention block was added to integrate prior knowledge of drusen features into the backbone. The evaluation metrics used were sensitivity, specificity and F1 score, which is the harmonic mean of precision and recall (sensitivity) and area under the receiver-operating characteristic (AUC). Fivefold cross-validation was performed, and the results compared with four other methods. RESULTS: Excellent discrimination results were obtained with the algorithm. On the public dataset (n = 6565), the proposed method achieved a mean (SD) sensitivity of 0.54 (0.09), specificity of 0.99 (0.00), F1 score of 0.62 (0.06) and AUC of 0.92 (0.02). On the private dataset (n = 1023), the proposed method achieved a sensitivity of 0.92 (0.02), specificity of 0.98 (0.01), F1 score of 0.92 (0.01) and AUC of 0.98 (0.01). CONCLUSION: The proposed drusen-aware model outperformed baseline and other vessel feature-based methods in F1 and AUC on the AMD/normal CFP classification task and achieved comparable results on datasets that included other diseases that often confound classification. The method also improved results when a five-category grading protocol was used, thereby reflecting discriminative ability of the algorithm within a real-life clinical setting.

Ensemble feature selection with data-driven thresholding for Alzheimer's disease biomarker discovery.

BACKGROUND: Feature selection is often used to identify the important features in a dataset but can produce unstable results when applied to high-dimensional data. The stability of feature selection can be improved with the use of feature selection ensembles, which aggregate the results of multiple base feature selectors. However, a threshold must be applied to the final aggregated feature set to separate the relevant features from the redundant ones. A fixed threshold, which is typically used, offers no guarantee that the final set of selected features contains only relevant features. This work examines a selection of data-driven thresholds to automatically identify the relevant features in an ensemble feature selector and evaluates their predictive accuracy and stability. Ensemble feature selection with data-driven thresholding is applied to two real-world studies of Alzheimer's disease. Alzheimer's disease is a progressive neurodegenerative disease with no known cure, that begins at least 2-3 decades before overt symptoms appear, presenting an opportunity for researchers to identify early biomarkers that might identify patients at risk of developing Alzheimer's disease. RESULTS: The ensemble feature selectors, combined with data-driven thresholds, produced more stable results, on the whole, than the equivalent individual feature selectors, showing an improvement in stability of up to 34%. The most successful data-driven thresholds were the robust rank aggregation threshold and the threshold algorithm threshold from the field of information retrieval. The features identified by applying these methods to datasets from Alzheimer's disease studies reflect current findings in the AD literature. CONCLUSIONS: Data-driven thresholds applied to ensemble feature selectors provide more stable, and therefore more reproducible, selections of features than individual feature selectors, without loss of performance. The use of a data-driven threshold eliminates the need to choose a fixed threshold a-priori and can select a more meaningful set of features. A reliable and compact set of features can produce more interpretable models by identifying the factors that are important in understanding a disease.

Understanding and detecting behaviours prior to a suicide attempt: A mixed-methods study.

OBJECTIVE: Prior research suggests there are observable behaviours preceding suicide attempts in public places. However, there are currently no ways to continually monitor such sites, limiting the potential to intervene. In this mixed-methods study, we examined the acceptability and feasibility of using an automated computer system to identify crisis behaviours. METHODS: First, we conducted a large-scale acceptability survey to assess public perceptions on research using closed-circuit television and artificial intelligence for suicide prevention. Second, we identified crisis behaviours at a frequently used cliff location by manual structured analysis of closed-circuit television footage. Third, we configured a computer vision algorithm to identify crisis behaviours and evaluated its sensitivity and specificity using test footage. RESULTS: Overall, attitudes were positive towards research using closed-circuit television and artificial intelligence for suicide prevention, including among those with lived experience. The second study revealed that there are identifiable behaviours, including repetitive pacing and an extended stay. Finally, the automated behaviour recognition algorithm was able to correctly identify 80% of acted crisis clips and correctly reject 90% of acted non-crisis clips. CONCLUSION: The results suggest that using computer vision to detect behaviours preceding suicide is feasible and well accepted by the community and may be a feasible method of initiating human contact during a crisis.

Imbalanced classification for protein subcellular localization with multilabel oversampling.

MOTIVATION: Subcellular localization of human proteins is essential to comprehend their functions and roles in physiological processes, which in turn helps in diagnostic and prognostic studies of pathological conditions and impacts clinical decision-making. Since proteins reside at multiple locations at the same time and few subcellular locations host far more proteins than other locations, the computational task for their subcellular localization is to train a multilabel classifier while handling data imbalance. In imbalanced data, minority classes are underrepresented, thus leading to a heavy bias towards the majority classes and the degradation of predictive capability for the minority classes. Furthermore, data imbalance in multilabel settings is an even more complex problem due to the coexistence of majority and minority classes. RESULTS: Our studies reveal that based on the extent of concurrence of majority and minority classes, oversampling of minority samples through appropriate data augmentation techniques holds promising scope for boosting the classification performance for the minority classes. We measured the magnitude of data imbalance per class and the concurrence of majority and minority classes in the dataset. Based on the obtained values, we identified minority and medium classes, and a new oversampling method is proposed that includes non-linear mixup, geometric and colour transformations for data augmentation and a sampling approach to prepare minibatches. Performance evaluation on the Human Protein Atlas Kaggle challenge dataset shows that the proposed method is capable of achieving better predictions for minority classes than existing methods. AVAILABILITY AND IMPLEMENTATION: Data used in this study are available at https://www.kaggle.com/competitions/human-protein-atlas-image-classification/data. Source code is available at https://github.com/priyarana/Protein-subcellular-localisation-method. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

A Federated Learning System for Histopathology Image Analysis With an Orchestral Stain-Normalization GAN.

Currently, data-driven based machine learning is considered one of the best choices in clinical pathology analysis, and its success is subject to the sufficiency of digitized slides, particularly those with deep annotations. Although centralized training on a large data set may be more reliable and more generalized, the slides to the examination are more often than not collected from many distributed medical institutes. This brings its own challenges, and the most important is the assurance of privacy and security of incoming data samples. In the discipline of histopathology image, the universal stain-variation issue adds to the difficulty of an automatic system as different clinical institutions provide distinct stain styles. To address these two important challenges in AI-based histopathology diagnoses, this work proposes a novel conditional Generative Adversarial Network (GAN) with one orchestration generator and multiple distributed discriminators, to cope with multiple-client based stain-style normalization. Implemented within a Federated Learning (FL) paradigm, this framework well preserves data privacy and security. Additionally, the training consistency and stability of the distributed system are further enhanced by a novel temporal self-distillation regularization scheme. Empirically, on large cohorts of histopathology datasets as a benchmark, the proposed model matches the performance of conventional centralized learning very closely. It also outperforms state-of-the-art stain-style transfer methods on the downstream Federated Learning image classification task, with an accuracy increase of over 20.0% in comparison to the baseline classification model.

Data augmentation with improved regularisation and sampling for imbalanced blood cell image classification.

Due to progression in cell-cycle or duration of storage, classification of morphological changes in human blood cells is important for correct and effective clinical decisions. Automated classification systems help avoid subjective outcomes and are more efficient. Deep learning and more specifically Convolutional Neural Networks have achieved state-of-the-art performance on various biomedical image classification problems. However, real-world data often suffers from the data imbalance problem, owing to which the trained classifier is biased towards the majority classes and does not perform well on the minority classes. This study presents an imbalanced blood cells classification method that utilises Wasserstein divergence GAN, mixup and novel nonlinear mixup for data augmentation to achieve oversampling of the minority classes. We also present a minority class focussed sampling strategy, which allows effective representation of minority class samples produced by all three data augmentation techniques and contributes to the classification performance. The method was evaluated on two publicly available datasets of immortalised human T-lymphocyte cells and Red Blood Cells. Classification performance evaluated using F1-score shows that our proposed approach outperforms existing methods on the same datasets.

Blood-based transcriptomic signature panel identification for cancer diagnosis: benchmarking of feature extraction methods.

Liquid biopsy has shown promise for cancer diagnosis due to its minimally invasive nature and the potential for novel biomarker discovery. However, the low concentration of relevant blood-based biosources and the heterogeneity of samples (i.e. the variability of relative abundance of molecules identified), pose major challenges to biomarker discovery. Moreover, the number of molecular measurements or features (e.g. transcript read counts) per sample could be in the order of several thousand, whereas the number of samples is often substantially lower, leading to the curse of dimensionality. These challenges, among others, elucidate the importance of a robust biomarker panel identification or feature extraction step wherein relevant molecular measurements are identified prior to classification for cancer detection. In this work, we performed a benchmarking study on 12 feature extraction methods using transcriptomic profiles derived from different blood-based biosources. The methods were assessed both in terms of their predictive performance and the robustness of the biomarker panels in diagnosing cancer or stratifying cancer subtypes. While performing the comparison, the feature extraction methods are categorized into feature subset selection methods and transformation methods. A transformation feature extraction method, namely partial least square discriminant analysis, was found to perform consistently superior in terms of classification performance. As part of the benchmarking study, a generic pipeline has been created and made available as an R package to ensure reproducibility of the results and allow for easy extension of this study to other datasets (https://github.com/VafaeeLab/bloodbased-pancancer-diagnosis).

Towards automated coronary artery segmentation: A systematic review.

BACKGROUND AND OBJECTIVE: Vessel segmentation is the first processing stage of 3D medical images for both clinical and research use. Current segmentation methods are tedious and time consuming, requiring significant manual correction and hence are infeasible to use in large data sets. METHODS: Here, we review and analyse available coronary artery segmentation methods, focusing on fully automated methods capable of handling the rapidly growing medical images available. All manuscripts published since 2010 are systematically reviewed, categorised into different groups based on the approach taken, and characteristics of the different approaches as well as trends over the past decade are explored. RESULTS: The manuscripts were divided intro three broad categories, consisting of region growing, voxelwise prediction and partitioning approaches. The most common approach overall was region growing, particularly using active contour models, however these have had a sharp fall in popularity in recent years with convolutional neural networks becoming significantly more popular. CONCLUSIONS: The systematic review of current coronary artery segmentation methods shows interesting trends, with rising popularity of machine learning methods, a focus on efficient methods, and falling popularity of computationally expensive processing steps such as vesselness and multiplanar reformation.

Transient wall shear stress estimation in coronary bifurcations using convolutional neural networks.

BACKGROUND AND OBJECTIVE: Haemodynamic metrics, such as blood flow induced shear stresses at the inner vessel lumen, are associated with the development and progression of coronary artery disease. Understanding these metrics may therefore improve the assessment of an individual's coronary disease risk. However, the calculation of such luminal Wall Shear Stress (WSS) using traditional Computational Fluid Dynamics (CFD) methods is relatively slow and computationally expensive. As a result, CFD based haemodynamic computation is not suitable for integrated and large-scale use in clinical settings. METHODS: In this work, deep learning techniques are proposed as an alternative method to CFD, whereby luminal WSS magnitude can be predicted in coronary bifurcations throughout the cardiac cycle based on the steady state solution (which takes <120 seconds to calculate including preprocessing), vessel geometry and additional global features. The deep learning model is trained on a dataset of 101 patient-specific and 2626 synthetic left main bifurcation models with 26 separate patient-specific cases used as the test set. RESULTS: The model showed high fidelity predictions with <5% (normalised against mean WSS magnitude) deviation to CFD derived values as the gold-standard method, while being orders of magnitude faster with on average <2 minutes versus 3 hours computation for transient CFD. CONCLUSIONS: This method therefore offers a new approach to substantially reduce the computational cost involved in, for example, large-scale population studies of coronary haemodynamic metrics, and may therefore open the pathway for future clinical integration.

Towards Building a Visual Behaviour Analysis Pipeline for Suicide Detection and Prevention.

Understanding human behaviours through video analysis has seen significant research progress in recent years with the advancement of deep learning. This topic is of great importance to the next generation of intelligent visual surveillance systems which are capable of real-time detection and analysis of human behaviours. One important application is to automatically monitor and detect individuals who are in crisis at suicide hotspots to facilitate early intervention and prevention. However, there is still a significant gap between research in human action recognition and visual video processing in general, and their application to monitor hotspots for suicide prevention. While complex backgrounds, non-rigid movements of pedestrians and limitations of surveillance cameras and multi-task requirements for a surveillance system all pose challenges to the development of such systems, a further challenge is the detection of crisis behaviours before a suicide attempt is made, and there is a paucity of datasets in this area due to privacy and confidentiality issues. Most relevant research only applies to detecting suicides such as hangings or jumps from bridges, providing no potential for early prevention. In this research, these problems are addressed by proposing a new modular design for an intelligent visual processing pipeline that is capable of pedestrian detection, tracking, pose estimation and recognition of both normal actions and high risk behavioural cues that are important indicators of a suicide attempt. Specifically, based on the key finding that human body gestures can be used for the detection of social signals that potentially precede a suicide attempt, a new 2D skeleton-based action recognition algorithm is proposed. By using a two-branch network that takes advantage of three types of skeleton-based features extracted from a sequence of frames and a stacked LSTM structure, the model predicts the action label at each time step. It achieved good performance on both the public dataset JHMDB and a smaller private CCTV footage collection on action recognition. Moreover, a logical layer, which uses knowledge from a human coding study to recognise pre-suicide behaviour indicators, has been built on top of the action recognition module to compensate for the small dataset size. It enables complex behaviour patterns to be recognised even from smaller datasets. The whole pipeline has been tested in a real-world application of suicide prevention using simulated footage from a surveillance system installed at a suicide hotspot, and preliminary results confirm its effectiveness at capturing crisis behaviour indicators for early detection and prevention of suicide.

A new and automated risk prediction of coronary artery disease using clinical endpoints and medical imaging-derived patient-specific insights: protocol for the retrospective GeoCAD cohort study.

INTRODUCTION: Coronary artery disease (CAD) is the leading cause of death worldwide. More than a quarter of cardiovascular events are unexplained by current absolute cardiovascular disease risk calculators, and individuals without clinical risk factors have been shown to have worse outcomes. The 'anatomy of risk' hypothesis recognises that adverse anatomical features of coronary arteries enhance atherogenic haemodynamics, which in turn mediate the localisation and progression of plaques. We propose a new risk prediction method predicated on CT coronary angiography (CTCA) data and state-of-the-art machine learning methods based on a better understanding of anatomical risk for CAD. This may open new pathways in the early implementation of personalised preventive therapies in susceptible individuals as a potential key in addressing the growing burden of CAD. METHODS AND ANALYSIS: GeoCAD is a retrospective cohort study in 1000 adult patients who have undergone CTCA for investigation of suspected CAD. It is a proof-of-concept study to test the hypothesis that advanced image-derived patient-specific data can accurately predict long-term cardiovascular events. The objectives are to (1) profile CTCA images with respect to variations in anatomical shape and associated haemodynamic risk expressing, at least in part, an individual's CAD risk, (2) develop a machine-learning algorithm for the rapid assessment of anatomical risk directly from unprocessed CTCA images and (3) to build a novel CAD risk model combining traditional risk factors with these novel anatomical biomarkers to provide a higher accuracy CAD risk prediction tool. ETHICS AND DISSEMINATION: The study protocol has been approved by the St Vincent's Hospital Human Research Ethics Committee, Sydney-2020/ETH02127 and the NSW Population and Health Service Research Ethics Committee-2021/ETH00990. The project outcomes will be published in peer-reviewed and biomedical journals, scientific conferences and as a higher degree research thesis.

Supervised and semi-supervised 3D organ localisation in CT images combining reinforcement learning with imitation learning.

Computer aided diagnostics often requires analysis of a region of interest (ROI) within a radiology scan, and the ROI may be an organ or a suborgan. Although deep learning algorithms have the ability to outperform other methods, they rely on the availability of a large amount of annotated data. Motivated by the need to address this limitation, an approach to localisation and detection of multiple organs based on supervised and semi-supervised learning is presented here. It draws upon previous work by the authors on localising the thoracic and lumbar spine region in CT images. The method generates six bounding boxes of organs of interest, which are then fused to a single bounding box. The results of experiments on localisation of the Spleen, Left and Right Kidneys in CT Images using supervised and semi supervised learning (SSL) demonstrate the ability to address data limitations with a much smaller data set and fewer annotations, compared to other state-of-the-art methods. The SSL performance was evaluated using three different mixes of labelled and unlabelled data (i.e. 30:70,35:65,40:60) for each of lumbar spine, spleen left and right kidneys respectively. The results indicate that SSL provides a workable alternative especially in medical imaging where it is difficult to obtain annotated data.