Publications
353 results found
Monteiro M, De Sousa Ribeiro F, Pawlowski N, et al., 2023, Measuring axiomatic soundness of counterfactual image models, International Conference on Learning Representations (ICLR)
We use the axiomatic definition of counterfactual to derive metrics that enable quantifying the correctness of approximate counterfactual inference models.Abstract: We present a general framework for evaluating image counterfactuals. The power and flexibility of deep generative models make them valuable tools for learning mechanisms in structural causal models. However, their flexibility makes counterfactual identifiability impossible in the general case.Motivated by these issues, we revisit Pearl's axiomatic definition of counterfactuals to determine the necessary constraints of any counterfactual inference model: composition, reversibility, and effectiveness. We frame counterfactuals as functions of an input variable, its parents, and counterfactual parents and use the axiomatic constraints to restrict the set of functions that could represent the counterfactual, thus deriving distance metrics between the approximate and ideal functions. We demonstrate how these metrics can be used to compare and choose between different approximate counterfactual inference models and to provide insight into a model's shortcomings and trade-offs.
Pati S, Baid U, Edwards B, et al., 2023, Author Correction: Federated learning enables big data for rare cancer boundary detection., Nature Communications, Vol: 14, Pages: 436-436, ISSN: 2041-1723
Batten J, Sinclair M, Glocker B, et al., 2023, Image To Tree with Recursive Prompting
Extracting complex structures from grid-based data is a common key step inautomated medical image analysis. The conventional solution to recoveringtree-structured geometries typically involves computing the minimal cost paththrough intermediate representations derived from segmentation masks. However,this methodology has significant limitations in the context of projectiveimaging of tree-structured 3D anatomical data such as coronary arteries, sincethere are often overlapping branches in the 2D projection. In this work, wepropose a novel approach to predicting tree connectivity structure whichreformulates the task as an optimization problem over individual steps of arecursive process. We design and train a two-stage model which leverages theUNet and Transformer architectures and introduces an image-based promptingtechnique. Our proposed method achieves compelling results on a pair ofsynthetic datasets, and outperforms a shortest-path baseline.
Rosnati M, Roschewitz M, Glocker B, 2023, Robust semi-supervised segmentation with timestep ensembling diffusion models, Pages: 512-527
Medical image segmentation is a challenging task, made more difficult by many datasets’ limited size and annotations. Denoising diffusion probabilistic models (DDPM) have recently shown promise in modelling the distribution of natural images and were successfully applied to various medical imaging tasks. This work focuses on semi-supervised image segmentation using diffusion models, particularly addressing domain generalisation. Firstly, we demonstrate that smaller diffusion steps generate latent representations that are more robust for downstream tasks than larger steps. Secondly, we use this insight to propose an improved ensembling scheme that leverages information-dense small steps and the regularising effect of larger steps to generate predictions. Our model shows significantly better performance in domain-shifted settings while retaining competitive performance in-domain. Overall, this work highlights the potential of DDPMs for semi-supervised medical image segmentation and provides insights into optimising their performance under domain shift.
Islam M, Glocker B, 2023, Frequency Dropout: Feature-Level Regularization via Randomized Filtering
Deep convolutional neural networks have shown remarkable performance on various computer vision tasks, and yet, they are susceptible to picking up spurious correlations from the training signal. So called ‘shortcuts’ can occur during learning, for example, when there are specific frequencies present in the image data that correlate with the output predictions. Both high and low frequencies can be characteristic of the underlying noise distribution caused by the image acquisition rather than in relation to the task-relevant information about the image content. Models that learn features related to this characteristic noise will not generalize well to new data. In this work, we propose a simple yet effective training strategy, Frequency Dropout, to prevent convolutional neural networks from learning frequency-specific imaging features. We employ randomized filtering of feature maps during training which acts as a feature-level regularization. In this study, we consider common image processing filters such as Gaussian smoothing, Laplacian of Gaussian, and Gabor filtering. Our training strategy is model-agnostic and can be used for any computer vision task. We demonstrate the effectiveness of Frequency Dropout on a range of popular architectures and multiple tasks including image classification, domain adaptation, and semantic segmentation using both computer vision and medical imaging datasets. Our results suggest that the proposed approach does not only improve predictive accuracy but also improves robustness against domain shift.
Roschewitz M, Glocker B, 2023, Distance Matters For Improving Performance Estimation Under Covariate Shift, Pages: 4551-4561
Performance estimation under covariate shift is a crucial component of safe AI model deployment, especially for sensitive use-cases. Recently, several solutions were proposed to tackle this problem, most leveraging model predictions or softmax confidence to derive accuracy estimates. However, under dataset shifts confidence scores may become ill-calibrated if samples are too far from the training distribution. In this work, we show that taking into account distances of test samples to their expected training distribution can significantly improve performance estimation under covariate shift. Precisely, we introduce a "distance-check"to flag samples that lie too far from the expected distribution, to avoid relying on their untrustworthy model outputs in the accuracy estimation step. We demonstrate the effectiveness of this method on 13 image classification tasks, across a wide-range of natural and synthetic distribution shifts and hundreds of models, with a median relative MAE improvement of 27% over the best baseline across all tasks, and SOTA performance on 10 out of 13 tasks. Our code is publicly available at https://github.com/melanibe/distance-matters-performance-estimation.
Puyol-Antón E, Feragen A, King AP, et al., 2023, FAIMI Preface, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol: 14242 LNCS, ISSN: 0302-9743
Piçarra C, Glocker B, 2023, Analysing Race and Sex Bias in Brain Age Prediction, Pages: 194-204, ISSN: 0302-9743
Brain age prediction from MRI has become a popular imaging biomarker associated with a wide range of neuropathologies. The datasets used for training, however, are often skewed and imbalanced regarding demographics, potentially making brain age prediction models susceptible to bias. We analyse the commonly used ResNet-34 model by conducting a comprehensive subgroup performance analysis and feature inspection. The model is trained on 1,215 T1-weighted MRI scans from Cam-CAN and IXI, and tested on UK Biobank (n=42,786), split into six racial and biological sex subgroups. With the objective of comparing the performance between subgroups, measured by the absolute prediction error, we use a Kruskal-Wallis test followed by two post-hoc Conover-Iman tests to inspect bias across race and biological sex. To examine biases in the generated features, we use PCA for dimensionality reduction and employ two-sample Kolmogorov-Smirnov tests to identify distribution shifts among subgroups. Our results reveal statistically significant differences in predictive performance between Black and White, Black and Asian, and male and female subjects. Seven out of twelve pairwise comparisons show statistically significant differences in the feature distributions. Our findings call for further analysis of brain age prediction models.
Jones C, Roschewitz M, Glocker B, 2023, The Role of Subgroup Separability in Group-Fair Medical Image Classification, Pages: 179-188, ISSN: 0302-9743
We investigate performance disparities in deep classifiers. We find that the ability of classifiers to separate individuals into subgroups varies substantially across medical imaging modalities and protected characteristics; crucially, we show that this property is predictive of algorithmic bias. Through theoretical analysis and extensive empirical evaluation (Code is available at https://github.com/biomedia-mira/subgroup-separability ), we find a relationship between subgroup separability, subgroup disparities, and performance degradation when models are trained on data with systematic bias such as underdiagnosis. Our findings shed new light on the question of how models become biased, providing important insights for the development of fair medical imaging AI.
Santhirasekaram A, Pinto K, Winkler M, et al., 2023, A Sheaf Theoretic Perspective for Robust Prostate Segmentation, Pages: 249-259, ISSN: 0302-9743
Deep learning based methods have become the most popular approach for prostate segmentation in MRI. However, domain variations due to the complex acquisition process result in textural differences as well as imaging artefacts which significantly affects the robustness of deep learning models for prostate segmentation across multiple sites. We tackle this problem by using multiple MRI sequences to learn a set of low dimensional shape components whose combinatorially large learnt composition is capable of accounting for the entire distribution of segmentation outputs. We draw on the language of cellular sheaf theory to model compositionality driven by local and global topological correctness. In our experiments, our method significantly improves the domain generalisability of anatomical and tumour segmentation of the prostate. Code is available at https://github.com/AinkaranSanthi/A-Sheaf-Theoretic-Perspective-for-Robust-Segmentation.git.
Dorent R, Kujawa A, Ivory M, et al., 2023, CrossMoDA 2021 challenge: Benchmark of cross-modality domain adaptation techniques for vestibular schwannoma and cochlea segmentation, Publisher: ELSEVIER
- Author Web Link
- Cite
- Citations: 3
Islam M, Li Z, Glocker B, 2023, Robustness Stress Testing in Medical Image Classification., Publisher: Springer, Pages: 167-176
Piçarra C, Winzeck S, Monteiro M, et al., 2023, Automatic localisation and per-region quantification of traumatic brain injury on head CT using atlas mapping, European Journal of Radiology Open, Vol: 10, Pages: 1-9, ISSN: 2352-0477
Rationale and objectivesTo develop a method for automatic localisation of brain lesions on head CT, suitable for both population-level analysis and lesion management in a clinical setting.Materials and methodsLesions were located by mapping a bespoke CT brain atlas to the patient’s head CT in which lesions had been previously segmented. The atlas mapping was achieved through robust intensity-based registration enabling the calculation of per-region lesion volumes. Quality control (QC) metrics were derived for automatic detection of failure cases. The CT brain template was built using 182 non-lesioned CT scans and an iterative template construction strategy. Individual brain regions in the CT template were defined via non-linear registration of an existing MRI-based brain atlas.Evaluation was performed on a multi-centre traumatic brain injury dataset (TBI) (n = 839 scans), including visual inspection by a trained expert. Two population-level analyses are presented as proof-of-concept: a spatial assessment of lesion prevalence, and an exploration of the distribution of lesion volume per brain region, stratified by clinical outcome.Results95.7% of the lesion localisation results were rated by a trained expert as suitable for approximate anatomical correspondence between lesions and brain regions, and 72.5% for more quantitatively accurate estimates of regional lesion load. The classification performance of the automatic QC showed an AUC of 0.84 when compared to binarised visual inspection scores. The localisation method has been integrated into the publicly available Brain Lesion Analysis and Segmentation Tool for CT (BLAST-CT).ConclusionAutomatic lesion localisation with reliable QC metrics is feasible and can be used for patient-level quantitative analysis of TBI, as well as for large-scale population analysis due to its computational efficiency (<2 min/scan on GPU).
Kori A, Sanchez P, Vilouras K, et al., 2023, A Causal Ordering Prior for Unsupervised Representation Learning., CoRR, Vol: abs/2307.05704
Santhirasekaram A, Winkler M, Rockall A, et al., 2023, Topology Preserving Compositionality for Robust Medical Image Segmentation, IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Publisher: IEEE COMPUTER SOC, Pages: 543-552, ISSN: 2160-7508
Mccradden MD, Odusi O, Joshi S, et al., 2023, What's fair is ... fair? Presenting JustEFAB, an ethical framework for operationalizing medical ethics and social justice in the integration of clinical machine learning, 6th ACM Conference on Fairness, Accountability, and Transparency (FAccT), Publisher: ASSOC COMPUTING MACHINERY, Pages: 1505-1519
Li Z, Kamnitsas K, Dou Q, et al., 2023, Joint Optimization of Class-Specific Training- and Test-Time Data Augmentation in Segmentation., CoRR, Vol: abs/2305.19084
Batten J, Sinclair M, Glocker B, et al., 2023, Image To Tree with Recursive Prompting., CoRR, Vol: abs/2301.00447
Rasal R, Castro DC, Pawlowski N, et al., 2023, Deep structural causal shape models, Computer Vision – ECCV 2022 Workshops, Publisher: Springer Nature Switzerland, Pages: 400-432, ISSN: 0302-9743
Causal reasoning provides a language to ask important interventional and counterfactual questions beyond purely statistical association. In medical imaging, for example, we may want to study the causal effect of genetic, environmental, or lifestyle factors on the normal and pathological variation of anatomical phenotypes. However, while anatomical shape models of 3D surface meshes, extracted from automated image segmentation, can be reliably constructed, there is a lack of computational tooling to enable causal reasoning about morphological variations. To tackle this problem, we propose deep structural causal shape models (CSMs), which utilise high-quality mesh generation techniques, from geometric deep learning, within the expressive framework of deep structural causal models. CSMs enable subject-specific prognoses through counterfactual mesh generation (“How would this patient’s brain structure change if they were ten years older?”), which is in contrast to most current works on purely population-level statistical shape modelling. We demonstrate the capabilities of CSMs at all levels of Pearl’s causal hierarchy through a number of qualitative and quantitative experiments leveraging a large dataset of 3D brain structures.
Dorent R, Kujawa A, Ivory M, et al., 2023, CrossMoDA 2021 challenge: Benchmark of cross-modality domain adaptation techniques for vestibular schwannoma and cochlea segmentation., Medical Image Anal., Vol: 83, Pages: 102628-102628
Osuala R, Kushibar K, Garrucho L, et al., 2023, Data synthesis and adversarial networks: A review and meta-analysis in cancer imaging., Medical Image Anal., Vol: 84, Pages: 102704-102704
Lekadir K, Feragen A, Fofanah AJ, et al., 2023, FUTURE-AI: International consensus guideline for trustworthy and deployable artificial intelligence in healthcare., CoRR, Vol: abs/2309.12325
Xu M, Islam M, Glocker B, et al., 2022, Confidence-Aware Paced-Curriculum Learning by Label Smoothing for Surgical Scene Understanding
Curriculum learning and self-paced learning are the training strategies thatgradually feed the samples from easy to more complex. They have captivatedincreasing attention due to their excellent performance in robotic vision. Mostrecent works focus on designing curricula based on difficulty levels in inputsamples or smoothing the feature maps. However, smoothing labels to control thelearning utility in a curriculum manner is still unexplored. In this work, wedesign a paced curriculum by label smoothing (P-CBLS) using paced learning withuniform label smoothing (ULS) for classification tasks and fuse uniform andspatially varying label smoothing (SVLS) for semantic segmentation tasks in acurriculum manner. In ULS and SVLS, a bigger smoothing factor value enforces aheavy smoothing penalty in the true label and limits learning less information.Therefore, we design the curriculum by label smoothing (CBLS). We set a biggersmoothing value at the beginning of training and gradually decreased it to zeroto control the model learning utility from lower to higher. We also designed aconfidence-aware pacing function and combined it with our CBLS to investigatethe benefits of various curricula. The proposed techniques are validated onfour robotic surgery datasets of multi-class, multi-label classification,captioning, and segmentation tasks. We also investigate the robustness of ourmethod by corrupting validation data into different severity levels. Ourextensive analysis shows that the proposed method improves prediction accuracyand robustness.
Gatidis S, Kart T, Fischer M, et al., 2022, Better together: data harmonization and cross-study analysis of abdominal MRI data from UK biobank and the German national cohort., Investigative Radiology, Vol: 58, Pages: 346-354, ISSN: 0020-9996
OBJECTIVES: The UK Biobank (UKBB) and German National Cohort (NAKO) are among the largest cohort studies, capturing a wide range of health-related data from the general population, including comprehensive magnetic resonance imaging (MRI) examinations. The purpose of this study was to demonstrate how MRI data from these large-scale studies can be jointly analyzed and to derive comprehensive quantitative image-based phenotypes across the general adult population. MATERIALS AND METHODS: Image-derived features of abdominal organs (volumes of liver, spleen, kidneys, and pancreas; volumes of kidney hilum adipose tissue; and fat fractions of liver and pancreas) were extracted from T1-weighted Dixon MRI data of 17,996 participants of UKBB and NAKO based on quality-controlled deep learning generated organ segmentations. To enable valid cross-study analysis, we first analyzed the data generating process using methods of causal discovery. We subsequently harmonized data from UKBB and NAKO using the ComBat approach for batch effect correction. We finally performed quantile regression on harmonized data across studies providing quantitative models for the variation of image-derived features stratified for sex and dependent on age, height, and weight. RESULTS: Data from 8791 UKBB participants (49.9% female; age, 63 ± 7.5 years) and 9205 NAKO participants (49.1% female, age: 51.8 ± 11.4 years) were analyzed. Analysis of the data generating process revealed direct effects of age, sex, height, weight, and the data source (UKBB vs NAKO) on image-derived features. Correction of data source-related effects resulted in markedly improved alignment of image-derived features between UKBB and NAKO. Cross-study analysis on harmonized data revealed comprehensive quantitative models for the phenotypic variation of abdominal organs across the general adult population. CONCLUSIONS: Cross-study analysis of MRI data from UKBB and NAKO as proposed in this work can be helpful for futur
Chalkidou A, Shokraneh F, Kijauskaite G, et al., 2022, Recommendations for the development and use of imaging test sets to investigate the test performance of artificial intelligence in health screening, LANCET DIGITAL HEALTH, Vol: 4, Pages: E899-E905
Kart T, Fischer M, Winzeck S, et al., 2022, Automated imaging-based abdominal organ segmentation and quality control in 20,000 participants of the UK Biobank and German National Cohort Studies, SCIENTIFIC REPORTS, Vol: 12, ISSN: 2045-2322
- Author Web Link
- Cite
- Citations: 2
Rosnati M, Soreq E, Monteiro M, et al., 2022, Automatic lesion analysis for increased efficiency in outcome prediction of traumatic brain injury, 5th International Workshop, MLCN 2022, Publisher: Springer Nature Switzerland, Pages: 135-146, ISSN: 0302-9743
The accurate prognosis for traumatic brain injury (TBI) patients is difficult yet essential to inform therapy, patient management, and long-term after-care. Patient characteristics such as age, motor and pupil responsiveness, hypoxia and hypotension, and radiological findings on computed tomography (CT), have been identified as important variables for TBI outcome prediction. CT is the acute imaging modality of choice in clinical practice because of its acquisition speed and widespread availability. However, this modality is mainly used for qualitative and semi-quantitative assessment, such as the Marshall scoring system, which is prone to subjectivity and human errors. This work explores the predictive power of imaging biomarkers extracted from routinely-acquired hospital admission CT scans using a state-of-the-art, deep learning TBI lesion segmentation method. We use lesion volumes and corresponding lesion statistics as inputs for an extended TBI outcome prediction model. We compare the predictive power of our proposed features to the Marshall score, independently and when paired with classic TBI biomarkers. We find that automatically extracted quantitative CT features perform similarly or better than the Marshall score in predicting unfavourable TBI outcomes. Leveraging automatic atlas alignment, we also identify frontal extra-axial lesions as important indicators of poor outcome. Our work may contribute to a better understanding of TBI, and provides new insights into how automated neuroimaging analysis can be used to improve prognostication after TBI.
Shehata N, Bain W, Glocker B, 2022, A Comparative Study of Graph Neural Networks for Shape Classification in Neuroimaging, Proceedings of Machine Learning Research, GeoMedIA Workshop
Graph neural networks have emerged as a promising approach for the analysisof non-Euclidean data such as meshes. In medical imaging, mesh-like data playsan important role for modelling anatomical structures, and shape classificationcan be used in computer aided diagnosis and disease detection. However, with aplethora of options, the best architectural choices for medical shape analysisusing GNNs remain unclear. We conduct a comparative analysis to providepractitioners with an overview of the current state-of-the-art in geometricdeep learning for shape classification in neuroimaging. Using biological sexclassification as a proof-of-concept task, we find that using FPFH as nodefeatures substantially improves GNN performance and generalisation toout-of-distribution data; we compare the performance of three alternativeconvolutional layers; and we reinforce the importance of data augmentation forgraph based learning. We then confirm these results hold for a clinicallyrelevant task, using the classification of Alzheimer's disease.
Satchwell L, Wedlake L, Greenlay E, et al., 2022, Development of machine learning support for reading whole body diffusion-weighted MRI (WB-MRI) in myeloma for the detection and quantification of the extent of disease before and after treatment (MALIMAR): protocol for a cross-sectional diagnostic test accuracy study, BMJ Open, Vol: 12, Pages: 1-9, ISSN: 2044-6055
Introduction Whole-body MRI (WB-MRI) is recommended by the National Institute of Clinical Excellence as the first-line imaging tool for diagnosis of multiple myeloma. Reporting WB-MRI scans requires expertise to interpret and can be challenging for radiologists who need to meet rapid turn-around requirements. Automated computational tools based on machine learning (ML) could assist the radiologist in terms of sensitivity and reading speed and would facilitate improved accuracy, productivity and cost-effectiveness. The MALIMAR study aims to develop and validate a ML algorithm to increase the diagnostic accuracy and reading speed of radiological interpretation of WB-MRI compared with standard methods.Methods and analysis This phase II/III imaging trial will perform retrospective analysis of previously obtained clinical radiology MRI scans and scans from healthy volunteers obtained prospectively to implement training and validation of an ML algorithm. The study will comprise three project phases using approximately 633 scans to (1) train the ML algorithm to identify active disease, (2) clinically validate the ML algorithm and (3) determine change in disease status following treatment via a quantification of burden of disease in patients with myeloma. Phase 1 will primarily train the ML algorithm to detect active myeloma against an expert assessment (‘reference standard’). Phase 2 will use the ML output in the setting of radiology reader study to assess the difference in sensitivity when using ML-assisted reading or human-alone reading. Phase 3 will assess the agreement between experienced readers (with and without ML) and the reference standard in scoring both overall burden of disease before and after treatment, and response.Ethics and dissemination MALIMAR has ethical approval from South Central—Oxford C Research Ethics Committee (REC Reference: 17/SC/0630). IRAS Project ID: 233501. CPMS Portfolio adoption (CPMS ID: 36766). Participants gave informe
Vasey B, Nagendran M, Campbell B, et al., 2022, Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI (May, 10.1038/s41591-022-01772-9, 2022), NATURE MEDICINE, Vol: 28, Pages: 2218-2218, ISSN: 1078-8956
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.