Jack Breen

Matthias Eisenmann, Annika Reinke, Vivienn Weru et al. · utoronto

The number of international benchmarking competitions is steadily increasing in various fields of machine learning (ML) research and practice. So far, however, little is known about the common practice as well as bottlenecks faced by the community in tackling the research questions posed. To shed light on the status quo of algorithm development in the specific field of biomedical imaging analysis, we designed an international survey that was issued to all participants of challenges conducted in conjunction with the IEEE ISBI 2021 and MICCAI 2021 conferences (80 competitions in total). The survey covered participants' expertise and working environments, their chosen strategies, as well as algorithm characteristics. A median of 72% challenge participants took part in the survey. According to our results, knowledge exchange was the primary incentive (70%) for participation, while the reception of prize money played only a minor role (16%). While a median of 80 working hours was spent on method development, a large portion of participants stated that they did not have enough time for method development (32%). 25% perceived the infrastructure to be a bottleneck. Overall, 94% of all solutions were deep learning-based. Of these, 84% were based on standard architectures. 43% of the respondents reported that the data samples (e.g., images) were too large to be processed at once. This was most commonly addressed by patch-based training (69%), downsampling (37%), and solving 3D analysis tasks as a series of 2D tasks. K-fold cross-validation on the training set was performed by only 37% of the participants and only 50% of the participants performed ensembling based on multiple identical models (61%) or heterogeneous models (39%). 48% of the respondents applied postprocessing steps.

Marc Aubreville, Nikolas Stathonikos, Christof A. Bertram et al.

The density of mitotic figures within tumor tissue is known to be highly correlated with tumor proliferation and thus is an important marker in tumor grading. Recognition of mitotic figures by pathologists is known to be subject to a strong inter-rater bias, which limits the prognostic value. State-of-the-art deep learning methods can support the expert in this assessment but are known to strongly deteriorate when applied in a different clinical environment than was used for training. One decisive component in the underlying domain shift has been identified as the variability caused by using different whole slide scanners. The goal of the MICCAI MIDOG 2021 challenge has been to propose and evaluate methods that counter this domain shift and derive scanner-agnostic mitosis detection algorithms. The challenge used a training set of 200 cases, split across four scanning systems. As a test set, an additional 100 cases split across four scanning systems, including two previously unseen scanners, were given. The best approaches performed on an expert level, with the winning algorithm yielding an F_1 score of 0.748 (CI95: 0.704-0.781). In this paper, we evaluate and compare the approaches that were submitted to the challenge and identify methodological factors contributing to better performance.

Jack Breen, Katie Allen, Kieran Zucker et al.

Purpose - To characterise and assess the quality of published research evaluating artificial intelligence (AI) methods for ovarian cancer diagnosis or prognosis using histopathology data. Methods - A search of PubMed, Scopus, Web of Science, CENTRAL, and WHO-ICTRP was conducted up to 19/05/2023. The inclusion criteria required that research evaluated AI on histopathology images for diagnostic or prognostic inferences in ovarian cancer. The risk of bias was assessed using PROBAST. Information about each model of interest was tabulated and summary statistics were reported. PRISMA 2020 reporting guidelines were followed. Results - 1573 records were identified, of which 45 were eligible for inclusion. There were 80 models of interest, including 37 diagnostic models, 22 prognostic models, and 21 models with other diagnostically relevant outcomes. Models were developed using 1-1375 slides from 1-776 ovarian cancer patients. Model outcomes included treatment response (11/80), malignancy status (10/80), stain quantity (9/80), and histological subtype (7/80). All models were found to be at high or unclear risk of bias overall, with most research having a high risk of bias in the analysis and a lack of clarity regarding participants and predictors in the study. Research frequently suffered from insufficient reporting and limited validation using small sample sizes. Conclusion - Limited research has been conducted on the application of AI to histopathology images for diagnostic or prognostic purposes in ovarian cancer, and none of the associated models have been demonstrated to be ready for real-world implementation. Key aspects to help ensure clinical translation include more transparent and comprehensive reporting of data provenance and modelling approaches, as well as improved quantitative performance evaluation using cross-validation and external validations.

Jack Breen, Katie Allen, Kieran Zucker et al.

Weakly-supervised classification of histopathology slides is a computationally intensive task, with a typical whole slide image (WSI) containing billions of pixels to process. We propose Discriminative Region Active Sampling for Multiple Instance Learning (DRAS-MIL), a computationally efficient slide classification method using attention scores to focus sampling on highly discriminative regions. We apply this to the diagnosis of ovarian cancer histological subtypes, which is an essential part of the patient care pathway as different subtypes have different genetic and molecular profiles, treatment options, and patient outcomes. We use a dataset of 714 WSIs acquired from 147 epithelial ovarian cancer patients at Leeds Teaching Hospitals NHS Trust to distinguish the most common subtype, high-grade serous carcinoma, from the other four subtypes (low-grade serous, endometrioid, clear cell, and mucinous carcinomas) combined. We demonstrate that DRAS-MIL can achieve similar classification performance to exhaustive slide analysis, with a 3-fold cross-validated AUC of 0.8679 compared to 0.8781 with standard attention-based MIL classification. Our approach uses at most 18% as much memory as the standard approach, while taking 33% of the time when evaluating on a GPU and only 14% on a CPU alone. Reducing prediction time and memory requirements may benefit clinical deployment and the democratisation of AI, reducing the extent to which computational hardware limits end-user adoption.

Jack Breen, Kieran Zucker, Katie Allen et al.

The rapid growth of digital pathology in recent years has provided an ideal opportunity for the development of artificial intelligence-based tools to improve the accuracy and efficiency of clinical diagnoses. One of the significant roadblocks to current research is the high level of visual variability across digital pathology images, causing models to generalise poorly to unseen data. Stain normalisation aims to standardise the visual profile of digital pathology images without changing the structural content of the images. In this chapter, we explore different techniques which have been used for stain normalisation in digital pathology, with a focus on approaches which utilise generative adversarial networks (GANs). Typically, GAN-based methods outperform non-generative approaches but at the cost of much greater computational requirements. However, it is not clear which method is best for stain normalisation in general, with different GAN and non-GAN approaches outperforming each other in different scenarios and according to different performance metrics. This is an ongoing field of study as researchers aim to identify a method which efficiently and effectively normalises pathology images to make AI models more robust and generalisable.

Jack Breen, Katie Allen, Kieran Zucker et al.

For many patients, current ovarian cancer treatments offer limited clinical benefit. For some therapies, it is not possible to predict patients' responses, potentially exposing them to the adverse effects of treatment without any therapeutic benefit. As part of the automated prediction of treatment effectiveness in ovarian cancer using histopathological images (ATEC23) challenge, we evaluated the effectiveness of deep learning to predict whether a course of treatment including the antiangiogenic drug bevacizumab could contribute to remission or prevent disease progression for at least 6 months in a set of 282 histopathology whole slide images (WSIs) from 78 ovarian cancer patients. Our approach used a pretrained Hierarchical Image Pyramid Transformer (HIPT) to extract region-level features and an attention-based multiple instance learning (ABMIL) model to aggregate features and classify whole slides. The optimal HIPT-ABMIL model had an internal balanced accuracy of 60.2% +- 2.9% and an AUC of 0.646 +- 0.033. Histopathology-specific model pretraining was found to be beneficial to classification performance, though hierarchical transformers were not, with a ResNet feature extractor achieving similar performance. Due to the dataset being small and highly heterogeneous, performance was variable across 5-fold cross-validation folds, and there were some extreme differences between validation and test set performance within folds. The model did not generalise well to tissue microarrays, with accuracy worse than random chance. It is not yet clear whether ovarian cancer WSIs contain information that can be used to accurately predict treatment response, with further validation using larger, higher-quality datasets required.

Jack Breen, Katie Allen, Kieran Zucker et al.

Computer vision models are increasingly capable of classifying ovarian epithelial cancer subtypes, but they differ from pathologists by processing small tissue patches at a single resolution. Multi-resolution graph models leverage the spatial relationships of patches at multiple magnifications, learning the context for each patch. In this study, we conduct the most thorough validation of a graph model for ovarian cancer subtyping to date. Seven models were tuned and trained using five-fold cross-validation on a set of 1864 whole slide images (WSIs) from 434 patients treated at Leeds Teaching Hospitals NHS Trust. The cross-validation models were ensembled and evaluated using a balanced hold-out test set of 100 WSIs from 30 patients, and an external validation set of 80 WSIs from 80 patients in the Transcanadian Study. The best-performing model, a graph model using 10x+20x magnification data, gave balanced accuracies of 73%, 88%, and 99% in cross-validation, hold-out testing, and external validation, respectively. However, this only exceeded the performance of attention-based multiple instance learning in external validation, with a 93% balanced accuracy. Graph models benefitted greatly from using the UNI foundation model rather than an ImageNet-pretrained ResNet50 for feature extraction, with this having a much greater effect on performance than changing the subsequent classification approach. The accuracy of the combined foundation model and multi-resolution graph network offers a step towards the clinical applicability of these models, with a new highest-reported performance for this task, though further validations are still required to ensure the robustness and usability of the models.

Jack Breen, Katie Allen, Kieran Zucker et al.

Large pretrained transformers are increasingly being developed as generalised foundation models which can underpin powerful task-specific artificial intelligence models. Histopathology foundation models show great promise across many tasks, but analyses have typically been limited by arbitrary hyperparameters that were not tuned to the specific task. We report the most rigorous single-task validation of histopathology foundation models to date, specifically in ovarian cancer morphological subtyping. Attention-based multiple instance learning classifiers were compared using three ImageNet-pretrained feature extractors and fourteen histopathology foundation models. The training set consisted of 1864 whole slide images from 434 ovarian carcinoma cases at Leeds Teaching Hospitals NHS Trust. Five-class classification performance was evaluated through five-fold cross-validation, and these cross-validation models were ensembled for hold-out testing and external validation on the Transcanadian Study and OCEAN Challenge datasets. The best-performing model used the H-optimus-0 foundation model, with five-class balanced accuracies of 89%, 97%, and 74% in the test sets. Normalisations and augmentations aided the performance of the ImageNet-pretrained ResNets, but these were still outperformed by 13 of the 14 foundation models. Hyperparameter tuning the downstream classifiers improved performance by a median 1.9% balanced accuracy, with many improvements being statistically significant. Histopathology foundation models offer a clear benefit to ovarian cancer subtyping, improving classification performance to a degree where clinical utility is tangible, albeit with an increased computational burden. Such models could provide a second opinion to histopathologists diagnosing challenging cases and may improve the accuracy, objectivity, and efficiency of pathological diagnoses overall.

Jack Breen, Kieran Zucker, Nicolas M. Orsi et al.

Breast cancer is the most commonly diagnosed cancer worldwide, with over two million new cases each year. During diagnostic tumour grading, pathologists manually count the number of dividing cells (mitotic figures) in biopsy or tumour resection specimens. Since the process is subjective and time-consuming, data-driven artificial intelligence (AI) methods have been developed to automatically detect mitotic figures. However, these methods often generalise poorly, with performance reduced by variations in tissue types, staining protocols, or the scanners used to digitise whole-slide images. Domain adaptation approaches have been adopted in various applications to mitigate this issue of domain shift. We evaluate two unsupervised domain adaptation methods, CycleGAN and Neural Style Transfer, using the MIDOG 2021 Challenge dataset. This challenge focuses on detecting mitotic figures in whole-slide images digitised using different scanners. Two baseline mitosis detection models based on U-Net and RetinaNet were investigated in combination with the aforementioned domain adaptation methods. Both baseline models achieved human expert level performance, but had reduced performance when evaluated on images which had been digitised using a different scanner. The domain adaptation techniques were each found to be beneficial for detection with data from some scanners but not for others, with the only average increase across all scanners being achieved by CycleGAN on the RetinaNet detector. These techniques require further refinement to ensure consistency in mitosis detection.