Artificial intelligence (AI) has shown promising results for computational pathology tasks. However, one of the limitations in clinical practice is that these algorithms are optimised for the distribution represented by the training data. For out-of-distribution (OOD) data, they often deliver predictions with equal confidence, even though these often are incorrect. In the pursuit of OOD detection in digital pathology, this study evaluates the state-of-the-art (SOTA) in computational pathology OOD detection, based on diffusion probabilistic models, specifically by adapting the latent diffusion model (LDM) for this purpose (AnoLDM). We compare this against post-hoc methods based on the latent space of foundation models, which are SOTA in general computer vision research. The approaches are not only evaluated on data from the same medical centres as the training set, but also on several datasets with data distribution shifts. The results show that AnoLDM performs similarly well or better than diffusion model based approaches published in previous studies in computational pathology but with reduced computational costs. However, our optimal configuration of an approach based on foundation models (kang_residual) outperforms AnoLDM on OOD detection on data not experiencing any covariate shifts, with an AUROC of 96.17 versus 91.86. Interestingly, AnoLDM is more successful at handling the data distribution shifts investigated in this study. However, both AnoLDM and kang_residual suffer substantial loss in the performance under the data distribution shifts, hence future work should focus on improving the generalisation of OOD detection for computational pathology applications.

Multiple-instance learning (MIL) is an attractive approach for digital pathology applications as it reduces the costs related to data collection and labelling. However, it is not clear how sensitive MIL is to clinically realistic domain shifts, i.e., differences in data distribution that could negatively affect performance, and if already existing metrics for detecting domain shifts work well with these algorithms. We trained an attention-based MIL algorithm to classify whether a whole-slide image of a lymph node contains breast tumour metastases. The algorithm was evaluated on data from a hospital in a different country and various subsets of this data that correspond to different levels of domain shift. Our contributions include showing that MIL for digital pathology is affected by clinically realistic differences in data, evaluating which features from a MIL model are most suitable for detecting changes in performance, and proposing an unsupervised metric named Fréchet Domain Distance (FDD) for quantification of domain shifts. Shift measure performance was evaluated through the mean Pearson correlation to change in classification performance, where FDD achieved 0.70 on 10-fold cross-validation models. The baselines included Deep ensemble, Difference of Confidence, and Representation shift which resulted in 0.45, -0.29, and 0.56 mean Pearson correlation, respectively. FDD could be a valuable tool for care providers and vendors who need to verify if a MIL system is likely to perform reliably when implemented at a new site, without requiring any additional annotations from pathologists.

Deep learning (DL) is a subfield of artificial intelligence (AI) focused on developing algorithms that learn from data to perform some tasks that can aid humans in their daily life or work assignments. Research demonstrates the potential of DL in supporting pathologists with routine tasks like detecting breast cancer metastases and grading prostate cancer. However, a widespread adoption of DL technology in pathology labs has been slow for several reasons. DL models often exhibit performance variations across medical centres, patient subgroups, and even within the same centre over time. While collecting more data and retraining the algorithms seems like a straightforward solution, it is a costly and time-consuming process. Moreover, retraining DL systems with regulatory approvals is complex due to existing regulations. Another limitation of DL models is their inability to provide confidence estimates for predictions, leaving users in the dark about their reliability. Finally, establishing a close collaboration between the research community, vendors, and pathology labs is crucial for producing effective DL systems for patient care. However, this collaboration faces challenges like miscommunication, misalignment of goals, and misunderstanding priorities.

This thesis presents various approaches that could tackle the generalisation and reliability challenges faced by diagnostic DL systems for digital pathology with a strong emphasis on the clinical needs. To address the generalisation issues, an unsupervised approach to quantify expected changes in a model’s performance between two datasets is proposed. This approach can serve as an initial validation step before deploying diagnostic DL systems in clinical practice, reducing annotation costs. Additionally, an unsupervised framework based on generative models is proposed to identify substantially different inputs, known as out-of-distribution (OOD) samples. Detecting OOD samples plays a crucial role in enhancing the reliability of DL algorithms. Furthermore, several studies are conducted to explore what benefits uncertainty estimation could bring. Firstly, various uncertainty estimation approaches are extensively evaluated, focusing on identifying incorrect predictions and generalisability issues between medical centres and specific patient groups. In addition, the results reveal that combining uncertainty estimation methods with DL outputs leads to a more robust classification score, enhancing the overall performance and reliability of the classification process. Another study demonstrates that spatial uncertainty aggregation improves the effectiveness of uncertainty estimation in tumour segmentation tasks. This is evaluated on the detection of false negatives which may reduce the risk of missing tumour cells. Finally, the clinical prerequisites for developing and validating diagnostic DL systems for digital pathology are discussed, along with an overview of explainable AI techniques.

In conclusion, multiple approaches to facilitate the adoption of DL systems in clinical practice, addressing reliability, generalisability, and clinical needs aspects are discussed in this thesis. I believe that the extensive efforts in the research community will have a positive impact on the development, validation, and deployment of DL systems in digital pathology labs, empowering pathologists with trustworthy AI tools.

Machine learning (ML) algorithms are optimized for the distribution represented by the training data. For outlier data, they often deliver predictions with equal confidence, even though these should not be trusted. In order to deploy ML-based digital pathology solutions in clinical practice, effective methods for detecting anomalous data are crucial to avoid incorrect decisions in the outlier scenario. We propose a new unsupervised learning approach for anomaly detection in histopathology data based on generative adversarial networks (GANs). Compared to the existing GAN-based methods that have been used in medical imaging, the proposed approach improves significantly on performance for pathology data. Our results indicate that histopathology imagery is substantially more complex than the data targeted by the previous methods. This complexity requires not only a more advanced GAN architecture but also an appropriate anomaly metric to capture the quality of the reconstructed images.

Artificial intelligence (AI) has shown great promise for diagnostic imaging assessments. However, the application of AI to support medical diagnostics in clinical routine comes with many challenges. The algorithms should have high prediction accuracy but also be transparent, understandable and reliable. Thus, explainable artificial intelligence (XAI) is highly relevant for this domain. We present a survey on XAI within digital pathology, a medical imaging sub-discipline with particular characteristics and needs. The review includes several contributions. Firstly, we give a thorough overview of current XAI techniques of potential relevance for deep learning methods in pathology imaging, and categorise them from three different aspects. In doing so, we incorporate uncertainty estimation methods as an integral part of the XAI landscape. We also connect the technical methods to the specific prerequisites in digital pathology and present findings to guide future research efforts. The survey is intended for both technical researchers and medical professionals, one of the objectives being to establish a common ground for cross-disciplinary discussions.