The high capacity of neural networks allows fitting models to data with high precision, but makes generalization to unseen data a challenge. If a domain shift exists, i.e. differences in image statistics between training and test data, care needs to be taken to ensure reliable deployment in real-world scenarios. In digital pathology, domain shift can be manifested in differences between whole-slide images, introduced by for example differences in acquisition pipeline -between medical centers or over time. In order to harness the great potential presented by deep learning in histopathology, and ensure consistent model behavior, we need a deeper understanding of domain shift and its consequences, such that a model's predictions on new data can be trusted.This work focuses on the internal representation learned by trained convolutional neural networks, and shows how this can be used to formulate a novel measure -the representation shift -for quantifying the magnitude of model-specific domain shift. We perform a study on domain shift in tumor classification of hematoxylin and eosin stained images, by considering different datasets, models, and techniques for preparing data in order to reduce the domain shift. The results show how the proposed measure has a high correlation with drop in performance when testing a model across a large number of different types of domain shifts, and how it improves on existing techniques for measuring data shift and uncertainty. The proposed measure can reveal how sensitive a model is to domain variations, and can be used to detect new data that a model will have problems generalizing to. We see techniques for measuring, understanding and overcoming the domain shift as a crucial step towards reliable use of deep learning in the future clinical pathology applications.

Recent technological advances have improved the whole slide imaging (WSI) scanner quality and reduced the cost of storage, thereby enabling the deployment of digital pathology for routine diagnostics. In this paper we present the experiences from two Swedish sites having deployed routine large-scale WSI for primary review. At Kalmar County Hospital, the digitization process started in 2006 to reduce the time spent at the microscope in order to improve the ergonomics. Since 2008, more than 500,000 glass slides have been scanned in the routine operations of Kalmar and the neighboring Linköping University Hospital. All glass slides are digitally scanned yet they are also physically delivered to the consulting pathologist who can choose to review the slides on screen, in the microscope, or both. The digital operations include regular remote case reporting by a few hospital pathologists, as well as around 150 cases per week where primary review is outsourced to a private clinic. To investigate how the pathologists choose to use the digital slides, a web-based questionnaire was designed and sent out to the pathologists in Kalmar and Linköping. The responses showed that almost all pathologists think that ergonomics have improved and that image quality was sufficient for most histopathologic diagnostic work. 38 ± 28% of the cases were diagnosed digitally, but the survey also revealed that the pathologists commonly switch back and forth between digital and conventional microscopy within the same case. The fact that two full-scale digital systems have been implemented and that a large portion of the primary reporting is voluntarily performed digitally shows that large-scale digitization is possible today.

Direct Volume Rendering has proved to be an effective visualization method for medical data sets and has reached wide-spread clinical use. The diagnostic exploration, in essence, corresponds to a tissue classification task, which is often complex and time-consuming. Moreover, a major problem is the lack of information on the uncertainty of the classification, which can have dramatic consequences for the diagnosis. In this paper this problem is addressed by proposing animation methods to convey uncertainty in the rendering. The foundation is a probabilistic Transfer Function model which allows for direct user interaction with the classification. The rendering is animated by sampling the probability domain over time, which results in varying appearance for uncertain regions. A particularly promising application of this technique is a "sensitivity lens" applied to focus regions in the data set. The methods have been evaluated by radiologists in a study simulating the clinical task of stenosis assessment, in which the animation technique is shown to outperform traditional rendering in terms of assessment accuracy.