Digital Pathology in the Clinic: Training, Validation and Patient Safety

Digital pathology is a technology with the potential to transform the way in which histopathological diagnoses are made and cancer diagnostic services are delivered. Despite this, clinical deployment of digital slides has lagged behind research and educational uses. This thesis describes some of the key barriers to widespread clinical adoption, which largely relate to a lack of guidance and information for pathologists regarding validation, training and patient safety. The evidence base for patient safety was analysed in a novel way to provide the basis for a validation and training protocol which was trialled in real world clinical settings, and guidance documents were developed and disseminated to the clinical pathology community to help with the transition from glass slide to digital slide reporting. In Chapter 1, background information and an overview of the published literature regarding clinical use of digital pathology is provided. In the second chapter, the results of a national survey on access to and usage of digital pathology hardware, in addition to attitudes to digital pathology, is presented. One significant barrier preventing digital pathology adoption has been a lack of widespread acceptance of digital slides as a safe alternative to conventional glass slides. Historically, validation literature investigating the safety of digital pathology as an alternative to conventional light microscopy has focussed on concordance metrics of glass and digital diagnoses, when arguably, it is appreciation of discordant cases that provides the clinical pathologist with the best opportunity to evaluate the scope of safe digital practice in their specialty. Chapter 3 describes a novel study to analyse diagnostic accuracy of whole slide imaging and identify key training and educational targets for novice digital pathologists. Chapter 4 presents the validation and training protocol developed by the author for Leeds Teaching Hospitals NHS Trust, which was subsequently adopted by the Royal College of Pathologists as an example of best practise in digital pathology implementation.1 Chapter 5 describes the deployment of this protocol to train and validate the primary digital diagnosis of cohorts of breast and neuro- pathologists. Chapter 6 introduces modifications of the protocol for use for more niche reporting scenarios: frozen section diagnosis and immunohistochemistry assessment. Chapter 7 responds to concerns in the pathology community regarding accreditation of digital services, and the use of WSI for primary assessment of screening programme specimens. The body of work presented in this thesis has generated multiple peer reviewed publications which have influenced national and international digital pathology guidance. In this time period, enormous progress has been made in converting digital pathology from a niche technology for the early adopter to a mainstream topic at clinical digital pathology conferences, and the number of deployments and planned deployments in the National Health Service and beyond has risen dramatically. The use of digital slides in routine clinical practice represents a major departure from conventional light microscopy working practices, and the author hopes this work will help the pathology community maintain diagnostic quality in a time of change