Digital technologies and Artificial Intelligence are set to revolutionise the way biobanks operate. Pathology – i.e. the study and diagnosis of disease through the examination of surgically removed organs and tissues – is one of the specialties that are undergoing this process of digital transformation. Dr. Yervand Karapetyan (YK), expert pathologist at IBBL, explains the notion of digital pathology and its applications.
Interview by Federica Amato (FA).
FA: The concept of ‘digital pathology’ reflects the remarkable evolution of the traditional microscopy work. Can you explain what it entails?
YK: Indeed, when people think about pathologists and their work, they often picture us analysing a slide under the microscope. Although this is of course still the case, our activities have evolved considerably with the advent of modern digital techniques. In particular, the Whole Slide Imaging (WSI) technology has enabled the digitalisation of pathological and histological glass slides, which were traditionally analysed under a microscope. This means that glass slides can be scanned and stored as digital images, which can then be shown and analysed on a computer screen through a dedicated software. In terms of resolution, the digital images are equivalent or, in some cases, even better than those obtained through the traditional microscope. Basically, the conversion of the information contained in physical glass slides into computer-based information through WSI has enabled the digital analysis of tissues – and this application is called ‘digital pathology’.
FA: What are the main utilisations of digital pathology in clinical practice and biobanking?
YK: In clinical practice, digital pathology is essentially used for sharing digitalised slides with pathology experts located remotely. This allows clinicians to obtain a second advice on the diagnosis of a disease, without the need for the pathologist to be located in the same establishment, a concept known as telepathology. Similarly, digital pathology allows for the exchange of images and associated data between biobanks and research organisations and other partners, without having to send the physical slides for analysis. Moreover, specialised software based on computational algorithms enables the analysis of the image for research purposes.
FA: And how is IBBL making use of digital pathology to add value to its services and support high-quality research?
YK: At IBBL, all our routine slides are digitalised by means of the Hamamatsu ‘Nanozoomer’ scanner. These could be slides containing sections of colon, lung or skin tissue. We then provide different services to our various partners and clients, depending on the specific needs of their projects. For instance, through the Calopix software, we can mark the tumour area on the digital image and share it with our partners, so as to allow them to identify and dissect the actual portion of the tissue that presents cancer cells. Once excised, the tumour section can be analysed. This typically entails the extraction of DNA from the cancerous cells, its amplification through Polymerase Chain Reaction (PCR) and sequencing through Next Generation Sequencing (NGS). This way, our partners are able to perform a ‘molecular profiling’ and mutational analysis of the DNA. In other words, they can ‘read’ the sequence of the DNA and identify the mutations that have occurred and that have an oncogenic impact. This allows them to diagnose the tumour and develop targeted therapies.
Through the Calopix software, we can also provide quantitative data pertaining to the tumour. For some of our projects, we supply our partners with data such as the measurements of the size of the tumour, the characteristics of its surface and the percentage of surface area of tumour and non-tumour components. We can essentially annotate the image with all the information related to the cancer and share it with researchers anywhere in the world.
FA: It is clear that digital pathology is lifting many barriers in the clinical and research world. Can you sum up the main benefits associated with it?
YK: There are indeed a great many advantages and opportunities associated with digital pathology. First, WSI technologies allow the visualisation of the entire image of the slide on the screen, as opposed to viewing only portions of the tissue through a traditional microscope, thus giving a comprehensive overview of the whole tissue at a glance. Moreover, it is possible to zoom in and out through the thickness of the tissue to view the different components. Second, it has considerably sped up the entire process of collaboration and consultation among clinicians and pathologists. As opposed to sending the physical slides, writing a report and sending it back – which could take weeks – digital pathology has allowed the remote analysis of the image and its data in just 24 hours – a real improvement in consistency, efficiency and productivity! Third, digital slides can be stored virtually forever, without the risk of traditional staining being lost over time. This means they can be consulted and analysed over and over again by many scientists worldwide. This may also relax the current restrictions encountered when accessing tissue slides. In other words, digital pathology increases the availability of a large numbers of high quality, well-annotated samples for research purposes.
FA: Looking ahead, what does the future have in store for digital pathology, both in biobanking and in clinical practice?
YK: It is evident that the benefits and potential of digital pathology technology are significant, and the market for it is growing. As scanners and related software become more affordable, we will witness the complete digitalisation of slides and pathology workflows, moving towards fully digitalised pathology labs. In the clinical setting, digital pathology will also improve the efficiency of tumour boards, allowing the various specialists involved – radiologists, oncologists, surgeons, etc. – to meet ‘virtually’ and facilitating the preparation and presentation of cancer cases. In the biobanking world, WSI ‘virtual’ repositories will spread rapidly and will increasingly accompany the ‘physical’ biorepositories, thus redefining current biobanking practices. Finally, Artificial Intelligence and deep learning will lead to the development of software and algorithms specifically ‘trained’ to identify cancerous tissues and recognise the specific types of cancer cells… a full AI-based diagnosis in less than a minute!
FA: That sounds enthralling! Do you see any major challenges to the widespread implementation of digital pathology?
YK: First of all, before any real AI-based diagnostic use can be made, it will be necessary to scan and digitalise a significant number of slides. Indeed, the algorithms need to be ‘trained’ on a big enough data set if we want to be sure that the analysis they provide is truly accurate. Second, as with all disruptive new technologies, there is still a certain degree of reluctance to the implementation of digital pathology. This can stem from legal, privacy, security and confidentiality considerations, to practical issues such as the amount of bench space needed to accommodate the equipment. Third, since pathology images require very high resolution, their large size can generate problems in terms of storage and compatibility between the various computer systems in different labs. Fourth, some pathologists feel they lack the time to follow new trainings and to test the digital solution. Finally, there is always the question of costs, which are still considerable especially for labs with fewer financial resources. However, there is an increasingly positive attitude among pathologists and I have no doubt that, with further simplification of the workflow and better standardisation of technology, ‘digital’ pathology will become the ‘new’ pathology!
FA: Ultimately, any technological advance in the biomedical field is driven by the aim of benefiting patients. How is digital pathology going to revolutionise patient care?
YK: Digital pathology has greatly facilitated the work of pathologists in research labs. For instance, at IBBL, we carry out tissue analysis through digital pathology for research purposes. This means that we do not ‘diagnose’ the type and characteristics of cancer for immediate feedback to clinicians, but rather to share such information with our research partners, with the aim of advancing biomedical research and contributing to the development of new therapies. Conversely, in the clinical setting, tissue biopsies are digitally analysed in specialised pathology labs with the purpose of providing the clinician with a diagnosis, which will then assist him in the selection of the most effective treatment option for the specific patient. In this context, digital pathology has enabled pathologists to make more accurate diagnoses, more rapidly, thus improving patient care and health outcomes. In conclusion, the benefits of digital pathology for patients are twofold: better research and better treatment, a winning combination!
FA: Thank you very much Yervand for taking the time to answer my questions!
 A multidisciplinary group of doctors that meet routinely to discuss cancer cases and determine the best possible cancer treatment for individual patients.