Digital Directions

The year 2024 marks The Pathologist’s 10th anniversary. To commemorate this milestone, we revisit some of the conversations and themes we hosted back in 2014. Bibiana Campos Seijo sits down with Garcia Rojo to discuss the advancements in the field of digital pathology. 

Our 2014 interview with García Rojo, The Digital Pathologist’s View, is available here. 

How has the field of digital pathology changed in the last decade? And how do you think it has affected your practice? 
 

Digital pathology has evolved considerably and continuously in these past 10 years, with better whole slide imaging (WSI) scanners, cost-effective storage solutions, interoperability, and more intuitive software, for example. The main impact of digital pathology in my clinical practice is related to increased efficiency, workflow optimization, better traceability, and decreased interobserver variability. Increased efficiency comes, for instance, from reducing the time that it takes to access previous studies or faster slide reading. Workflow optimization is mainly a consequence of redefining many processes in the lab, such as improvements to and standardization of microtomy, staining and coverslipping, or reducing the need to repeat slides. 

With digital pathology, it is easier to guarantee that the slides correspond to the correct patient, and we can make sure that the complete slide has been revised. Additionally, teleconsultation with the rest of the team in the department or with external colleagues is significantly easier and faster. 

When we interviewed you back in 2014, you talked about the need for validation and standardization in digital pathology. Has anything changed since then?
 

In validation, the main changes are related to the clinical guidelines released by scientific societies and digital pathology associations, and the introduction of In Vitro Diagnostic Regulation (IVDR) in the European Union – a framework governing the production and marketing of in vitro diagnostic medical devices with the goal of ensuring safety and effectiveness. 

In relation to standardization, we have Digital Imaging and Communications in Medicine (DICOM) WSI standards working and implemented by scanner and picture archiving and communication system (PACS) vendors. However, only a few hospitals worldwide are using DICOM to store their digital slides nowadays. The situation with regards to the “Integrating the Healthcare Enterprise” (IHE) guidelines, which define the standards we should use between different actors (pathology information system, scanners, viewer, storage repositories), is even worse. I’m not aware of any institution that has implemented the IHE Anatomic Pathology General Workflow Technical Framework, even though it has been available since 2010.

What are the main technological hurdles that remain for the implementation of digital pathology?
 

Today, technology cannot be considered a barrier for the digital transformation in pathology. 

The initial costs of digital pathology implementation are a main barrier. Making the case for digital pathology and convincing our managers of its cost efficiency and how it saves lives, is still difficult, but not impossible. 

Focusing on technology, apart from interoperability issues, we still need to optimize storage technology. WSI scanners should be able to accept slide racks from stainers. Scanning speed could also improve. Correct tissue detection or focusing is still an occasional issue with some scanners, and the generation of DICOM files natively from the scanners using patient and specimen data obtained from the pathology information system (scanners using DICOM modality worklist) is still under development. 

What are your views on the pace of integration of digital technology? In our 2014 interview, you acknowledged that scanners remained difficult to implement, for example. Is this still the case?
 

Many things have changed since 2014. Scanning technology is better today. Some private and public health institutions are aware of the importance of digital pathology to make precision and personalized medicine a reality. Nowadays, we see on average about 15 percent of hospitals with fully digital pathology in surgical pathology, and the figure is steadily increasing. In cytology, complete digitization is becoming a reality now thanks to new cytology scanners. 

Where do you see your lab in 10 years? “I think I’ll be fully digital,” you said in 2014. Has this prediction become true?
 

In my case, frankly, no. We suffered a delay due to the COVID-19 pandemic, and we expect to finish the digitization process this year. But we have many examples in Spain that have become fully digital in surgical pathology. First, it was Granada, with four hospitals working in an interconnected manner. Then, it was Catalonia, with eight hospitals, and the Quiron Hospital group successfully connected their hospitals in Madrid and Catalonia. Recently, complete digitization has become a reality in three northern regions in Spain – Castilla y León, Navarra, and the Basque Country. 

How do you see the field of digital pathology evolving over the next decade?
 

I expect that, in the next 10 years, digital pathology will be a requirement for pathology department accreditation – and most of these departments will be fully digital, including cytology. 

Part of the process will also be implementation of artificial intelligence (AI) algorithms. Otherwise, we run the risk of increasing the technology gap between pathology departments (and between patients).

I believe AI will become a driver for the adoption of digital pathology – especially for those who are reluctant to admit the benefits obtained from the digital transformation without AI. 

Are you hopeful about the future of digital pathology?
 

Ten years ago, we were not able to fully foresee the disruptive technology of artificial intelligence, and we were focusing on image analysis applied to digital slides that were difficult to develop with conventional technology. Deep learning has opened our eyes, and now we can think of the integration of AI algorithms that are able to detect automatically if a tissue slide contains cancer, to perform an efficient cervical cancer screening in cytology, to help in grading and classifying tumors, to predict the prognosis for each patient, or to estimate the probability of a tumor to express specific biomarkers or to detect nucleic acid mutations – all this with a digital haematoxylin-eosin. So, yes, I’m hopeful!