Long Live the (Digital) Revolution!

Pathology has witnessed many changes over the last few decades– it seems like we’re applying new technologies and developing new molecular studies every day. Digital pathology is just one such example: a dynamic, image-based technique that enables the acquisition, management, and interpretation of diagnostic information obtained from a digitized glass slide. Healthcare applications for digital pathology are incredibly wide-ranging: primary and intraoperative diagnosis, diagnostic consultation, medical student and resident training, manual and semi-quantitative review of immunohistochemistry (IHC), clinical research, diagnostic decision support, peer review, tumor boards, and more.

In the last 10 years, digital pathology has rapidly expanded as an essential tool to support medical education, tissue-based research, drug development, and the practice of clinical pathology. Recently, the field has begun integrating artificial intelligence (AI) and machine learning software as tools for efficiency and accuracy within pathology.

Applying AI and machine/deep learning to high-resolution digital tissue images should ultimately allow pathologists to automate the analysis of routine clinical slides. And the pathologist may use software to create algorithms to spot cancerous areas – or even identify individual cancer cells – within such a digitized tissue sample. How? Such software uses annotated areas to memorize and create criteria that can help pathologists narrow down morphologic diagnosis and interpretation. In fact, we already use similar software algorithms to evaluate percentage expression of IHC staining – a function that has proven especially useful for evaluating ER, PR, and Ki67 in breast pathology services. Software like this helps to standardize data and improve patient care.

In my view, the integration of AI and machine learning software into the daily clinical workflow of pathologists will cut down work time, reduce misdiagnoses, and create more comfortable work environments. At the same time, I anticipate that it will make pathology departments more efficient and cost-effective. In our digital pathology laboratory, we use our system to cover intraoperative frozen section evaluation, liver and kidney transplants, tumor boards, and clinical education for two hospitals. Our data on the application of digital pathology to evaluating frozen sections has garnered great interest at national and international meetings and, in the near future, we plan to expand our laboratory so that we can digitize all slides for daily base morphologic interpretation.

In short, I believe that new technology is set to help pathologists get out of the basement and take up the challenge of our increasingly important role in personalized health care.