April 10, 2024

Embracing AI: The Third Revolution in Pathology

Even as pathology labs around the globe have begun to adopt digital pathology and artificial intelligence (AI), many pathologists remain skeptical. Will AI take their jobs? Can AI reliably support them in their challenging work? Can they trust AI in the clinic?

These concerns stem from the idea that AI is a breed of technology of its own, disruptive beyond what has been introduced in pathology before. Certainly, this is true in some ways; AI can understand data and uncover insight from that data better than existing tools can, unlocking a host of possibilities in the field. However, AI is ultimately another tool in the pathologists’ toolkit, much like any diagnostic adjunct used today.

Pathologists should view AI not as something to fear but instead as the next evolution of the practice. Like additional staining, genomic profiling, and other commonly used diagnostic tests, AI offers support for pathologists in efficiently and effectively understanding the tissue they are viewing and delivering a confident final diagnosis. Should pathologists embrace AI in the same way they have embraced the now trusted tools they rely on daily, the entire cancer care continuum will benefit.

Transformation in Pathology to Date

Pathology has so far undergone two major revolutions that enhanced the practice beyond what was previously thought possible. The first was the introduction of immunohistochemistry (IHC) staining in the 1980’s and 90’s, which empowered pathologists to visualize protein expression in tissue samples. Analysis of this information helped pathologists to not only deliver more precise diagnoses, such as a cancer subtype, but to understand key prognostic and predictive information that could inform more accurate treatments.

Then came genomic medicine, which provided further insight into the molecular basis of the disease. With so many treatment options available today that specifically target a cancer’s genomic dependencies, this information again empowered pathologists to provide more detailed diagnoses that could support the downstream care team in delivering personalized care.

While these examples may not seem revolutionary today, at the time they were introduced, they truly redefined the practice of cancer diagnosis. By expanding upon the information pathologists had available for interpretation and highlighting deeper individualized insight into the disease, both the first and second revolutions in pathology allowed physicians to make more informed, efficient, and confident decisions.

This is exactly what AI promises, and why it is posed as a third revolution.

AI: The Third Revolution

AI, at its core, is a tool for learning large amounts of data and recognizing patterns within it. This is especially useful in fields like pathology, because as pathologists know, cancer is extremely diverse and nuanced. While no single pathologist in their lifetime could see millions of slides containing millions of cancer variants, AI can. In this way, AI can further improve or streamline trusted diagnostic techniques and unlock never-before-seen functionalities and possibilities.

For example, AI can automatically identify tissue areas that are suspicious for cancer, even when they are small, challenging, or rare. It can also support cancer grading and quantification by reducing subjectivity and the tedium of associated tasks such as measurement or mitotic counting. Further, it can perform some of the tasks that would typically require an additional test or stain – and therefore time and cost – on H&E-stained slides alone, enabling the review of all relevant cellular, protein, and molecular information together at the point of diagnosis.

In sum, AI introduces efficiencies that allow pathologists to better focus their attention on the areas of diagnosis that most need it, offers an instant second opinion to enhance confidence, and can reduce total turnaround times or pathologist burnout. Yet like the revolutions that came before, it is only meant to supplement pathologist skill. The final diagnostic decisions still always remain in the hands of pathologists.

When AI is employed safely and responsibly, not only is day-to-day practice less tedious and challenging, but patients receive more confident and personalized diagnoses faster, which plays a role in improving their care or even outcomes. Pathologists today then are faced with a choice, not as to whether to employ AI, but when. With the magnitude of impact AI can make for practice and patients, the answer is clear – embracing AI today means pathologists can directly play a role in changing patient lives for the better and propelling the industry forward. The time to embrace AI as the third revolution is now.