Article
2025: The future of cancer diagnostics in the pathology lab
Q&A with Katie Robertson, Ph.D., oncology network lead at Roche Diagnostics
May 12, 2025
Since the founding of VENTANA® Medical Solutions (now Roche Tissue Diagnostics) 40 years ago, Roche has been at the forefront of driving much needed innovation for labs and patients. Looking forward, Katie Robertson, disease area network lead for oncology at Roche, shares her perspective on some of the exciting developments that are shaping the future of tissue diagnostics.
Where is the pathology laboratory headed in cancer detection?
There are several key trends and innovations we can expect over the next two to five years – AI integration with digital pathology and also personalized medicine evolving through companion diagnostics, both of which are influenced by successful development collaborations between organizations. And earlier detection of cancer is key for these trends.
What does the future hold for digital pathology in oncology?
There’s a shift from interpretation of traditional glass slides with the microscope to digital image analysis, and that’s accelerating.This has been greatly assisted by recent FDA clearances, including Roche’s, for digital pathology systems. Additionally, this transition will allow pathologists to analyze and interpret images with the help of software tools, potentially improving efficiency and accuracy.1
Integration of AI in digital pathology can assist in pattern recognition, improve scoring subjectivity,2 maximize patient identification, automate routine tasks3 and support diagnostic decision-making. AI image analysis may ultimately improve patient care by increasing accuracy of diagnosis, grading, staging and classification.4 We are just starting to see some of this happen now, and it will ramp up significantly over the next five years as more and more algorithms are developed and more pathologists become comfortable with the technology.
For the research space in digital pathology, we created the Roche open environment, which allows users to seamlessly access algorithms developed within Roche, as well as fully integrated third party tools – from longstanding companies to smaller startups that may only specialize in AI. Our broad menu of research-use-only algorithms give laboratories access to the latest algorithms.
Where will personalized medicine and companion diagnostics (CDx) be trending?
Right now, there are more than 60 FDA-approved companion diagnostic tests in hematology and oncology, and so many more are needed. One of the great things we see for patients are new indications for targeted therapies, and a recent example is HER2-low and HER2-ultralow. In the past what was classified as HER2 negative may now be classified as HER2-low or HER2-ultralow giving patients access to targeted therapy following endocrine therapy.
Roche, specifically has numerous companion diagnostics in the pipeline, including those to diagnose lung, gastric and prostate cancers. A Roche Kappa and Lambda test recently cleared by the FDA can potentially help treat patients faster, as it allows pathologists to accurately differentiate lymphoma from an infection, which is really important as the symptoms can be similar.
CDx and digital pathology will also begin to intersect with the future development of new CDx algorithms. There will also be some CDx that will only be used to score with digital pathology because the slide can’t be evaluated with the naked eye. We should expect to see improvement in biomarker development and CDx performance with AI-based technology.
Is there an upcoming new oncology advancement that will bridge the gap for a huge unmet medical need?
As of early 2025, there are no approved cancer therapies for NSCLC patients with the overexpression of c-MET protein. Among NSCLC tumors, 35% to 72%5 may overexpress c-MET. We’re working with two companies on two different diagnostic tests for c-MET protein overexpression that are targeted at two different patient populations.
Another big unmet need is gastric cancer diagnostics, especially with GI tumors, diagnosed much earlier now in younger adults. Of patients with gastric cancer, 61%6 already have advanced disease when they are diagnosed. We’re working on a companion diagnostic for FGFR2b, which is an emerging biomarker and expressed in 20% to 30%7 of gastric cancers. The five-year survival rate8 for metastatic gastric cancer is only 7%.

Further out in the pipeline, we’re looking at the PTEN biomarker in prostate cancer. What is interesting about PTEN is its loss or deficiency fuels the growth of cancer cells, leading to dysregulation of the PI3K/AKT pathway, and is associated with poor outcomes in patients with prostate cancer.
What other trends are happening in pathology oncology testing?
Two trends are converging at the same time. The first is that NCCN (National Comprehensive Cancer Network) is updating guidelines faster than in the past. It took NCCN just a couple of months to update guidelines to include CLDN18 testing for samples in which advanced/metastatic disease is documented or suspected. And at the same time, labs are also trying to gain greater insight into pipelines – looking two years out rather than the six to eight months as they were previously. This is true not only for companion diagnostics, but also for innovations in digital pathology.
References
- Drug Discovery and Development. Last accessed February 19, 2025.
- Cureus. Last accessed February 19, 2025.
- Archives of Pathology and Laboratory Medicine. Last accessed February 19, 2025.
- Modern Pathology. Last accessed, April 22, 2025.
- The Journal of Cancer Drug Discovery and Preclinical Development. Last accessed February 19, 2025.
- American Cancer Society. Last accessed February 19, 2025.
- The Lancet Oncology. Last accessed February 19, 2025.
- FGFR2B. Last accessed February 19, 2025.
Contributor
Katie Robertson, Ph.D., is a disease area network lead for oncology at Roche Diagnostics. She leads a team that provides disease area expertise and clinical input into medical strategies across pathology and sequencing for oncology.