January 25, 2023

Fulfilling the Promise of Precision Medicine: Why 2023 is the Year AI and Digital Pathology Make their Debut

By Andy Moye, PhD, CEO, Paige

I have been very fortunate in my career to watch the evolution of precision medicine in oncology from the front row. I worked with the legendary Dr. Daniel Von Hoff as he speculated that one could identify molecular and genomic targets for certain drugs for cancer patients, particularly for refractory and salvage patients, and apply those findings in clinical settings.1 I spoke with oncologists in the early days of molecular profiling who told me story after story where a Stage IV cancer patient with widespread metastases was almost miraculously cured after looking for molecular targets and prescribing something that was otherwise off-label. I am a huge believer in the power of precision medicine. Yet unfortunately, not enough patients are getting the benefit. Recently I attended the San Antonio Breast Cancer Symposium, and at the end of the plenary session with the finest minds in oncology, pathology, and precision medicine, a question was asked to the thousands of attendees – what is the biggest obstacle to overcome to find solutions for breast cancer? The answer from the audience? Slow implementation of precision medicine.2

The FDA defines precision medicine as “matching the right drugs or treatments to the right people, based on a genetic or molecular understanding of their disease. It’s an approach to patient care that’s based on the idea that one person’s disease isn’t necessarily exactly the same in someone else who seemingly has the same disease.”3 For twenty years, the full promise of precision medicine in oncology has eluded us, despite the overwhelming number of precision drugs to hit the market since the FDA approval in 2001 of Imatinib, which targets the Philadelphia chromosome mutation (Ph+) in Chronic Myeloid Leukemia (CML)4 Biomarker clinical trials in oncology now make up the majority of all trials, and in 2020, 79 total approved oncology drugs require or are recommended to be used after a biomarker test. 37 such drugs have been approved from 2016 to 2021, and every year biopharma discovers more and more new active substances requiring somatic biomarker testing.5

There is no doubt that these precision drugs and the biomarkers that define them have saved countless lives, but are we doing enough to ensure that all these amazing drugs are fulfilling the promise of the right drug, given to the right patient, at the right time? Not so much. Multiple practical issues plague both pathologists and oncologists in this area. One issue is that oncologists are unable to keep pace with all these new developments and tests and have shown that they are not following guidelines when it comes to timely ordering of even the most common biomarker tests. Labs also struggle to get timely reports and information back to oncologists. For example, one study of over 20,000 patients showed the rate of test results for common biomarker testing in Colorectal Cancer (CRC) was only 46% for NRAS, 56% for KRAS, 46% for BRAF, and 59% for MMR/MSI testing.6 Similar practice pattern studies for Non-Small Cell Lung Cancer (NSCLC) showed that only 46% of patients received results for all five biomarkers of EGFR, ALK, ROS1, BRAF, and PD-L1.7 Lastly, a recent study looking at NSCLC concluded that almost half of the time, precision medicines aren’t prescribed for patients due to factors associated with getting biomarker testing results.8 These are all biomarkers that are in guidelines and greatly impact the kind of care a patient would receive.

The reality is that not enough patients are getting tested appropriately. So, how do we improve patients getting tested at the right time, with the right test, so that oncologists can give the right drug? I believe the answer lies in digital pathology and AI. There is an old saying that no one has cancer until a pathologist says they do. Pathologists are the gatekeepers of the primary diagnosis as well as any follow-up testing. Today, AI can help pathologists find patterns in the Hematoxylin & Eosin (H&E)-stained slide and predict biomarkers at the same time a primary diagnosis is made. By adopting digital pathology and AI, pathologists can begin screening patients much earlier and more cost-effectively than comprehensive genomic profiling or other techniques, improving overall diagnostic yield and increasing the number of patients who can benefit from precision medicine. To that end, I believe we will see three key developments in 2023 that will greatly impact precision medicine:


Oncologists and pathologists will turn to digital pathology and AI to help screen for biomarkers or help find more patients who would benefit from precision medicines. 

With the approval of Trastuzumab deruxtecan, sold under the brand name Enhertu®*, for breast cancer and the realization that current methods for reporting HER2 expression were inadequate, oncologists and pathologists are seeking answers for how best to stratify patients as HER2 low. Some researchers have suggested as low as 60% concordance between pathologists on HER2 0 versus HER2 1 expressed patients.9 AI applications like HER2Complete™, which was given CE & UKCA marks in 2022 and trained on both mRNA expression and IHC status, may be the key to helping pathologists give the best answer on how to treat those patients. Another example is in lung cancer, where screening of clinical patients for something like EGFR with the use of AI could increase the number of patients who might benefit from precision therapeutics by 3x.10


The adoption of Digital Pathology and AI more broadly in the clinic will lead to earlier diagnosis and better, more cost-efficient outcomes.

Earlier diagnosis almost always leads to better outcomes. In Paige’s pivotal study leading to the FDA marketing authorization of Paige Prostate Detect, pathologists using AI made 70% fewer diagnostic errors.11 Paige is also working with the NHSx and Oxford University Hospitals NHS Foundation Trust on a two-year study of the use of Paige Prostate Detect for clinical diagnosis.12 Ultimately, the goal is to see improved outcomes for patients and a more cost-effective solution for payors. I believe many more studies will begin to shed light on the benefits of AI in both diagnosis and prediction across tumor types, and how the use of these tools leads to better outcomes for patients and payors.


More and more AI biomarkers will be used in clinical trials and studies to help stratify patients who are most likely to benefit from those targeted agents.

Paige works with Janssen Research & Development, LLC (Janssen) to deploy their AI-powered biomarker test in clinical trials to predict the presence of certain actionable alterations in the fibroblast growth factor receptor (FGFR) genes in patients with bladder cancer.13 My prediction is that more pharmaceutical companies who are developing these targeted agents will turn to digital pathology and AI to better enrich the cohort of patients who will benefit from their drugs in clinical trials. Over 55% of all oncology clinical trials today are biomarker driven – patients will benefit greatly from the use of AI to help get them the right drug in those trials.14


The Impact to Patients

In 2023, digital pathology and AI should play a much larger role in helping to diagnose patients, find the right patients for clinical trials, and ensure the right patient receives the right targeted agent.  AI can help empower pathologists by giving them information at the point of diagnosis that was never available before in an analog setting.  It can help reduce errors and enrich patient cohorts by looking at the tissue itself. At Paige, we have a vision of transforming cancer diagnostics through AI and truly helping patients defeat this terrible disease called cancer. We see a future where AI plays a huge role in fulfilling the promise of precision medicine, and I hope that we all move to adopt these technologies faster than ever – patients deserve it.



About the Author

Andy Moye has been at the forefront of the molecular diagnostics, biotechnology, and life sciences industries for over 18 years. During his tenure as CEO of Paige, he has driven the company’s mission of advancing patient care through the development of cutting-edge AI, including the U.S. launch of Paige Prostate Detect, the first and only AI-based pathology product to receive FDA marketing authorization for in vitro diagnostic (IVD) use in detecting cancer in prostate biopsies.



1Von Hoff DD, Stephenson JJ Jr, Rosen P, et al. Pilot study using molecular profiling of patients’ tumors to find potential targets and select treatments for their refractory cancers.  J Clin Oncol. 2010;28(33):4877-4883. doi:10.1200/JCO.2009.26.5983 

2Pancholi NJ. SABCS 2022: Panelists Discuss Strategies to Overcome Top Obstacles in Breast Cancer Management. American Association for Cancer Research. Published December 15, 2022. Accessed January 18, 2023.,their%20intracellular%20and%20extracellular%20environments 

3Understanding Precision Medicine. US Food & Drug Administration. Published 2019. Accessed January 18, 2023. 

4How Imatinib Transformed Leukemia Treatment and Cancer Research. National Cancer Institute. Updated April 11, 2018. Accessed January 18, 2023. 

5The IQVIA Institute for Human Data Science. Global Oncology Trends 2021: Outlook to 2025. IQVIA. Published June 2021. Accessed January 18, 2023. 

6Iyer P, Deng M, Handorf EA, et al. Assessing Oncologists’ Adoption of Biomarker Testing in Metastatic Colorectal Cancer Using Real-World Data. JNCI Cancer Spectrum. 2022;6(6). 

7Robert NJ, Espirito JL, Chen L, et al. Biomarker testing and tissue journey among patients with metastatic non-small cell lung cancer receiving first-line therapy in The US Oncology Network. Lung Cancer. 2022;166:197-204. doi:10.1016/j.lungcan.2022.03.004 

8Sadik H, Pritchard D, et al. Impact of Clinical Practice Gaps on the Implementation of Personalized Medicine in Advanced Non-Small-Cell Lung Cancer. JCO Precision Oncology. 2022;6. DOI: 10.1200/PO.22.00246  

9SABCS 2022: Researchers Share Insights on HER2-Low Breast Cancer. Breast Cancer Research Foundation. Published January 5, 2023. Accessed January 18, 2023. 

10Campanella G, Ho D, Häggström I, et al. H&E-based Computational Biomarker Enables Universal EGFR Screening for Lung Adenocarcinoma. arXiv. 2022. 

11Raciti P, Sue J, Retamero JA, et al. Clinical Validation of Artificial Intelligence-Augmented Pathology Diagnosis Demonstrates Significant Gains in Diagnostic Accuracy in Prostate Cancer Detection [published online ahead of print, 2022 Dec 20]. Arch Pathol Lab Med. 2022;10.5858/arpa.2022-0066-OA. doi:10.5858/arpa.2022-0066-OA 

12Nuffield Department of Surgical Sciences. The ARTICULATE PRO Project. University of Oxford. Accessed January 18, 2023. 

13Paige. Paige Announces Collaboration to Deploy a Novel AI-Based Biomarker Test for Advanced Bladder Cancer in Clinical Settings. Businesswire. Published June 15, 2022. Accessed January 18, 2023. 

14Vadas A, Bilodeau TJ, Oza C. Special Report: The Evolution of Biomarker Use in Clinical Trials for Cancer Treatments. The Journal of Precision Medicine. Accessed January 18, 2023. 

*Enhertu® is a trademark of AstraZeneca