This article has been commissioned and funded by AstraZeneca
Translational science plays a pivotal role in developing precision medicines. This rapidly developing field is expanding our knowledge of tumor biology and the mechanisms of action of therapeutic agents, pioneering state-of-the-art biomarkers, and informing the design of innovative clinical trials that have the potential to transform outcomes for patients.
Developments in artificial intelligence (Al) and computer vision are enabling a transformation in how biomarkers are discovered, developed, and, ultimately, delivered clinically. Computational pathology is becoming invaluable not only to replicate the tasks that pathologists do or to infer the presence of specific molecular alterations that can be identified by genomics, but also to understand mechanisms of action of targeted therapies and to identify patients who are most likely to benefit from specific treatments. In fact, computational pathology allows for a more holistic characterization of the tumor with precision, accuracy, and reproducibility, bringing together features to both target biology and tumor biology.
Recognizing the criticality of deeper biological understanding in precision medicine for oncology, AstraZeneca is committed to developing advanced computational pathology tools. We connected with Jorge Reis-Filho, Vice President of Cancer Biomarker Development in Oncology R&D, to learn how AstraZeneca is bridging the gap between early-stage drug development, biomarker identification, and clinical practice.
What priorities are driving AstraZeneca's approach to precision medicine in oncology?
Patients can have vastly different responses to treatment - despite sharing the same cancer diagnosis. These differences in response can be due in part to variations in their genetic make-up, epigenomic landscape, target expression, and more.
Our approach to precision medicine is based on understanding these unique variations between patients and developing treatments that could ultimately benefit patients at the right time during their treatment journey.
How is AstraZeneca using computational pathology to identify novel digital biomarkers for patient selection?
The vast majority of the therapeutic agents in our early oncology pipeline at AstraZeneca apply a precision medicine approach. This involves leveraging advancements in computer vision and Al to derive target and tumor biology insights. With this approach, we achieve levels of precision, reproducibility and accuracy that far surpass those attained by the human eye - while also increasing our understanding of tumor biology. These new biomarkers are transforming patient selection and enabling the delivery of more personalized treatments.
Our most advanced computational pathology biomarker solution is Quantitative Continuous Scoring (QCS) - a fully supervised Al-based platform that assesses the expression of targets in specific subcellular compartments of cancer cells. By applying QCS to digital whole-slide images of samples analyzed by immunohistochemistry (IHC), we can assess the number and distribution of biomarker-positive tumor cells. The technique allows us to calculate the staining intensity of biomarkers both on the surface of and inside every tumor cell.
QCS can enable enhanced decision making by pathologists - in a more precise, objective, reproducible manner than with traditional methods. It can also detect and quantify the presence of target expression heterogeneity in cancer cells within tumors. Based on the assessment of seven classes of human interpretable features, QCS helps develop new biomarkers that identify patients who are most likely to respond to certain treatments.
These QCS-based biomarkers also have the potential to identify new patient populations - such as patients who express low levels of a biomarker that may not be detected using traditional pathologist scoring.
How is multiomics being used to interrogate tumor biology at the molecular level?
At AstraZeneca, we are accelerating innovation to inform the development of next-generation diagnostics that have the potential to improve patient outcomes. We are developing multimodal biomarkers through Al-powered deep multiomic analyses - integrating molecular, genetic, transcriptomics, and single cell genomics data, with data derived from computational pathology and radiomics.
Unlike traditional biomarkers, which typically assess only a single variable, multimodal biomarkers provide a panoramic view of the disease, enabling an integrated approach to diagnosis and treatment. Our approach for the delivery of these multimodal biomarkers, however, is to reduce them to the most parsimonious set of biomarkers required for treatment decisions. This allows us to better define treatment strategies that are tailored to a patient's unique tumor profile, while keeping the assays clinically deployable.
How is AstraZeneca using circulating tumor DNA (ctDNA) technologies to enhance precision medicine approaches in cancer treatment?
There is no one-size-fits-all solution to cancer diagnosis and treatment. And that's why we have assembled a systematic framework for the technical benchmarking and deployment of ctDNA technologies. These technologies will allow clinicians to detect and intercept cancer earlier, track tumor dynamics, and monitor treatment responses. They also offer unique opportunities to identify patients whose tumors may be responsive or resistant to specific treatments in much greater molecular detail.
By detecting cancers earlier and being able to characterize their genomic and epigenomic features, we may be able to enable physicians to intervene sooner. In this way, they can offer personalized treatments that may be effective and durable, helping to transform the treatment journey of our patients.
How do you envision AstraZeneca's pursuit of transformational technologies shaping the future of cancer care?
Given the incredible development pace of new Al-driven technologies, it is clear that we are at an inflection point in oncology and pathology. Al applied to pathology is undoubtedly making an impact, but generative Al and foundation models will result in greater generalizability and accuracy of the biomarker solutions developed. Eventually, this will change the way we integrate data to understand patient trajectories even further - allowing for more personalized treatment based on predictions with a much greater level of accuracy.
A key challenge in our industry is navigating the multitude of therapeutic agents available to patients and how they can be combined or sequenced to maximize benefits. Traditionally, the pharmaceutical industry has focused on a one-biomarker-per-drug approach. At AstraZeneca, we are now using transformational technologies to develop the next generation of biomarkers for all classes of therapeutic agents in our pipeline. This advanced approach allows us to explore how to optimize combination therapies to attack cancer effectively from multiple angles.
We are pursuing exciting new, transformative technologies, including:
Foundation models - with the goal of developing a multi-cancer, cross-indication QCS, potentially allowing us to deliver QCS solutions across our portfolio for assets that target cell surface markers.
Multiplex biomarker solutions - potentially enabling us to quantify the expression of multiple targets and mechanisms of resistance on the same tissue section in a precise, reproducible, and quantitative manner.
How do you see these innovations transforming patient outcomes over the next few years?
AstraZeneca is leading the digital and computational pathology revolution - and we firmly believe that in addition to providing more precise, accurate and reproducible biomarkers, it will drive the democratization of access to biomarkers. This is the next step in transforming how precision oncology is practiced and enabling the delivery of our potentially life-saving drugs to patients at the right time in their treatment journey.
You can learn more about AstraZeneca's approach to computational pathology here.
This article has been commissioned and funded by AstraZeneca.
Newsletters
Receive the latest pathology news, personalities, education, and career development – weekly to your inbox.
