How transformational technologies are shaping the future of cancer care
This article was initiated and funded by AstraZeneca
Transformational technologies are reshaping our understanding of tumor biology, uncovering novel biomarkers and aiding the development of innovative clinical trial designs, all of which have the potential to improve outcomes for cancer patients. In this article, Jorge Reis-Filho and Maurizio Scaltriti (both AstraZeneca, MD, USA) discuss the possibility of computational pathology, ctDNA and multi-omics in advancing a precision medicine approach to cancer care. They also delve into what is required to successfully integrate transformational tools into routine clinical practice.
Jorge Reis-Filho
Vice President, Cancer Biomarker Development, Oncology R&D
AstraZeneca (MD, USA)
Jorge Reis-Filho leads a team focused on the development of mechanistically-informed but clinically-deployable biomarkers for patient selection and therapeutic response prediction. Their goal is to ensure patients receive the right treatments or treatment combinations at the optimal points in their therapeutic journey.
With a background in molecular pathology, bioinformatics, functional genomics and AI, Jorge and his team have pioneered innovative methods to decipher cancer genomics, intra-tumor genetic heterogeneity and DNA repair defects in cancers and worked on developing next generation of AI-driven computational pathology biomarkers.
In his previous role as the Chief of Experimental Pathology and as a consultant for Goldman Sachs (NY, USA), Jorge leveraged his expertise in pathology, genomics and AI to bridge academic medical research with biotechnology and pharmaceutical industries.
Jorge is passionate about nurturing the next generation of scientists and promoting diversity within the scientific community. Throughout his career, he has collaborated with talented individuals from diverse backgrounds, united by a shared commitment to harnessing science for positive change and improving patients’ lives.
Maurizio Scaltriti
Vice President, Translational Medicine, Early Oncology, Oncology R&D
AstraZeneca (MD, USA)
Maurizio Scaltriti leads a team of translational scientists who work on projects in AstraZeneca’s tumor drivers and resistance, DNA damage response and epigenetics portfolio. The team provides the scientific rationale for the design of novel combinatorial strategies and identify biomarkers of response to therapy.
Before joining AstraZeneca in October 2020, Maurizio was an Associate Professor and the Associate Director of Translational Science at Memorial Sloan Kettering Cancer Center (NY, USA). In this role, he worked closely with early drug development physicians – enabling new findings and hypotheses from the clinic to be quickly evaluated in the laboratory setting.
Maurizio’s research interests have been focused on elucidating tumor vulnerabilities to improve patient selection and how the selective pressure imposed by targeted therapy impacts drug resistance.
Historically, what is the process of biomarker development for new cancer treatments? What are the limitations?
Jorge Reis-Filho: Traditionally, when developing cancer treatments such as chemotherapy and radiotherapy, a one-biomarker-per-drug approach has often been used for patient selection and to stratify patients into different groups. While this approach has led to the development of effective treatments, the complexity of cancer means that single biomarker tests are often insufficient to fully understand and target an individual’s cancer.
The molecular profile of the same cancer type varies between individuals and this can influence the response to therapy. We seek to deploy a holistic approach for the development of biomarkers, incorporating target and tumor biology aspects, as well as biologic characteristics of the patients. Our increasing awareness of how the genetic make-up, tumor mutational status, target expression and tumor microenvironment epigenomics vary across patients, has led to a shift in the focus of research and development away from general one-size-fits-all therapies toward precision medicine and personalized treatment combinations.
The unprecedented pace of technological advancement and the development of AI, together with its increasing adoption, are offering novel opportunities to utilize biomarkers to help guide the development of combinatorial treatments based on the mechanisms of action of not only the individual agents but also their synergistic effects in combination.
How is the landscape of cancer treatment changing? How are transformational technologies being leveraged to help improve cancer treatment?
Jorge: We are undergoing a fundamental paradigm shift in terms of our ability to collect and process data through AI approaches. In fact, novel classes of biomarkers, including computational pathology, radiomics and digital health, are rapidly emerging, and being utilized to understand cancer in a more holistic manner. The opportunities these approaches offer to identify new targets for novel cancer medicines, predict which molecules have the best chance of providing clinical benefit, and even define which patients have the best chance of benefiting from a single treatment or combination approach are truly unprecedented.
Now, we are at an inflection point where these technologies are maturing and beginning to have a positive impact for patients. AI is poised to change fundamentally the way we understand and integrate data.
Maurizio Scaltriti: These transformational technologies are the basis for applying translational science, which seeks to convert clinical and preclinical findings into actionable insights and tools that improve care. Translational science plays a key role in identifying which patients are likely to respond to treatment, understanding the factors that influence their response and deepening our understanding of the drug’s mechanism of action in the process. By studying the elimination of malignant cells during treatment and the changes in the tumor microenvironment upon pharmacological pressure, we can gain valuable insights into drugs’ mechanisms of action and how a patient’s unique molecular profile responds to treatment.
How is AstraZeneca utilizing transformational technologies to help advance a precision medicine approach in oncology?
Mauri: At AstraZeneca, translational science plays a pivotal role in developing precision medicines, acting as the link between early scientific breakthroughs, their application in clinical decision making and the development of next-generation diagnostics. This rapidly developing field is expanding our knowledge of tumor biology, pioneering state-of-the-art biomarkers and informing the design of innovative clinical trials, which have the potential to transform outcomes for patients.
Jorge: We’re applying a precision medicine approach to the majority of our early oncology pipeline. This is intimately linked to our transformational technologies:
Computational pathology. While traditional pathology approaches rely on a human’s assessment of tissue samples, advances in computer vision and AI are now allowing digital analyses of samples, which have the potential to transform how tissue-based biomarkers are developed. We are driving this approach in computational pathology with our novel Quantitative Continuous Scoring (QCS) platform. QCS is a fully supervised deep learning model for detecting and quantifying the presence and sub-cellular location of biomarkers to generate rich datasets, based on seven classes of human interpretable features to build a robust picture of what is happening inside a tumor, so we can select patients for clinical trials that are most likely to respond to treatment. We are currently pioneering the use of QCS in our clinical trial portfolio, aiming to fundamentally change the way we connect treatments to patients, allowing for more personalized treatments that are based on predictions with a much greater level of accuracy, precision and reproducibility.
ctDNA. Advances in sequencing now allow us to interpret circulating tumor DNA (ctDNA), which has the potential to transform the early detection of cancer, define which patients are most likely to have aggressive disease, monitor responses to specific treatments and characterize the genomic features and heterogeneity of cancer cells upon disease progression. This tells us about the presence and burden of cancer during the patient’s treatment journey and can potentially identify specific genetic mutations that help understand and predict mechanisms of response and resistance in a personalized manner. We have developed a systematic framework for the deployment of ctDNA assays that are fit for purpose and help define the optimal treatment for patients in different stages of their therapeutic journey.
Multi-omics. AI can integrate all elements of tumor biology including genomic, transcriptomic, proteomic and metabolomic datasets, allowing for the generation of multimodal biomarkers. Unlike traditional biomarkers, which assess a single variable, we are now leveraging foundation models to devise multiplexed and even multimodal biomarkers, which incorporate several sources of data to give a more sophisticated picture of an individual cancer. Combining data in this way also simplifies and streamlines clinical decision making, leading to the identification of effective biomarker solutions for a specific treatment combination.
Beyond therapeutic advancements, where do you see the greatest potential for transformational technologies in oncology?
Mauri: Transformational technologies have value not only for improving diagnosis and prognosis but also for clinical trial design and operation.
AI combined with digital health solutions plays a major role in the automation of processes, which will reduce the administrative burden on investigators, freeing up more time to drive innovation.
We are advancing our capabilities to improve data collection, accessibility and retention in clinical trials. Our digital solutions include diagnostics that can detect the presence and type of disease and biomarkers that remotely measure patients’ biologic and pathologic processes. Through these capabilities, we aim to improve the ability to gather clinical data in real time and better understand patients’ disease and their response to treatment.
Jorge: We see great potential for our computational pathology solution, QCS. We believe it has the potential to help us uncover new ways of selecting patients by deriving more precise, quantitative and continuous assessment of biomarkers on a single cell level. It aims to identify patients who are likely to be the most responsive to certain treatments, and identify new patient populations, for example, patients who express low levels of a biomarker that may not be detected using traditional pathologist scoring. We will continue to leverage novel advancements in computational pathology to deliver QCS models and biomarkers for not only the antibody drug conjugates in our portfolio, but also for other treatment modalities we are developing.
What is needed to successfully integrate transformational technologies into routine clinical practice?
Mauri: Collaboration across industry, healthcare systems and regulators is essential if these technologies are to realize their potential and become incorporated into regular clinical practice. We must also consider that the integration of technology into healthcare networks requires ethical and regulatory questions to be addressed in areas such as data privacy. The only way to overcome these challenges is through an open dialogue between industry and regulators.
As technologies are rolled out, awareness and education among clinicians and patients is vital so that tools are used effectively, and patients understand how their use could benefit their treatment. In all these discussions, patient centricity must be at the forefront; improving patient outcomes and experiences should be a primary goal.
Jorge: As the digital pathology revolution advances across the diagnostic ecosystem, it is inevitable that pathology labs will require updates in terms of laboratory and digital infrastructure, workflow, platforms and the training of their staff. We have the ability to quantify the expression of targets in subcellular compartments of cancer cells at single cell resolution, to apply AI-based computer vision approaches to digital images of millions of cells and to derive biological information not only in relation to the cancer cells themselves but also their spatial relationships in exquisite detail. To do that at scale, for all patients who may need it, the number of labs capable of digital pathology must also expand. Screening patients at scale is one of the unique opportunities that computational pathology offers and helps us move toward democratizing access to specialized pathology expertise and biomarkers.
How do you envision the evolution of these technologies over the next decade and what are the key challenges that must be overcome?
Mauri: Our transformational technologies have real potential to support disease understanding, patient diagnostics and finding the right treatment combinations, and we are working to understand and implement the ones that will make the biggest difference to patients. I believe it is a question of when, and not if, there will be an expansion of the use of adjuvant screening technologies and the integration of AI solutions into the process.
We are engaged in many clinical trials to put our transformational technologies to the test, even as patient enrollment, trial size and expense remain a challenge. Through these trials, we’re hoping to answer critical translational questions so we can maximize the real-world benefit.
Jorge: I agree that translational science powered by AI will offer novel and transformative opportunities to make discoveries and translate them into impactful solutions to improve cancer care for patients. I am certain that we will see greater integration of different technologies into a holistic, yet individualized approach to patient care, and I think this is where the real potential lies. The key to achieving success with this approach will be the development of assays and approaches that are clinically deployable as healthcare systems evolve.
This article was initiated and funded by
The opinions expressed in this interview are those of the author and do not necessarily reflect the views of Oncology Central or Taylor & Francis Group.