At the 39th annual meeting of the Society for Immunotherapy of Cancer (SITC), Flatiron’s Dr. Kristi M. Zimmerman Savill, Director of Scientific Engagement, Precision Oncology, discussed the important role that omnics-based technology, testing, and linked real-world data will play in ushering in a new era of precision immune-oncology research.
Since the FDA first began approving molecular targeting therapies — starting in 1977 with the approval of the estrogen receptor modulator tamoxifen — cancer research and oncology practice have significantly evolved. However, while patient outcomes have improved, it is becoming increasingly clear that treating cancer with a one-size-fits-all approach based largely on the tumor’s organ of origin, histology and disease stage does not lead to optimal outcomes for many patients. Today, research and practice are shifting toward precision oncology.
Understanding the potential of precision oncology
As Dr. Zimmerman Savill explained, precision oncology allows a patient’s anti-cancer therapy to be customized based on the molecular characteristics of their disease with the ultimate goal of targeting the right drug, at the right dose, to the right patient, at the right time. Already, precision oncology is leading to improved patient outcomes — and advances in molecular profiling have unlocked more and more precision oncology-based approaches. To date, over 40 targeted therapies acting against specific molecular pathways have been approved by the FDA.
Yet Dr. Zimmerman Savill raised the important point: while progress has been made, in order to unlock precision oncology’s full potential, more research is needed to uncover new approaches, and biological markers that are predictive of response are still needed in order to further improve outcomes for more patients.
The existing limitations, challenges and opportunities within precision oncology research
Research must address the critical challenges and opportunities in the precision oncology space. For instance, Dr. Zimmerman Savill noted that critical questions remain about the complex and dynamic processes driving tumor progression and treatment resistance among broad, diverse patient populations.
This is where real-world data (RWD) can play a crucial role in advancing such research, filling in gaps in data from populations typically underrepresented in clinical trials. Dr. Zimmerman Savill noted several key examples of traditional trial data limitations, including:
- ~7% or less of adult patients with cancer participate in cancer treatment trials in the US
- Black and Latinx patients are less likely to participate in trials than White patients
Fewer patients in economically or socially marginalized neighborhoods participate in trials - Barriers to trial participation can include those related to narrow trial I&E criteria, trial access, awareness, perceptions, financial, language, etc.
High quality and comprehensive RWD can address such limitations and lead to the optimization and development of novel precision oncology approaches. However, existing multi-modal datasets often still come with limitations, including data quality issues, lack of complete longitudinal clinical data, limitations in the depth and breadth of molecular data, data being sourced mostly from academic medical centers, small cohorts once all inclusion criteria are applied, and other factors. These challenges have limited researchers’ ability to address critical research questions and to integrate this type of data into drug development decision-making.
Introducing the Clinical Molecular Database (CMBD).
Dr. Zimmerman Savill described how Flatiron Health and Caris Life Sciences are partnering to address many of the limitations and gaps in existing clinical-molecular datasets with the creation of the Clinical Molecular Database (CMBD).
Dr. Zimmerman Savill discussed what distinguishes the CMDB from other linked datasets. Importantly, Flatiron owns the oncology electronic health record software that has transformed oncology care, from which most of their research-grade data is pulled from. This allows traceability and ensures consistency and auditability. This also allows Flatiron to process the full EHR end-to-end and make data usable for research purposes. Additionally, this increases data quality and allows for capture of the longitudinal patient journey with impressive completeness.
Additional factors that set Flatiron and Caris apart include:
How researchers can utilize the CMBD
Dr. Zimmerman Savill covered potential use cases for the CMDB throughout the research and development life cycle stages — from discovery and translational research, through clinical development and beyond. Such powerful use cases from a non-exhaustive list include:
- Uncovering cohorts with unmet need to inform drug development planning.
- Identifying and profiling novel targets and the biological and clinical correlates with alterations in or increased expression of those targets.
- Informing a target product profile and help to make go/no go decisions on advancing a program or product
- Uncovering mechanisms of treatment resistance to inform combination strategies, biomarker development, etc.
- Guiding clinical trial design (including patient selection and stratification approaches)
- Informing label expansions to bring therapies to patients likely to benefit
Dr. Zimmerman Savill also explained that the CMDB can be leveraged to make better informed decisions more quickly, including around program prioritization. Hypotheses generated from preclinical studies and clinical trials can be validated in the broader real-world population using the CMDB. Furthermore, insights generated from the CDMB can be validated and explored in the lab via reverse translation.
Importantly, the breadth, depth, and scale of the CMDB unlocks use cases that could not be sufficiently addressed with clinical trial datasets or other sources. For example, the CMBD may unlock opportunities to identify very rare biomarker-defined subtypes of patients via unsupervised clustering analyses and characterization in terms of biology and clinical endpoints.
Compared with other multi-omics datasets, the CMDB provides more in-depth and longitudinal clinical data critical for addressing many research objectives, as demonstrated in this example:
Dr. Zimmerman Savill also shared how the CMBD can improve researchers’ ability to evaluate tumor biology, sharing a case study that examined outcomes after first-line treatment with immunotherapy plus chemo in advanced non-small cell lung cancer. The study successfully demonstrated that the CMBD can be used to not only replicate known associations, but also generate new insights. Additionally, the study underscored that the CMBD can:
- Enable fast data-driven decision making (e.g., target/program prioritization)
- Fuel target discovery
- Guide combination therapy approaches
- Identify predictive biomarkers
- Inform clinical development strategy (e.g., patient selection/I&E criteria, stratification)
Key takeaways
To conclude, Dr. Zimmerman Savill summarized the key benefits and potential impact that the CMDB can have on the future of precision oncology and cancer treatment. The CMDB will allow researchers to build on recent advances in molecular profiling and serve to further progress oncology research and treatments, specifically fueling more precision oncology-based approaches. The CMDB will continue to address unanswered questions and limitations of other multi-model datasets, leading to the optimization and development of novel therapies, and the ushering in of a new era in precision oncology research and development.
