Integrating genomics into population-based cancer surveillance in the era of precision medicine

Population-based cancer surveillance provides a quantitative measurement of cancer occurrence in the United States and globally. Core activities of surveillance include measuring cancer incidence and characterizing each cancer with regard to histopathology, stage, and treatment in the context of survival. Cancer surveillance has been crucial in informing policy and practice, as well as clinical and public health efforts to reduce the cancer burden. Surveillance also provides information for generating research hypotheses on cancer causes and outcomes, and for developing and evaluating interventions for cancer prevention and treatment.

Cancer surveillance traditionally is conducted based on tumor anatomic location, histologic features, size, involvement of lymph nodes and distant metastasis (i.e., anatomic stage) at diagnosis. However, within and across cancers, biomarkers can identify heterogeneous subgroups associated with different risk factors, treatment responses, recurrences and survival patterns. Population-based cancer registries have already been integrating these important predictive and prognostic factors into data collections. Some of these go into anatomic stage determination (e.g. Gleason score and prostate specific antigen) others such as gene expression profiling for breast cancer are used to predict response to therapy and guide treatment decisions. However, there are many new biomarkers that have been incorporated into the AJCC 8^th Edition as a component of prognostic stage and will be collected starting in January 2018, if they are not already part of routine surveillance (such as breast cancer gene expression profiling and human papilloma virus). The use of these biomarkers in refining surveillance data is increasingly important. For example, the 2015 annual report showed breast cancer incidence by molecular subtypes using tumor biomarkers for hormone receptor (HR) and human growth factor-neu receptor (HER2) expression. The report showed that HR+/HER2- breast cancers, the subtype with the best prognosis, were the most common with highest rates among non-Hispanic white women, local stage cases, and low poverty areas. HR+/HER2- breast cancer incidence rates were strongly correlated with mammography use, especially among non-Hispanic white women. Triple-negative breast cancers, the subtype with the worst prognosis, were highest among non-Hispanic black women.

The surveillance community must continue to evolve in their characterization of cancers according to biomarkers for subtype classifications. With rapidly evolving clinical applications in sequencing of the human genome as well as the genomes of tumors, the traditional anatomic descriptions of cancer types will be supplemented by molecular classification based on tumor genetic aberrations. Other tumor-related genome markers are rapidly maturing providing prognostic indicators for survival and response to therapy (e.g. gene expression profiling in prostate cancer). In addition, conducting surveillance for inherited causes of cancer, which account for about 5-10% of all cancers, will allow us to stratify reporting and tracking of cancers by underlying genetic causes. Examples include inherited mutations in BRCA1/2 in breast and ovarian cancer, and mutations in mismatch repair genes (Lynch syndrome) in colorectal and endometrial cancer.

Patients with inherited cancers can also respond differently to different treaments and may have different outcomes. (see one example here)  Such integration will allow monitoring of incidence, response to treatments and survivorship, evaluating trends and uncovering gaps in interventions across subgroups of heterogeneous cancer types and subgroups of the population based on age, race/ethnicity, geographic locations and other factors. The NCI’s Surveillance, Epidemiology, and End Results (SEER) cancer surveillance program is currently supporting an important pilot in which BRCA mutation panels have been linked to breast and ovarian cancer cases in California and Georgia. This linkage represents the first population-based set of information on testing of women with breast and ovarian cancer. This type of linkage exemplifies the new approaches that can improve cancer surveillance in today’s rapidly evolving oncology practice.

The launch of the U.S. Precision Medicine initiative in 2015 includes a large cancer component, featuring acceleration of NCI precision medicine trials. These efforts, will, over time, lead to molecularly targeted interventions that maximize benefits and minimize harms and costs of interventions. A conceptual shift toward precision medicine has the potential to change diagnostic categories, treatment strategies, and enable early detection and prevention. The approach has led to some new treatments that are individually tailored based on genomics (read recent review for example), but much more lies ahead, especially in using evidence- based recommendations for use of genomic markers as prognostic and predictive factors.

The long-term implementation and success of cancer genomics in improving cancer care and survival will depend on our ability to track the population impact of new genome-based discoveries in the “real” world and not only in the context of controlled clinical trials. As a deluge of information is generated in the next decade, it will become increasingly crucial to integrate evidence-based genomic information relevant to cancer etiology, response to interventions, and long-term outcomes in cancer surveillance. While much of cancer genomics is still in research mode, several genomic applications are used today in clinical oncology and cancer prevention. The CDC Public Health Genomics Knowledge Base (PHGKB) displays an updated searchable list of genomic tests and applications by their level of evidence based on systematic reviews and guidelines. Tier 1 applications feature tests with highest level of evidence. As of July 27, 2017, PHGKB lists 47 tier 1 genomics applications, most of which (35/47 or 74% are cancer-related).

As the field of cancer genomics matures further, there will be many scientific and logistical challenges to integrate new findings into surveillance and surveillance research. These include, among others, the rapid pace and volume of biomarker discovery and applications; the need for standardized definition of biomarkers; and capture of testing, especially when specimens are sent to reference labs and/or results may not be sent to the settings where cancer registrars collect surveillance data.

In addition to the ongoing challenges of surveillance for cancers related to traditional risk factors such as smoking and obesity, there will be an increasing need for a scientific-based approach to choose relevant genomic applications for cancer surveillance as well as the need to conduct pilot demonstration projects. In addition, the inclusion of these clinically relevant data will require the surveillance community to think in novel ways and work with nontraditional partners such as genomic labs to include these data on a population level. In the long run, cancer genomics, along with other scientific fields such as informatics, geographic analysis, and data linkages, will contribute to a new era of cancer precision surveillance.

As always, we welcome our readers’ input and feedback.