Evidence for the Benefit of Targeted Proteomics in the Era of the "Big Data" Approach

Evidence-Based Diabetes Management, March 2016, Volume 22, Issue SP4

The availability of a novel predictive biomarker in diabetes care could have benefits for patients, payers, and pharma.

There is much discussion today concerning our ability to analyze and interpret large data sets with the goal of better understanding complex multigenic diseases and drug effects. These data sets are based on the careful col­lection of disease cohorts and controls, and include standard demographic and clinical information that can be analyzed with data from high-throughput biomarker platforms. Most such studies have been based on genetic analysis from a sin­gle nucleotide polymorphism or sequencing data and many include transcriptome data. Patients, pharma, and payers are all benefiting from this approach:

Patients benefit through early detection of disease when intervention is most successful and with “personaliza­tion” of treatments based on targeted therapies and pre­dictive diagnostic tests.

• Enormous data sets provide pharma the opportunity to better understand disease and, therefore, guide more ef­fective drug development, including optimal stratification of patients.

Payers benefit from the treatment and care efficiencies that result from these novel therapeutics and diagnostics. Big data studies will increasingly be a key component of the strategy to transition to population-based payments, or “volume to value.”

To date, the most strikingly successful application of big data approaches leading to dramatic improvements in health outcomes for patients is in oncology. Starting about 15 years ago, researchers identified candidate drug targets in specific tumor types in big data studies that relied on the use of high-throughput sequencing technology and large patient cohorts. This approach led to the targeted therapy drug revolution for personalized treatment in cancer, with dramatic improve­ments in quality of life and remission rates.

These therapeutic developments have been followed by advances in diagnostic tests, such as for KRAS and HER2/ NEU, which identify potential responders to targeted thera­py, and patients with increased cancer risk, as exemplified by the BRCA1 and BRCA2 tests for breast and ovarian cancer. We have now entered the era of gene panels replacing single gene tests for hereditary risk assessment in oncology risk and treatment. For example, Myriad Genetics’ myRisk Hereditary Cancer test is a 25-gene panel that identifies an elevated risk for 8 important cancers. However, the big data advances in other disease areas have not been as dramatic, with little progress being made in prevalent, chronic diseases, such as diabetes, cardiovascular disease, autoimmune disease, and neurodegenerative disease.


A recent issue of Circulation highlights a big data approach that includes high-throughput quantitation of blood-based serum protein biomarkers using a multiplexed immunoassay platform perfectly suited for big data studies.1 The research­ers hoped to identify biomarkers that could predict cardiovas­cular events in a population of 8401 carefully characterized individuals in the Outcome Reduction with Initial Glargine Intervention (ORIGIN) trial. To our knowledge, this is the most ambitious targeted proteomic study ever performed, with nearly 2 million protein measurements.

The 7-year trial, sponsored by Sanofi and managed by the Population Health Research Institute (PHRI), was designed to investigate the effects seen in a population of patients with type 2 diabetes who were on Sanofi’s long-acting in­sulin glargine, Lantus.2 Blood samples were drawn at 3 time points over the 7-year period, and baseline serum samples were subjected to targeted proteomic testing that generated a data set containing 2 million points. The generation of this large data set was made possible using a targeted proteomic platform, DiscoveryMAP®, developed and commercialized by Myriad RBM, a wholly owned subsidiary of Myriad Genetics.3 This testing platform provided quantitative concentrations on 237 individual protein targets in the serum using less than 1.0 mL of sample. The 237 targets cover many biochemical path­ways and mechanisms of therapeutic intervention. This study would simply not have been possible using conventional im­munoassay technologies due to cost and sample volume re­quirements to measure so many protein biomarker targets.

Using the protein measurements generated in this study, and a sophisticated data mining approach, Hertzel C. Ger­stein, MD, MSc, FRCPC, deputy director of PHRI and lead study investigator, correlated 10 specific blood proteins with a fu­ture cardiovascular event or death over the 7-year period of the study. The 10 markers, along with the predictive values for the clinical risk factors, identified individuals with dysglyce­mia who are at greater risk of a cardiovascular event. Some of the markers were already known to be associated with cardio­vascular disease, adding confidence to the quality of the study and approach. Others were more novel and require additional validation.

In addition to these 10 biomarkers, 5 other markers had the greatest impact to predict death over the 7 years in this pa­tient population. These are remarkable findings that provide a window into the future of these individuals. But from the perspectives of the patient, pharma, and payer, what will be the utility of such a biomarker test?


For patients, cardiovascular disease is still the number-one killer in the world, with 3 in every 10 deaths being attribut­able to the disease. Any improvement in risk prediction, pre­vention, and treatment can have a dramatic impact on saving patients from a premature death and improve their quality of life. By further improving our ability to stratify patients, pa­tient management can be more personalized and tailored to optimize prevention and treatment, thus reducing morbidity and mortality. Prevention measures for cardiovascular dis­ease, including lifestyle changes, are beneficial to all. Howev­er, changing behavior patterns is very challenging. Individuals are more likely to change behavior with direct feedback about their condition.4,5 This newly reported test for predicting car­diovascular events should be a powerful motivator for life­style changes. Beyond prevention, the other obvious benefit of such a test for patients is confidence in stratifying individu­als. Those at high risk would be eligible for more aggressive prevention and therapy programs, while lower-risk patients would be best maintained on less aggressive and less costly intervention and therapy.

For pharmaceutical companies, the novel serum multi-bio­marker test for cardiovascular events has several applications. These novel biomarkers open up avenues for understanding and exploring new drug targets and mechanisms of action. Preclinical and exploratory human research into the biomark­ers will further our understanding of their relationship to car­diovascular disease. The new test could also have significant implications for cardiovascular clinical trials. New drugs for cardiovascular disease and diabetes have to undergo large and lengthy clinical trials to monitor cardiovascular events and death. Use of the cardiovascular risk panel could help re­duce the size and complexity of such trials. Such a test could also potentially be used to correlate decreased cardiovascular risk with therapeutic intervention.

For the payer, the rapid formation and consolidation of population-based health systems are increasingly turning to big data analytics to mine ever more sophisticated electronic health records (EHRs). This provides the cost-effective man­agement of their patients as payers shift from fee-for-service to a capitated system of value-based payment. Reducing acute cardiovascular events is one of the most effective ways of reducing healthcare costs and improving a health system’s quality scores. Each year, a larger percentage of value-based reimbursements are tied to a provider’s patient population quality metrics. As structured, the additional bonus payments by payers to healthcare systems for improvements in patient quality scores is more than offset by the savings to the payer in reduced costs for healthcare services. Providers are incen­tivized to use data from EHRs to more directly and personally interact with their patients through innovations such as the patient-centered medical home.6,7

This personalized approach by providers can have profound improvements on prevention via additional incentives for pa­tients to alter their behavior and adhere to prescribed thera­pies. The new cardiovascular risk panel discovered by this big data study could be a powerful new addition to the EHR for managing a subpopulation of patients with dysglycemia who are at a higher risk for acute cardiovascular events and death.

The implementation of “big data” analyses should include not only molecular genetics approaches but also a targeted proteomics approach as documented by Gerstein et al. This study revealed a set of protein biomarkers in the blood whose concentrations could predict cardiovascular events and/or death within a 7-year period. Going forward, the impact of these types of studies to the patient, pharma, and payer will be to decrease morbidity and mortality in the higher risk pop­ulations, help develop more effective drugs, and, ultimately, save the healthcare system money through more efficient, tai­lored treatments.


1. Gerstein HC, Pare G, McQueen MJ, et al. Identifying novel biomarkers for cardiovascular events or death in people with dysglycemia. Circulation. 2015; 132(24):2297-2304. doi: 10.1161/CIRCULA­TIONAHA.115.015744.

2. Sanofi announces results of ORIGIN, the world’s longest and largest randomized clinical trial in pre- and early diabetes [press release]. Paris, France: PRNewswire, June 11, 2012. http://www. news.sanofi.us/2012-06-11-Sanofi-Announces-Results-of-ORIGIN-the-Worlds-Longest-and-Largest- Randomized-Clinical-Trial-in-Pre-and-Early-Diabetes.

3. DiscoveryMAP. Myriad RBM website. http://rbm.myriad.com/discoverymap/. Accessed February 12, 2016.

4. Chi CL, Street NW, Robinson JG, Crawford MA. Individualized patient-centered lifestyle recommen­dations: an expert system for communicating patient specific cardiovascular risk information and pri­oritizing lifestyle options. J Biomed Inform. 2012;45(6):1164:1174. doi: 10.1016/j.jbi.2012.07.011.

5. Lin JS, O’Connor EA, Evans CV, Senger CA, Rowland MG, Groom HC. Behavioral counseling to promote a healthy lifestyle for cardiovascular disease prevention in persons with cardiovascular risk factors: an updated systematic evidence review for the US Preventive Services Task Forces. National Center for Biotechnology Information website. http://www.ncbi.nlm.nih.gov/books/NBK241537/. Accessed February 12, 2016.

6. How can providers use health IT to create a medical home? Health Resources & Services Admin­istration website. http://www.hrsa.gov/healthit/toolbox/Childrenstoolbox/BuildingMedicalHome/ healthitmedicalhome.html. Accessed February 15, 2016.

7. Kern LM, Edwards A, Kaushal K. The patient-centered medical home, electronic health record, and quality of care. Ann Intern Med. 2014;160(11):741-749. doi: 10.7326/M13-1798.