Since the inception of the Human Genome Project in the late 1980s there has been unprecedented growth in the technologies to assay nucleic acids, proteins, and metabolic analytes at scale. One might have expected a concomitant increase in the development and delivery for use in the clinic of genome-inspired biomarkers and diagnostics; but, as this special cardiovascular issue of Clinical Chemistry demonstrates, the field continues to be denominated by old friends such as cardiac troponin I (cTnI),2 B-type natriuretic (BNP), and other single-analyte markers, albeit with greater analytical sensitivity and, as a result, broader potential applicability. It is disappointing not to see more from genomics given that the field of cardiovascular medicine aspires to be more precise in risk assessment, early and accurate diagnosis, and in the treatment and monitoring of disease. Most cardiovascular biomarker and diagnostics have been introduced using disease hypotheses based on known biology. BNP and cTnI are great examples that have revolutionized our ability to distinguish cardiac from pulmonary etiologies of dyspnea, to risk stratify patients with heart failure, and guide therapies in acute coronary syndromes. Unbiased and hypothesis-generating approaches using genomics, transcriptomics, proteomics, and metabolomics have great potential to deliver novel biomarkers and assays for the practice of cardiovascular precision medicine, but have challenges in their path to becoming accepted and useful clinically.
Prior to the advent of the first genome-wide association studies (GWAS) in 2005 (1, 2), the literature was littered with association studies using single nucleotide polymorphisms that failed to replicate. The reasons for this were many, including poor study design, insufficient power, the use of “samples of convenience” resulting in confounding by unmeasured variables, and the lack of replication as a requirement for publication. Over the past decade, GWAS have defined the association of genetic variants with a variety of cardiovascular phenotypes, including myocardial infarction, coronary artery disease, dysrhythmias, aortic aneurism, and dyslipoprotieinemias. These GWAS could be conducted, in part, because of the assembly of large consortia (e.g., CARDIoGRAMplusCDR [Coronary ARtery DIsease Genome wide Replication and Meta-analysis (CARDIoGRAM) plus The Coronary Artery Disease (C4D) Genetics], CHARGE [Cohorts for Heart and Aging Research in Genomic Epidemiology]) with DNA collected and similar phenotypic data. These cohorts not only provided a standardized approach to consent, sample collection, and clinical data acquisition but also unsurpassed sample size and opportunities for replication of results. The results of GWAS studies during the past 11 years have generated unparalleled insights into the biology of disease with an unbiased reveal of >4000 variants associated with >300 clinically relevant phenotypes. Arguably this is one of the more stunning scientific accomplishments of the past decade—giving us a window into disease mechanisms that were heretofore never contemplated to be associated with those diseases. From a clinical perspective, however, since the odds ratio of the risk allele was <2 in most cases, these studies yielded disappointingly little that could be used to assess risk and/or that was superior to conventional risk models. Where DNA diagnostics have made progress is in the area of Mendelian diseases such as hypertrophic cardiomyopathies, and arrhythmias such as long QT syndromes. Panels of variants that have been reproducibly found in families with these phenotypes are now available commercially and used in highly selected cases. But outside of these rare diseases, few DNA-based genetic tests have made it to routine clinical use.
One area of genetics with perhaps the greatest potential in the near term for clinical use is the use of gene variants to guide the use of drugs. While few pharmacogenetic tests have been adopted into clinical guidelines, they are becoming increasingly used in clinical trials and in patient care. Genetic variants guiding the use of warfarin, clopidogrel, and statins are reaching the evidentiary threshold for clinical use. Some believe it is already there but others are looking for more studies with demonstrable impact on outcome. The Clinical Pharmacogenomics Implementation Consortium has published guidance on these drug–gene pairs and it behooves the cardiovascular and primary care communities who are the most frequent prescribers of these medications to be familiar with this guidance.
Perhaps surprisingly, RNA- and protein-based tests have had more success in their development, commercialization, and approval for reimbursement. The development of the Allomap® test for monitoring patients for rejection following cardiac transplantation is a notable example of what is possible and what is required for development of these types of assays for diagnostics. The test was developed using genomics and bioinformatics technologies. DNA microarrays were used to discover 252 RNA transcripts in blood samples associated with rejection defined by histology from endomyocardial biopsies. Quantitative real-time PCR technology confirmed 68 of the transcripts from which a 20-gene gene expression panel was selected. Two clinical studies were key to its development and US Food and Drug Administration approval: the development and clinical validation of the test used patient samples and clinical data obtained during the Cardiac Allograft Rejection Gene Expression Observational Study (3). From 2001 to 2005, 737 patients from 9 US transplant centers enrolled in the study and contributed 5834 blood samples and associated clinical data. A comparative effectiveness study, the Invasive Monitoring Attenuation through Gene Expression Study, compared clinical outcomes of patients managed with AlloMap to outcomes of patients managed with endomyocardial biopsy (4). The study, which ran from 2005 to 2009, included 602 patients from 13 US centers who were at least 6 months posttransplant. The results showed that AlloMap was not inferior to endomyocardial biopsy with respect to clinical outcomes when used to monitor stable, asymptomatic heart transplant patients. Another blood-based RNA test, Corus CAD®, has also been commercialized. Here again, DNA microarrays were used to profile blood RNA and the results were correlated with the degree of obstructive coronary artery disease by coronary angiography, the gold standard. Several clinical studies—PREDICT (patients with renal impairment and diabetes undergoing computed tomography) (5) and IMPACT (Improving Pediatric and Adult Congenital Treatment) (6)—have been used to establish the clinical utility of Corus CAD.
The development of both of these complex RNA-based genomic tests highlights the rigorous process required to bring a test from genome to patient (Figure 1). As with the development of therapeutics, the development of diagnostics is a process with many stage gates and opportunities for attrition. Few genomically derived cardiovascular biomarkers have been put through the entirety of this process. Most investigators stop at the validation stage, which is often associated with a journal publication. Taking the process forward to implementation and clinical studies is a substantial investment that most researchers are unable to carry out without a commercial partner.
Discovery and development of genomic biomarkers from discovery to application.
Challenges in the development of genomics-derived cardiovascular biomarkers and diagnostics remain. Below are 6 areas the cardiovascular biomarker community needs to reckon with so that genomics can have the greatest impact on the field.
Patient engagement and proper consent. Patients and the public need to be engaged as partners in the research process, so as to mitigate concerns surrounding privacy, trust, and informed consent, and ensure an ongoing research partnership. Standardized informed consent should be adopted that includes the use of a broad range of genomic assays and that enable future studies.
Standards of phenotypes and longitudinal studies linked to biorepositories. Development of a standard nomenclature and measures both within the electronic medicine record (EMR) and also for data acquired outside of the EMR is essential for harmonization of studies and for future validation. As illustrated by the NIH eMERGE (Electronic Medical Records and Genomics) network, EMRs linked to biorepositories becomes a key strategy for large-scale discovery and validation studies.
Data ownership, privacy, and sharing. Barriers to sharing data—within healthcare industries and systems, and the patient population—as well as concerns surrounding data ownership and privacy hamper the acceleration of the science and rapid replication and validation of biomarker results.
Data analytics and data science. The assembly of genomic data with various types of phenotype data from EMRs, patient reports, and mobile health requires rigorous standards and the development of architectures that will allow for data housing and access. Novel methods—e.g., machine learning and deep learning—are required as is a workforce of data scientists trained in them in order to realize the full potential of these data.
Implementation. Enhanced investment is needed in information technology infrastructure, data standards and interoperability, decision support technology, and implementation science as part of a translational health research continuum. NIH has historically been focused on discovery research (7, 8); however, with the Dissemination and Implementation initiative and the National Human Genome Research Institute's formation of the IGNITE (Implementing Genomics in Practice) network, that appears to be changing.
Evidence generation. Research and studies are needed that demonstrate that genomic and precision medicine innovations will improve patient outcomes, increase cost-effectiveness, and reduce the need for unnecessary testing and therapies. Genomic technologies are delivering ever-increasing numbers of molecular tools for clinical decision-making. The pace of discovery far outweighs the ability to evaluate the value of these using traditional randomized controlled clinical trials. Alternative strategies of evidence generation, perhaps integrated with healthcare delivery, should be considered (9). Policy frameworks should address when a randomized control trial is needed vs other types of evidence (e.g., observational via registries) for clinical adoption, regulatory approval, and reimbursement by payers.
The future is bright for the advent of precision cardiovascular medicine using a novel pallet of biomarkers that are being mined from genomic data linked to robust clinical phenotypes. The cardiovascular community must address certain key challenges to truly enable discovery through clinical application. This will require partnerships between the academic community, industry, payers, and regulators. The cardiovascular community needs to come together in order for it to be scaled. The next Clinical Chemistry cardiovascular biomarkers focus issue should be replete with genomic-inspired assays making their way to the clinic.
2 Nonstandard abbreviations
- cTnI
cardiac troponin I
- BNP
B-type natriuretic
- GWAS
genome wide association studies
- EMR
electronic medicine record.
Author Contributions:All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest:Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: G. Ginsburg, guest editor, Clinical Chemistry, AACC, Duke University, and CardioDx.
Consultant or Advisory Role: None declared.
Stock Ownership: G. Ginsburg, CardioDx.
Honoraria: None declared.
Research Funding: G. Ginsburg, National Institutes of Health.
Expert Testimony: None declared.
Patents: None declared.
References
