INTRODUCTION — Personalized medicine (also termed personalized genomics, genomic medicine, or precision medicine) refers to the application of patient-specific profiles, incorporating genetic and genomic data as well as clinical and environmental factors, to assess individual risks and tailor prevention and disease-management strategies.
This topic reviews concepts in personalized medicine, including the use of genetic testing marketed directly to consumers (DTC testing).
Principles of genetic testing, genetic counseling and pharmacogenomics are discussed separately.
●Testing – (See "Genetic testing".)
●Counseling – (See "Genetic counseling: Family history interpretation and risk assessment".)
●Pharmacogenomics – (See "Overview of pharmacogenomics".)
TERMINOLOGY — Personalized medicine involves the use of an individual's clinical information or genetic profile to guide decisions regarding prevention, diagnosis, and treatment of disease [1]. The definition encompasses a broad range of clinical practices in which genetic test results are used to guide patient care.
Additional terms related to genetic testing are defined in the table (table 1).
BENEFITS AND LIMITATIONS
Potential benefits — Advocates for personalized medicine cite its potential to yield significant health benefits for patients, practitioners, and society, including:
●Improved medical decision-making
●Delivery of appropriate therapies that are tailored to a patient's sequence variants or genotype rather than the general population
●Optimized disease prevention strategies, including lifestyle and behavioral modification, as well as pharmaco-prevention
●Avoidance of medications of lower efficacy
●Reduced exposure to medications that have the potential for greater toxicity, with resulting lower incidence of treatment-related side-effects and complications
In turn, this could lead to enhanced patient satisfaction with the treatment process, improved adherence to therapy, and reduced healthcare costs.
Several studies have demonstrated that the availability of genetic information enhances patient adherence to behavior modification and other disease prevention strategies [2-6].
●Smoking cessation – Alpha-1 antitrypsin (AAT) deficiency is known to cause severe lung disease in smokers. In a study of 199 smokers, individuals who were homozygous for an AAT null allele with severe deficiency of the resulting protein, were significantly more likely to attempt to quit smoking (59 percent) than either mutation carriers (heterozygotes; 34 percent) or individuals with a normal genotype (26 percent) [2].
In a study of 261 smokers who were asked to consider one of two randomly assigned hypothetical scenarios, individuals who were informed of having genetic risk factors for heart disease were more likely to report an inclination to quit smoking than individuals in a high-risk (but non-genetic) group [5]. A substantial proportion of respondents reported that their decisions were motivated by the belief that smoking cessation would translate to reductions in heart disease, suggesting that genetic risk factors would not negatively influence the ability to modify behavior and improve risk factor avoidance.
●Cholesterol lowering – In a study of 781 individuals who were carriers of gene variants that cause familial hypercholesterolemia (ascertained through an affected relative), cholesterol-lowering medication use in at-risk individuals increased from 51 to 81 percent two years after learning genotype status [6]. Significant reductions in low-density lipoprotein C levels were noted, although normal levels were not achieved in most subjects.
What is not clear is whether genetic information for individuals who have no known increased risk prior to genetic testing will lead to change in lifestyle behaviors. In a sample of over 2000 patients who completed genome-wide testing through a direct-to-consumer vendor, testing did not result in any short-term change in diet or exercise behavior [7].
Momentum for the implementation of personalized medicine in clinical practice is increasing. The price of exome sequencing and whole genome sequencing continues to fall, and the catalog of disease-associated copy number variants or deleterious sequence variants, produced from genome-wide association studies, array comparative genomic hybridization (array CGH), and next-generation sequencing (including exome sequencing and whole genome sequencing), is rapidly increasing [4].
Limitations to widespread use — Skeptics also argue that, while there are selected examples where specific biomarkers or genetic tests can help guide medical decision-making, more widely implemented profiling remains challenging in clinical practice. Factors cited to support this viewpoint include the high costs of testing (which may decline), the lack of reliable predictive biomarkers for most conditions, the lack of clear therapeutic alternatives (based on genetic differences) for many conditions, and the lack of knowledge and expertise among most clinicians regarding genetics, risk prediction, and genetic counseling [8-10]. These issues are discussed in more detail below. (See 'Obstacles for implementation' below.)
Despite obstacles, there is strong support for the development of personal genomics among healthcare policy makers and research funding agencies [5,6].
Need for greater diversity in genomic databases — As stated in an editorial about the field of genomic medicine, participants (and investigators) in genomic databases are predominantly of European origin [11].
This lack of representation of individuals with non-European ancestry can adversely impact care and potentially cause harm, such as when a variant common in African Americans was misclassified as disease-causing for hypertrophic cardiomyopathy when it is in fact a benign variant. (See "Secondary findings from genetic testing", section on 'Underrepresented ethnicities'.)
PERSONALIZED MEDICINE INITIATIVES — Available biomarker assays are impacting the practice of several medical specialties, most notably in oncology.
The use of a personalized medicine approach to other diseases has also been advanced by some federal initiatives. As examples:
●In 2007, the Department of Health and Human Services (HHS) in the United States launched the Personalized Health Care Initiative (PHCI) [12]. This initiative proposed a set of goals "for achieving gene-based medical care combined with health information technology." The PHCI aims to accelerate the development of personalized treatment strategies and transform the practice of medicine towards individualized patient care.
Components of the program include translational research initiatives to develop high-throughput sequencing technologies (ie, next-generation sequencing), enhanced mapping of the genetic determinants of disease and drug responsiveness, development of an informatics infrastructure to promote electronic medical records, and broad implementation of genomic data.
●In 2015, a precision medicine initiative was proposed in the United States, which would promote enhanced use of existing and new databases of genomic information to improve diagnosis and therapy [13]. In the near-term, the initiative will focus on cancer; a longer-term aim is to generate knowledge related to a broader range of diseases.
Cancer detection — The concept of a biomarker test to detect cancer in asymptomatic individuals is especially appealing because it could lead to early diagnosis; reduce the use of more aggressive, more toxic, more costly therapy; and potentially result in a greater number of individuals being cured of their disease.
CancerSEEK is a multi-analyte blood test that detects common, cancer-associated DNA variants using a panel of 61 amplicons for cancer "driver" mutations (for which only one abnormal copy would be sufficient to cause cancer) as well as 41 protein biomarkers associated with common cancers [14]. In a series of 1005 individuals already known to have nonmetastatic cancer of one of eight common cancer types (ovary, liver, stomach, pancreas, esophagus, colon/rectum, lung, or breast), the median sensitivity for detecting cancer was 70 percent. Among individual tumor types, sensitivity was highest for ovarian cancer (98 percent) and lowest for breast cancer (33 percent). Sensitivity increased with more advanced-stage cancers (43 percent for stage I, 73 percent for stage II, and 78 percent for stage III). Specificity was over 99 percent; only 7 of 812 healthy individuals without known cancer tested positive, and it is possible that these individuals had early cancer that was not clinically detectable. The test was also able to determine the type of cancer in the majority of individuals (median, 63 percent). Limitations of the study included the presence in case patients of more advanced cancers than would be expected in the general population, and an absence of chronic inflammatory conditions in the controls that might lead to more false positive tests. Additional validation is needed before this test is incorporated into clinical practice.
Cancer treatment — In some cases, gene expression profiling may help to stratify the need for therapy or the type of therapy in patients with early-stage cancer. Examples include breast, lung, and colon cancer. (See "Adjuvant therapy for resected stage III (node-positive) colon cancer", section on 'Molecular markers and genomic profiling' and "Systemic therapy in resectable non-small cell lung cancer", section on 'Predictive biomarkers' and "Prognostic and predictive factors in early, non-metastatic breast cancer", section on 'Receptor status' and "Deciding when to use adjuvant chemotherapy for hormone receptor-positive, HER2-negative breast cancer".)
Gene expression profiling (see "Tools for genetics and genomics: Gene expression profiling") has been especially useful in classifying lymphomas, which allows distinction among a number of subtypes of disease that cannot be reliably distinguished histologically, which may lead to differences in management. (See "Pathobiology of diffuse large B cell lymphoma and primary mediastinal large B cell lymphoma" and "Prognosis of diffuse large B cell lymphoma", section on 'Gene expression profiling' and "Clinical manifestations, pathologic features, and diagnosis of peripheral T cell lymphoma, not otherwise specified", section on 'Genetic features'.)
The SHIVA trial was an early randomized trial applying personalized medicine to cancer treatments [15]. This trial randomly assigned 195 patients with metastatic solid tumors, for which standard treatments were ineffective, to receive a molecularly-targeted agent (based on the molecular profile of the tumor) versus standard of care. The median progression-free survival was similar between the two groups (2.3 versus 2.0 months). One conclusion from this study was that further research into personalized medicine is needed. Additional research may focus on combining molecularly targeted agents to overcome resistance or incorporating information about tumor evolution using circulating tumor DNA. Additional studies involving matching therapies to tumor profiles are ongoing [16-20].
The role of genetic testing in solid tumors and gene expression profiling in cancer of unknown primary site is discussed separately. (See "Next-generation DNA sequencing (NGS): Principles and clinical applications", section on 'Cancer screening and management' and "Poorly differentiated cancer from an unknown primary site", section on 'Molecular cancer classifier assays'.)
Pharmacogenetic testing — The earliest clinical implementations of genetic profiling have been in the area of pharmacogenetics, also referred to as pharmacogenomics. Pharmacogenetics is the study of variability in drug response due to genetic factors and includes the prediction of a patient's response to a specific therapy and susceptibility to toxicity and adverse events. Pharmacogenetic data may inform both the selection of a particular treatment and the individualized dose and dosing schedule for that treatment. This subject is described in more detail separately. (See "Overview of pharmacogenomics".)
Drug labels for numerous drugs include information regarding pharmacogenetic biomarkers that can be tested. Though most notably impacting dosing of medications used to treat hematologic malignancies and solid tumors, pharmacogenetic markers are also available for medications used in the treatment of infectious, cardiac, rheumatologic, and pulmonary diseases [21]. Label content includes warnings regarding genotype-specific contraindications or toxicities, dosing recommendations, or information regarding the availability of genetic tests (without specific recommendations for testing). An updated listing of pharmacogenetic biomarkers cited in drug labels in the United States is available on the US Food and Drug Administration website [22].
The most extensively studied pharmacogenetic variants are those of the cytochrome P450 drug metabolizing liver enzymes (CYPs). Fifty-eight CYPs have been characterized in humans, and functional single nucleotide polymorphisms (SNPs) that alter functional activity have been identified for many CYPs. These variants influence the metabolism of a wide range of commonly prescribed medications, including 33 with pharmacogenetic biomarker labels. However, this test has not been widely adopted for clinical use, since the clinical value of this array has not been validated in prospective studies and insurance reimbursement is infrequently available. (See "Overview of pharmacogenomics", section on 'Online resources for clinicians'.)
Prenatal testing — With the recognition that a sufficient quantity of fetal DNA is present in the maternal circulation for clinical testing, it is becoming increasingly easier and safer to accurately assess the structural integrity and sequence variation of fetal genes. Consequently, prenatal genetic testing is playing an increasing role in obstetric care, with a move towards implementation of personalized approaches in fetal medicine [23]. (See "Prenatal screening for common aneuploidies using cell-free DNA".)
An example is the implementation of a non-invasive prenatal test (NIPT) of cell-free fetal DNA (cf-DNA), also called free fetal DNA (ff-DNA) sequences from maternal blood samples for the RhD allele of the Rh blood group. In approximately 10 percent of pregnancies in non-Hispanic White individuals, the mother is RhD-seronegative and carries an RhD-positive fetus. These mothers are at risk of RhD sensitization during delivery, with consequential risk to subsequent RhD-positive pregnancies for life-threatening hemolytic anemia. Maternal sensitization is preventable with timely administration of anti-D immune globulin, and anti-D prophylaxis of RhD-negative mothers is standard in prenatal care. However, approximately 40 percent of RhD-negative mothers (those carrying RhD-negative fetuses) receive anti-D unnecessarily. (See "RhD alloimmunization in pregnancy: Management".)
Non-invasive genotyping assays have been developed that type the RHD allele in ff-DNA circulating in maternal blood. The assay has a diagnostic accuracy approaching 100 percent [24], with better accuracies noted with newer assays that include rigorous positive controls to confirm sufficient quantities of circulating fetal DNA [25]. However, the test performs less well in non-White populations, due to the presence of alternative sequence variants. In a population with European ancestry, homozygous deletion of the RHD gene is the most common cause of the RhD-negative phenotype, whereas in 82 percent of RhD-negative patients of African American ancestry, the RhD-negative phenotype can be due to one copy or two copies of RHD variant genes, an RHD pseudogene or an RHD-CE-Ds hybrid gene [25]. As neither of these genetic variants produces any epitopes of RhD, it is important that false-positive results that might result from the presence of these variant genes in populations of mixed ancestry are considered [25]. Non-invasive screening of RhD-negative pregnancies using this method could spare about 40 percent of mothers from unnecessary exposure to anti-D therapy and avoid more invasive testing (ie, amniocentesis) in previously sensitized mothers during subsequent pregnancies. Many European countries have implemented routine ff-DNA screening for RhD-sensitized mothers. Fewer support the more generalized screening in all RhD-negative mothers, largely due to the high costs of mass screening. For example, a cost-effectiveness analysis in the United Kingdom concluded that at current test and immunization costs, routine ff-DNA screening of all RhD-negative mothers would not result in cost-savings or appreciable health benefits [26]. However, strategies to lower the costs of testing, including bundling of this test with other testing (eg, other blood groups, structural genetic aberrations), may result in broader implementation.
Other applications of cf-DNA in testing (eg, for trisomy 21) are presented separately. (See "Prenatal screening for common aneuploidies using cell-free DNA".)
DIRECT-TO-CONSUMER TESTING
Evolution of DTC testing — The availability and scope of direct-to-consumer (DTC) genetic testing services is evolving, and various legislative and regulatory bodies are actively developing policies regarding these services. Some companies are seeking approval for testing without physician involvement, whereas others are obtaining a physician order based on an online survey. Personal genome testing has been available to the general public in the United Kingdom and Canada since late 2014 [27].
In the Impact of Personal Genomics Study survey, 54 of 961 respondents (5.6 percent) who had recently undergone DTC testing reported either starting a new medication or changing a medication they were already taking based on the results of DTC tests. Most of these patients (83.3 percent) reported discussing the change with a health care provider. Importantly, the vast majority of respondents (875 people, 91.2 percent) had at least one pharmacogenetic variant indicated on their report, and there was a direct association between the number of positive pharmacogenetic tests received and the likelihood of a change in medication prescription (1.57-fold increase in the odds of a change for each positive test result, 95% CI 1.17, 2.11) [28].
The following illustrates the history of the availability of genetic testing services by the company 23andMe, which has been the most active in seeking regulatory approval for DTC genetic testing from the US Food and Drug Administration (FDA):
●Initial marketing – In approximately 2006, 23andMe began directly marketing genetic testing services to the general public; this included testing for risk variants associated with certain medical conditions, as well as genealogy/ancestry.
●FDA ruling – In late 2013, the FDA requested that the company discontinue marketing of their health-related personal genome service (PGS; ie, disease-risk prediction) in the United States, because the clinical validity of the service had not been demonstrated by the company [29]. The FDA considered the PGS kit to be a device and hence under their regulatory purview. The company suspended its clinical services in the United States but continued to provide services related to raw genetic data, information about carrier status, and ancestry testing, as well as testing services in the United Kingdom.
●FDA authorization for carrier states – In February 2015, the FDA authorized for marketing DTC carrier testing by 23andMe in the United States for Bloom syndrome (an autosomal recessive disorder) [30]. Simultaneously, the FDA classified all carrier-screening tests as class II medical devices, subject to general controls such as misbranding. The FDA subsequently exempted these tests from premarket review. With these regulatory changes, 23andMe and several other companies have begun offering DTC carrier screening for a variety of conditions. As an example, in early 2017, 23andMe began providing DTC testing for increased risk of a predetermined set of 10 conditions that include celiac disease, factor XI deficiency, G6PD deficiency, and late-onset Alzheimer disease; results of this testing might result in lifestyle modifications and/or discussions with a clinician [31]. It should be noted that these tests improve patient access to information about carrier status, but it is unclear how well they will be used, particularly in the absence of pre- or post-test genetic counseling. Results are returned to patients using reporting formats approved by the FDA.
●FDA authorization for cancer risk genes – In March 2018, the FDA authorized marketing DTC testing for three BRCA1 and BRCA2 mutations that are common in individuals of Ashkenazi Jewish descent [32]. These represent a very small subset of BRCA mutations, which number over 1000, and they are not common in the general population. Customers can elect to receive or not to receive these results when they register for non-clinical (eg, ancestry) testing. The FDA advised that information from this cancer risk testing "should not be used…to determine any treatments."
In January 2019, the FDA authorized marketing for two MUTYH mutations. Biallelic deleterious variants in MUTYH are associated with adenomatous polyps and an increased risk of early-onset colon cancer; individuals with one deleterious variant in MUTYH have a twofold to threefold increase in their risk for colorectal cancer at an age similar to that in the general population [33,34]. Similar caveats for this testing apply as those for BRCA testing, including the much larger number of colon cancer risk genes that are not being tested, and concerns about validity of the results, possible misinterpretation by patients (eg, assuming their colon cancer risk is low based on analysis of only two variants in a single gene), and unclear implications for management.
DTC testing commonly uses a kit for saliva-based testing that can be sent through the mail.
Concerns about value, accuracy, and interpretation — Results of DTC testing should not be used to make a new diagnosis or initiate a new treatment without first confirming the results of testing in a Clinical Laboratory Improvement Amendments (CLIA)-certified laboratory with a proper "chain of evidence" that ensures the result applies to the tested individual. Conceptually, this is analogous to a home pregnancy test performed soon after a missed menstrual period; the results may be correct, but standard practice is to obtain a confirmatory test before initiating a new pregnancy-related medication or instituting additional prenatal care [35,36]. It is imperative that physicians are aware of this caveat and are familiar with the potential limitations of DTC, which are outlined below. There is evidence, however, that some practitioners are making clinical decisions based on DTC results without first confirming the results in a CLIA-certified laboratory.
A number of concerns have been raised about the accuracy, interpretation, and value of DTC genetic testing [37]. The following examples illustrate some of these concerns:
●Reliability and reproducibility – Given the heterogeneity among DTC vendors in the markers tested and the predictive models used to develop risk estimates, questions have been raised regarding the reliability and reproducibility of these services. The accuracy of identifying mutations or disease-associated single nucleotide polymorphisms (SNPs) is one concern; the accuracy of using the resulting information to predict the risk of disease is a separate concern of equal or greater importance.
Two studies have evaluated these issues; both have found high concordance between companies in determining genotype, but significant variability in interpretation of disease risk, despite agreement on genotype:
•One study compared genotyping panels from two DTC companies: 23andMe and Navigenics [38]. DNA from five individuals was sent to each company for testing, and risk predictions for 13 diseases were compared. Genotype calls (ie, the genotypes reported at each locus based on SNP analysis) were in excellent agreement (>99 percent concordance). In contrast, risk prediction for diseases was discordant in about a third of cases. Disease risk predictions from the two companies disagreed more than half of the time for seven conditions (eg, systemic lupus erythematosus, heart disease, Crohn disease, type 2 diabetes); risk predictions were concordant in four conditions. Conditions demonstrating strong agreement were those in which an identical SNP with a very strong genetic effect was tested by both companies (for example, the SNP for celiac disease that confers a sevenfold increase in risk). SNPs with such strong effects are rarely observed in common complex diseases.
•A second study evaluated the concordance across three companies (23andMe, Navigenics, and deCODE), although only one sample was tested [39]. Genotype concordance was high (99.6 percent), but significant variability in risk prediction was noted, with undetermined clinical validity and utility. The variability in risk prediction was influenced by the SNPs that were genotyped for each condition and the reference population used [39]. Diseases for which genotyped SNPs have strong predictive value were more likely to receive similar risk estimates from different DTC companies [40]. This study also pointed out the need for risk data that is based upon client ethnicity, as the majority of genome-wide association studies have focused on populations with European ancestry, with uncertain applications to other ethnicities, including people with ancestry from Asia, Africa, and other continents [39,40].
These reports suggest that risk prediction remains unreliable, and interpretation of results for most SNPs should be approached with caution. CLIA-certified laboratories also have variability in interpretation, but they are probably more concordant than DTC laboratories.
●Generalizability – In most cases, data about genetic risk has been obtained from testing individuals with a known personal or family history of disease. Data from testing a healthy population has not yet been obtained for most conditions, and it has not been demonstrated that risk is similar when a variant is identified in an individual with a negative personal and family history of disease.
●Implications for management – Results from genetic testing, both normal and abnormal, may have major implications for management. In their 2013 notification to the 23andMe company regarding screening for medical risk variants, the US FDA expressed concerns regarding the potential misuse of reported genetic information leading to inappropriate management. As an example, patients might make treatment decisions regarding prophylactic mastectomy, chemoprevention, or aggressive surveillance based on false-positive or false-negative BRCA1 genotypes. An editorialist suggested that the FDA may not have intervened if the genome service had been available only to an ordering physician, rather than marketed directly to consumers [41]. The implications of results should be discussed with a clinical geneticist, genetic counselor, or clinician with expertise in the condition in question, but as the amount of DTC testing increases, resources to provide such counseling may become strained.
Role in lifestyle modification — To the extent that results of DTC testing promote healthy lifestyle and health maintenance interventions such as exercise, healthy diet, and recommended screenings (eg, cancer screening, risk assessment for cardiovascular disease), they may be likely to be of value to the individual. For disorders in which other interventions may be appropriate (eg, medication, surgery), the value of speaking with a clinical geneticist, genetic counselor, or expert clinician should be emphasized regardless of the results of DTC testing. These experts may help with interpretation of results, additional testing not provided in the DTC product, and interventions most likely to be helpful. (See "Genetic counseling: Family history interpretation and risk assessment".)
In addition, patients should be made aware that:
●For all but a limited number of SNPs of large effect, the majority of SNPs tested in DTC test kits provide only incremental changes in a patient’s risk profile.
●Prospective studies of the predictive accuracy of these products have not been performed, thus precluding the provision of effective counseling or reliable decision-making for most results.
●There is substantial inter-company variability in the risk estimates reported [39,40].
●The influence of non-genetic factors, including race and lifestyle factors, on the interpretation of these results is unknown for many of the SNPs tested.
These caveats make the point that for most conditions, a negative test result (ie, susceptibility variant not identified) does not guarantee low risk for disease development, and positive lifestyle modifications should be encouraged regardless of DTC testing results.
OTHER PERSONALIZED MEDICINE PLATFORMS — The early and intuitive focus for personalized medicine has been the development of genetic-based tests. Other "omic" approaches are being developed that will provide a more complete characterization of risk that includes variation between individuals in gene regulation, epigenetics, and cellular metabolism [42]. Such approaches, which are under development as part of research and rarely provided clinically, include:
●Gene expression profiling (also referred to as transcriptomics) – Analysis of mRNA (of either individual genes or panels of gene targets), representing gene expression patterns; often uses microarray technology, although whole transcriptome analysis (also known as RNA-Seq) is also possible and can be used as an aid for variant interpretation. Gene expression is dynamic and influenced by a range of cellular, genetic and environmental factors, which makes gene expression a particularly attractive target for profiling malignant cells. (See "Tools for genetics and genomics: Gene expression profiling".)
●Proteomics – Qualitative and quantitative analysis of the collection of protein constituents in a biological sample. Typically performed using modification of polyacrylamide gel electrophoresis (PAGE) or matrix-assisted laser desorption/ionization (MALDI) approaches, these methods provide measures of the types and abundance of proteins in a biological sample. Proteomics assays are under investigation in certain tumors.
●Metabolomics – The characterization of metabolic profiles; typically consists of a collection of assays that characterize panels of metabolites related to specific pathways. These studies can be static (cross-sectional profiling at a given time-point) or dynamic (assessing the change in profile patterns following a specific metabolic challenge) [43]. In combination with separation methods such as high-performance liquid chromatography or gas chromatography, metabolites are typically characterized either by mass spectrometry (MS) or nuclear magnetic resonance spectroscopy (NMR).
●Lipidomics – Characterization of the complete collection of lipids. Lipid structures, like metabolites, can be differentiated by MS or NMR [44,45]. These methods are being applied towards the development of diagnostic tests that assess the lipid composition of cell membranes [46].
●Epigenomics – Profile of the modifications to DNA (often, methylation) that control gene expression. Unlike genomic changes, epigenomic changes are affected by the environment and may change with age, stress, or exposures to the individual or earlier generations. (See "Principles of epigenetics".)
●Exposomics – The sum of exposures an individual incurs over a period of time. These may include nutrients, foods, toxins, stresses, exercise, vaccinations, medications, and other exposures. The exposome is highly dynamic and malleable over an individual’s life.
●Microbiomics – Characterization of the microbes (typically, bacteria) that reside in or on an individual. A common example is the gut microbiome, which might influence adiposity and/or immunity. (See "Anaerobic bacteria: History and role in normal human flora".)
These applications are being increasingly applied to clinical practice, and development of reliable, clinically adaptable assays for these platforms is actively being pursued. One example is in the field of oncology, where gene expression profiling of malignant cells or tissue is becoming an important diagnostic and prognostic tool. Some of the earliest and most successful implementations of gene expression profiling in oncology focused on single genes, such as determining estrogen receptor expression status in breast cancer for informing prognosis and chemotherapeutic options [47]. Subsequently, whole transcriptome expression profiling technologies have been applied to tumor samples, with notable success [48,49]. (See "Tools for genetics and genomics: Gene expression profiling", section on 'Overview of clinical applications'.)
OBSTACLES FOR IMPLEMENTATION — Despite early successes in the clinical introduction of a limited number of pharmacogenetic assays, multiple barriers preclude the widespread implementation of personalized medicine as standard clinical practice across all medical fields. The development of validated biomarkers and genetic assays represents an important bottleneck, as does the analysis of the masses of data than can result from next-generation sequencing (NGS) technologies, although it is possible that the prospect of affordable exome and whole genome sequencing will remove this obstacle in the future. Other ethical and data storage challenges will remain.
●Limited predictive value of most tests – There are a limited number of examples where genetic testing provides substantial gains in guiding therapeutic recommendations for healthy individuals, although diagnostic genetic testing may be very useful for patient management and recurrence risk estimation. However, the number of genetic variants associated with disease susceptibility and pharmacogenetic response is increasing, especially in the oncology field.
●Lack of physician knowledge – The lack of general knowledge of genetics among many medical practitioners is one of the most pressing challenges preventing broad implementation of personalized medicine. Many medical practitioners do not feel adequately prepared to provide counseling for genetic testing [50] and the number of genetic counselors and clinical geneticists in North America is considered to be lower than required to meet expected demands. However, medical schools are actively developing curricula content in the areas of personalized medicine to prepare the next generation of physicians, including several programs that have developed case-studies of whole genome sequence data [51,52]. However, there is little infrastructure in place to facilitate the education of clinicians already in practice.
Definitions of commonly used genetics terms are presented separately in UpToDate. (See "Genetics: Glossary of terms".)
Several online pharmacogenetics resources are available to assist with clinical decision-making, as listed in a separate topic review. (See "Overview of pharmacogenomics", section on 'Use of genomic biomarkers to guide therapy'.)
●Inadequate informatics infrastructure – The record-keeping for clinical laboratory data resides with the health care providers (and their affiliated institutions) who originally requested the tests. This model is inadequate to accommodate NGS data, and newer approaches that include an easily accessible electronic medical record (EMR) are under development. Although adoption of EMRs has increased, the majority of those in place would require substantial enhancement to accommodate the data storage and processing needs anticipated for genomic data. The Personalized Health Care Initiative has made the nationwide development of EMRs a top priority [12], and the Office of National Coordination for Health Information Technology (ONC) was established in 2004 by executive order to facilitate the establishment of an electronic health record (EHR) across the United States that would coordinate information across multiple EMRs.
●Information provenance and patient privacy issues – Many questions arise when considering how personal genomic information should be managed.
•Who would be responsible for ordering a pharmacogenetic genotyping panel (ie, the major CYP loci) that profiles responsiveness for a wide array of drugs?
•How would such information be transmitted to other clinicians and pharmacists?
•When should genomic profiling be ordered? If personal genomic profiling is performed as part of a neonatal screening program, who will be the purveyors of such data?
•Would pediatricians be responsible for interpreting data and facilitating the development of patient management plans related to conditions of adult onset?
Many geneticists do not recommend testing for adult onset disorders in childhood unless there are diagnostic or management implications relevant to the child at the age of testing. (See "Genetic testing", section on 'Ethical, legal, and psychosocial issues' and "Genetic testing", section on 'Which relatives should be tested?' and "Genetic testing", section on 'Obtaining informed consent'.)
•At what age and how would children be informed of their specific liabilities?
•Should information be provided for conditions for which there is no known cure or treatment?
●Inconsistent standardization and oversight of testing – As illustrated by the preliminary concordance studies of direct-to-consumer (DTC) genetic tests (see 'Direct-to-consumer testing' above), there is considerable inconsistency in the prediction of genetic risk [38,39]. These inconsistencies are largely due to differences in the risk-estimation models used and in the single nucleotide polymorphism (SNP) content on the array platforms. Standardization of these models and the SNP content will help improve this problem. The US FDA is developing policies regarding standardization and oversight of these tests, and policies have been discussed in Australia and the United Kingdom; some have advocated for international certification of quality standards [53].
●Reimbursement issues – Changes in the reimbursement policies will be needed to promote a personalized medicine initiative. Several attempts have been made to modify the Medicare clinical laboratory fee schedule to accommodate the reimbursement for complex genetic tests.
●Genetic discrimination – Discrimination based on the results of genetic testing or even on the basis of having obtained a test is a concern (eg, denial of insurance coverage or other services). Many jurisdictions have tried to put protections in place, such as the Genetic Insurance Non-discrimination Act (GINA) in the United States, which provides some protections (eg, for health insurance but not for life insurance). (See "Genetic testing", section on 'Genetic discrimination'.)
●Societal issues and misconceptions – Numerous societal challenges must be addressed prior to widespread implementation of personalized medicine [54,55].
•Acceptance of genetically-based treatment recommendations may be low, particularly among certain groups [48].
•Concern that misconception of genetic risk could result in a medicalization of society, whereby healthy individuals become preoccupied with disease prevention based on their profile or undergo unnecessary tests or procedures. One study of individuals who underwent direct-to-consumer genome profiling found no measurable evidence for increased anxiety, comparing baseline and post-testing anxiety scores [7].
•Genome sequencing will likely identify a variety of "high-risk" alleles in all individuals. With the limited available understanding of risk concepts or penetrance of mutations and risk alleles, the identification of risk alleles or recessive traits could inappropriately influence reproductive decisions.
•Individuals may be adversely impacted when advised of increased risk for sensitive issues, such as psychiatric disorders or behavioral traits.
•The desirability of learning if an individual has an increased risk for conditions for which there is no available treatment or prevention is questionable. In one study, more than 80 percent of at-risk individuals who had no symptoms chose not to pursue testing for Huntington disease [56].
•Family dynamics might be affected by disclosure, or failure to disclose, genetic information.
INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)
●Basics topics (see "Patient education: Genetic testing (The Basics)")
SUMMARY
●Definition – Personalized medicine (also called precision medicine) involves the use of an individual's genetic profile to guide decisions made with regards to the prevention, diagnosis, and treatment of disease. (See 'Terminology' above.)
●Rationale – Potential benefits of personalized medicine include customized treatment plans that can include targeted pharmacotherapy to improve drug response and reduce toxicity and expense, targeted recommendations for lifestyle modifications and other disease-prevention strategies, and enhanced patient satisfaction with healthcare. Genetic information can enhance patient compliance with behavior-modification recommendations. (See 'Personalized medicine initiatives' above.)
●Drug dosing – The earliest implementation of genetic profiling into clinical practice has been in the area of pharmacogenetic testing. Pharmacogenetic biomarkers are available for over 70 drugs; many of these markers involve polymorphisms in cytochrome drug metabolizing activity. (See 'Pharmacogenetic testing' above and "Overview of pharmacogenomics".)
●DTC testing – The use of direct-to-consumer (DTC) genetic testing services is evolving, and various legislative and regulatory bodies are actively developing policies regarding these services. Genetic testing for carrier status is considered a class II medical device and does not require premarket approval by the US Food and Drug Administration (FDA). Clinicians and patients should be aware of the concerns that have been raised about the accuracy, interpretation, and value of DTC genetic testing. (See 'Direct-to-consumer testing' above.)
●Other "omics" initiatives – Analyses based upon gene expression profiling, proteomics, metabolomics, or lipidomics may enhance the predictive value of testing used for personalized medicine. Gene expression profiling is used for some tumors, reflecting cellular, genetic, and environmental factors. (See 'Other personalized medicine platforms' above.)
●Challenges to implementation – Multiple issues will need to be addressed to successfully implement personalized medicine. Beyond the development of validated biomarkers, areas of concern include physicians’ knowledge base of genetics and risk interpretation; infrastructure to store and confidentially retrieve an individual’s genetic data; timing of testing; issues of providence in terms of clinicians responsible for ordering and transmitting information; test standardization and quality; reimbursement; and overcoming public misperceptions related to genetic data. (See 'Obstacles for implementation' above.)
●Related considerations – Comprehensive reviews of genetic testing and counseling are provided separately. (See "Genetic testing" and "Genetic counseling: Family history interpretation and risk assessment".)
