Technologies in Personalized Medicine
High-throughput technologies in genomics were valuable in completing the human genome. Further improvement in bioinformatics and prediction models made it possible to conduct molecular classification of different tumor types.[7,8] Currently, multiple biomarkers can be followed in the same sample. The following section highlights major technological advancements and their implications in personalized medicine in cancer.
Genomics in Personalized Medicine
Major advances in the field of molecular genetics, especially following completion of the human genome, expanded our ability to identify genetic events that underlie the pathogenesis of diseases that follow Mendelian inheritance. Although hereditary cancers account for only 8–10% of cancers overall, considering the high incidence and prevalence of cancer worldwide, genetic information can be useful in personalizing cancer medicine. Pharmacogenomics and pharmacokinetics information is useful in determining proper drug doses for the treatment of cancer. Genetic alterations that have been studied for use in personalizing cancer medicine include: CYP polymorphisms, UGT1A1 polymorphisms, thiopurine methyltransferase (TPMT) polymorphisms, dihydropyrimidine dehydrogenase (DPYD) polymorphisms, cytidine deaminase (CDA) polymorphisms and CXR polymorphisms. The US FDA has recommended using irinotecan and mercaptopurine in patients with polymorphisms in the UGT1A1 and TPMT genes, respectively. Funding agencies such as the NIH showed great enthusiasm for developing programs to conduct genome-wide association studies to determine phenotypic characteristics during disease development. Mutations and other genomic alterations are being cataloged in the International Cancer Genome Consortium; and cancer specimens are being characterized by genomic, transcriptomic and epigenomic analysis in the Cancer Genome Atlas project of the National Cancer Institute (NCI), NIH.[14,15] These programs have reduced previous concerns about the cost, time, complexity and availability of tissues.
Elements of risk assessment, diagnosis, prognosis and treatment are part of the applications of genetics and genomics in personalized medicine and the path from idea to clinical implication involves basic, translational and regulatory science (with guidelines provided by the FDA). Four phases of genomics and medicine in personalized medicine from the start to outcome have been proposed: discovery and replication of findings (T1 phase), evaluation of markers for validity and utility (T2 phase), evaluation of the best approaches for the diffusion and dissemination of markers (test/assay; T3 phase) and translation from bench to bedside (T4 phase). All these phases will involve research addressing the efficacy and impact of these tests on improved health. Here it is emphasized that genetics refers to the study of single genes in contrast to genomics which encompasses all the genes of a person. A few examples of genetic and genomics testing in personalized medicine include Lynch syndrome testing for hereditary colon cancer, BRCA1/2 testing for breast cancer, fusion genes and rearrangement of BCR–ABL, TEL-AML1 and MLL in pediatric leukemia, hepatitis virus for hepatocellular carcinoma, papillomavirus for cervical cancer, EGFR point mutations in lung cancer and glioblastoma and cetuximab, gefitinib, erlotinib, panitimumab, lapatinib treatment, BRAF mutations in melanoma treated by RAF inhibitors, PARP inhibitors in BRCA mutant breast, ovarian, prostate and pancreatic cancer.
Non-Mendalian patterns of inheritance of disease susceptibility are seen in case of genomic imprinting (in cancer), de novo germline mutations and epigenetic inheritance. The de novo mutation rate may go as high as 30–50% in neurofibromatosis and hereditary endocrine tumors whereas in other tumors the rate varies too much. Mismatch repair genes exhibit epimutations in colorectal cancer and promoter silencing by methylation has been observed in a number of tumor types. There is always scope of precaution while practicing personalized medicine and the National Institute of Human Genome Sciences suggested that even an analytically and clinically validated genetic test may fail as a tool unless it is translated into a behavioral action by the at-risk individual.
Compared to the candidate gene approach, molecular profiling and whole-genome sequencing are more powerful technologies because their results are unbiased. Furthermore, whole-genome sequencing can detect structural variants (deletions, amplifications, translocations and inversions) and allow for confirmation of a suspected diagnosis, even if caused by a rare or unusual mutation. Genome-wide association studies in cancer have identified potential polymorphisms associated with different cancers.[16–20] Because cancer is considered a genetic and epigenetic disease, epigenome-wide association studies may also be helpful in personalized medicine.[21–24] These association studies are useful in identifying individuals who are at high risk of developing cancer and should be treated first. Offit has suggested that clinical validation of genetic markers may take place even in the absence of a complete understanding of their functional biological significance. At the same time, the author cautioned against the potential harms of premature translation of research findings.
Epigenomics in Personalized Medicine
Epigenomic alterations do not include nucleotide changes and affect only the profiling of methylated regions, miRNA, and histones. In normal individuals, the entire genome becomes hypomethylated with age, whereas gene-specific (and cancer-specific) sites become hypermethylated. Histones also are modified by post-translation due to acetylation, phosphorylation, ubiquitination and other modifications. Tumor-specific miRNAs also have been identified that could be targeted for therapy. Considerable progress has been made during the past 5 years in determining the clinical epigenetics of cancer, and outcomes from the NIH Roadmap Epigenomics Program have indicated the importance of epigenetic mechanisms in cancer development. Methylation, histone and miRNA profiling of samples from cancer patients have generated numerous biomarkers for use in following the effectiveness of treatment.[4,5,26] Epigenomic profiling provides great promise for improving the efficacy of patient treatment evaluation and for following outcomes. NCI's Epidemiology and Genomics Research Program recently evaluated epigenetic biomarkers in cancer and their validation in the epidemiology setting.[26–28] Because the epigenotype represents the effects of external stimuli on the genomic background, epigenomic profiling provides information about gene–environment interaction and offers greater potential for making decisions about personalized cancer treatment. Four epigenetic drugs have been approved by the FDA, but correlation of the efficacy of these drugs against the epigenomic background of individuals has not yet been systematically studied.[26–28,30–35] The methyltransferase inhibitors azacitidine and decitabine have been approved by the FDA for clinical therapy in patients with myelodysplastic syndrome. Among others, zebularine and isothiocyanate also are potential methyltransferase inhibitors. 5-azacytidine (Vidaza) forms a covalent complex with cytosine (C-5)-specific DNA methyltransferases and inhibits their activity. This compound also is activated by uridine–cytidine kinase and thus can be incorporated into both RNA and DNA. Epigenetic inhibitors, either alone or in combination, work effectively in cancer treatment. Two examples are described below where epigenetic inhibitors were used. Epstein–Barr virus infection has been observed in Burkitt lymphoma and nasopharyngeal carcinoma, and it has been determined that hypermethylation of latency genes allows Epstein–Barr virus to stay integrated in the main genome and contribute to these cancers. In myelodysplastic syndrome, with an incidence rate of 20,000 new cases per year in the USA, the standard treatment is weekly blood transfusion, which is considered aggressive treatment. Treatment with Vidaza, an epigenetic inhibitor, was successful in these patients, and most of them did not need transfusion. The survival time and quality of life of these patients increased substantially.
Histone methylation in infants is different than in childhood. Methylation of the IGF2 and long interspersed integrated nuclear element sequences in infants of nonsmoker mothers differed from those with smoker mothers.[8,36] At this stage, carcinogens do not show genomic alterations but epigenetic alterations can be observed. Future studies may predict which infants will develop lung or other cancers.
Metabolomics & Other Technologies in Personalized Medicine
Based on the results from mass spectrometry and nuclear magnetic resonance-based spectrometry, metabolomic profiling provides a comprehensive overview of the metabolites that lead to the characterization of a disease phenotype.[37–46] Metabolite profiling is considered very close to the phenotype compared with genomic and transcriptomic profiling. Metabolomic biomarkers have been used in cancer diagnosis and in following response to treatment in bladder, breast, colon, esophageal, lung, pancreatic, prostate and urogenital cancers.[47–64] There are many potential opportunities to evaluate and validate potential metabolomic biomarkers that could enhance drug development and clinical decision-making about treatment. The only significant challenge in metabolomics is that few epidemiology studies have been conducted to date, and further research is needed to validate metabolomics biomarkers and establish correlations between these and other biomarkers. For better outcome results, longitudinal studies are needed in which patients are followed for their clinical outcomes.
Imaging Technologies in Personalized Medicine
In oncology science, MRI, PET and computerized tomography (CT) are the most common technologies for diagnosis. The PET and CT technologies are much more sensitive than MRI, although the radiation doses involved in MRI are much lower. In practical terms, treatment follow-up can be accomplished using MRI, but PET and CT scanning are essential for initial diagnosis. Bone metastasis in specific prostate cancers and tumor hypoxia can be followed by PET scanning. Metabolomic disorders during cancer development can be followed by MRI-based metabolite characterization in blood and urine samples of patients. These technologies are crucial in determining effective dose and follow-up outcome from surgery. Furthermore, these technologies, especially MRI which has the least radiation exposure, are also suitable for screening populations to identify high-risk individuals. A recently developed digital immunohistochemistry algorithm has been automated for determining the estrogen receptor, progesterone receptor and HER2 status in breast cancer samples. Automated microscopy, computerized processing and digital imaging have provided increased accuracy in quantification and standardization of values which can be used in the clinic.
The characterization of cancer stem cells (CSCs) has not been explored extensively but has implications for personalized medicine. These cells are few in number in a tumor (one to three in 1 million cells) and initiate cancer. CSCs are generally resistant to chemotherapy. Characterizing CSCs for specific tumor types and targeting these cells for elimination holds potential for advancing cancer control therapy. CSCs also can be useful in detecting cancer early, which is advantageous because combinations of therapies (chemotherapy, radiation therapy and new-generation targeted immunotherapy) can be planned and applied to individual cancer patients.
Tumor-associated antigens are specifically expressed by malignant cells that are recognized by the host immune system. Identifying antibody characteristics may help in early diagnosis and individualized anticancer treatment. Biopanning technology, used to enrich the immune response, has shown promising results in head and neck cancer.
Circulating tumor cells (CTCs), sometimes referred as 'liquid biopsies,' are shed from primary or secondary tumors. CTCs could be a good source for conducting next-generation sequencing and may be useful in personalized medicine, as has been suggested in prostate cancer. Because CTCs are collected from blood, physicians can monitor response to treatment closely for the most effective outcome. The presence of CTCs in breast, prostate, colon and other cancers has implications for understanding metastasis and for predicting progression-free and overall survival. The major challenge in CTC technology is the origin of CTCs. To confirm the evolution of a tumor, other biochemical and molecular data should be included in treatments based on CTC analyses. Enrichment technologies for CTCs are available, which facilitate molecular characterization of these cells.
The consensus among scientists is that understanding the mechanisms and pathways that contribute to primary or secondary resistance to therapy will yield biomarkers that predict response to therapy in individuals. Abernethy and Basch discussed the role of clinical cancer informatics in creating pathways for personalized cancer care. Considering the high incidence rate of most cancers, personalized therapy at both the individual and personal level should be developed. For proper implementation of personalized medicine, a cancer patient's records should be maintained electronically, and bioinformatic algorithms should be developed that can integrate molecular data with clinical and epidemiologic data.[27,34,69–71] The clinical utility of tumor markers in cancer has been evaluated and published as the NCCN Task Force Report. Guidelines and suggestions in this report should be followed in developing new assays for cancer diagnostic and prognostic biomarkers.
Personalized Medicine. 2014;11(8):761-771. © 2014 Future Medicine Ltd.