Personal Genomics and Genome-Wide Association Studies: Novel Discoveries but Limitations for Practical Personalized Medicine
- First Online:
- Cite this article as:
- Roukos, D.H. Ann Surg Oncol (2009) 16: 772. doi:10.1245/s10434-008-0109-6
- 427 Views
To the Editor:
The recent article by Ellsworth et al. in this journal1 concludes that genomic alterations of the 8q24 region, including c-MYC, drive lymph node metastasis in early-stage breast cancer. The authors examined levels and patterns of chromosomal alterations in primary breast tumors from node-negative (n = 114) and node-positive (n = 115) patients. They found higher frequencies of allelic imbalance (AI) at chromosome 8q24 in patients with positive lymph nodes. This finding suggests, according to the authors, that genetic changes in this region are important to the process of metastasis. If this finding could be validated, then it could have important implications. For example, levels of AI determined in the primary breast tumor could be used as prognostic markers.
The 8q24 region may play an important role in cancer development. Recent genome-wide association studies (GWAS) have identified several genetic variants associated with significantly increased risk of breast, prostate, and colorectal cancer.2–4 A most recent study has shown that there are at least five independent loci within the 8q24 locus associated with different cancer types, including breast, prostate, colorectal, and ovarian cancer.5 However, Ghoussaini et al.5 note that it is not known whether the identified associations of the single-nucleotide polymorphisms (SNPs) with cancer are causal genetic variants or are simply markers that are correlated with the causal variants in each region.
Genetic variation—differences in the coding and noncoding portions of our DNA—is what makes each of us unique. It also can contribute to our personalized susceptibility to disease. Exhaustive analysis of human SNPs has led to the identification of interesting SNP markers for certain disorders. Genetic variants beyond SNPs include copy number variations (CNVs)—gain or loss of segments of genomic DNA relative to a reference—which have also been shown to be associated with several complex and common disorders like cancer.6
There has been an explosion in biomedical research for the identification of genetic-variation-based personalized management of complex diseases. Over the last 2 years GWAS have identified more than 100 new chromosomal regions at which more than 165 novel DNA variants influence risk of common human diseases and clinical phenotypes.7 But at the present time, despite initial optimism, the clinical implication of these risk variants is still being debated.8
Personal genomics promise to discover prognostic and predictive tools toward achieving personalized prevention and treatment of cancer,9,10 But despite the flood of risk alleles discoveries by GWAS,11 many scientists are now skeptical about whether these DNA variants could have clinical utility.12 Single, common, low-penetrance genes have limited clinical use because the risk conferred by each of these variants is modest. Thus, the clinical use of these genetic variants as markers for risk prediction is limited. However, there is some suggestion that, when combined, these risk alleles could stratify people in the general population into high-risk and low-risk categories. This polygenic model of diseases still remains to be proven by large-scale clinical studies.
Empirical GWAS have identified six breast-cancer susceptibility alleles that are common in the general population.2,13,14 Most recently, Pharoah and colleagues examined the implications of seven SNPs, both for individualized disease prevention and for public health policy.15 Each SNP has a modest influence on a woman’s risk of disease; none act through well-understood mechanisms.12 Pharoah et al. consider the potential usefulness of these markers in targeting patients who would benefit from screening for early detection of disease and argue that, as more associated loci are identified, risk prediction algorithms will need to be based on the best available risk estimates. The authors conclude that stable algorithms may eventually be useful in identifying groups of women with clinically meaningful differences in risk.15
Personalized prevention of breast cancer is currently a dream. Risk stratification of the general population into high-, moderate-, and low-risk categories for the vast majority of women is infeasible. Carriers of mutations in BCA1 BRCA2 high-penetrance cancer susceptibility genes are at high risk of breast cancer and may benefit from aggressive preventive strategies which include prophylactic surgery, chemoprevention, and magnetic resonance imaging (MRI).16–18 But this specific subpopulation accounts for only 2% of the general population.15 In contrast, low-risk women benefit from normal mammographic screening, appropriate lifestyle, and efforts to modify environmental risk factors. There was an initial optimism that GWAS would provide sufficient discriminatory accuracy through identification of genetic risk variants.11 But these identified SNPs only modestly increase risk, and a most recent study has demonstrated that this genetic information is no better than the more simple and inexpensive clinical information based on family history and “classical” risk models.19
The way toward personalized management of cancer is too long and there are multiple hurdles and challenges to be overcome. The perspectives for the completion of cancer genetic mapping through new technologies with the current availability of genotyping platforms with more than 1 million SNPs and CNVs are realistic. But a cancer genetic atlas alone has limitations. Functional studies to explore the interactions between genetic risk variants are needed. Furthermore, the understanding of genetic and environmental risk factor interactions is prerequisite for a risk-stratification-based prevention strategy.
A very large-scale, prospective, population-based cohort study recording genetic variation, family and medical history, and environmental exposures, might be a great step toward sophisticated algorithm-based personalized prevention. But the costs of such a study proposed by Dr. Francis Collins, director of The National Human Genome Research Institute, are too high, approximately US $400 million a year to run.20