Personalized medicine has become the mantra for the 21st
century—as if somehow we never before customized the care we provide to the individual patient’s diagnosis and personal needs. But the reality of choosing therapies rationally for patients with cancer has never been more apparent. In the past decade, the lessons from understanding the molecular underpinnings of specific cancer have become clear; the specific mutations in the cancer cells of an individual’s tumor can be ascertained, the mutations that drive that cancer can be isolated, and, in the best of cases, targeted with highly specific molecules which interfere with the cellular machinery that allows the cancer cell to survive, proliferate, and spread. Such is the logic, and the development of a growing list of drugs based on this paradigm is proof that this will be the near-term future of cancer therapeutics. Such molecularly targeted therapies have revolutionized the management of several leukemias and solid tumors. Even the highly resistant melanoma is yielding a bit; for patients with metastatic melanoma that harbor activating mutations in BRAF
(present in nearly half of melanomas), selective inhibitors of BRAF
have produced stunning responses.
The molecular targets of tumors affecting our patients have been the primary focus of research; less attention has been paid to the host and the potential for analogous investigations to have an impact on what we prescribe. Yet, the patient (non-tumor) side of the equation may provide key insights into personalizing care of patients with cancer. Two recent studies are pertinent. A study1
presented at the ASCO Annual Meeting in June examined genetic associations with taxane-induced neuropathy. Using genome-wide association (GWAS), several SNPs were found to be associated with shorter time to neuropathy among women receiving paclitaxel for breast cancer. Neuropathy is a common and potentially debilitating toxicity of taxane therapy. The ability to predict which patients are more likely to develop this side effect would allow better selection of therapy for the individual patient to minimize potential toxicity.
The development of secondary malignancy in patients cured of cancer is arguably the most devastating late effect of cancer therapy. Although it is known that women (especially young women and adolescents) treated for Hodgkin lymphoma are at substantially elevated risk for development of secondary breast cancer in the field of irradiation, the complication is relatively rare, and our ability to predict which patients are at risk would greatly help in planning therapy. Many recent studies of Hodgkin lymphoma have focused on eliminating or reducing the use of irradiation in an effort to avoid late complications. However, radiation is particularly important in some patients with Hodgkin lymphoma (notably those with bulky disease in the mediastinum), and withholding irradiation for such patients in an effort to avoid late effects likely increases their risk of recurrence and the need for salvage therapies and associated additional toxicities.
If we could reliably predict which women are at risk for radiation-induced breast cancer, we could more intelligently design primary therapy for women with Hodgkin disease, using radiation where it is needed for women who are low risk for this complication, while reserving different, radiation-sparing approaches for women at higher risk.
A study published recently in Nature Medicine2
provides some important early insights into addressing this issue. Investigators again utilized GWAS to identify variants associated with second cancers induced by radiation therapy among survivors of Hodgkin lymphoma. Patients surviving Hodgkin lymphoma treated with radiotherapy who did and did not develop secondary solid tumors were compared, and three SNPs achieved genome-wide significance—two of which mapped to chromosome 6q21. When an independent cohort of survivors of Hodgkin lymphoma was tested, these same two SNPs on chromosome 6q21 were significantly associated with the development of second cancers. Further studies implicated the PRDM1
gene known to be involved in cellular proliferation, differentiation, and apoptosis. The SNPs identified in this study appear to be associated with the level of expression of the PRDM1 protein in the basal state or after radiation exposure; cells with the “high-risk” haplotype were associated with lower PRDM1 expression and failure to upregulate expression of the gene after exposure to radiation. This study represents an important step toward understanding genetic predisposition to radiation-induced breast cancer.
While “personalized medicine” has always been the goal of what we prescribe for our patients, we may soon have the tools to deliver it. The prospect of “tailoring therapy” to the patient’s individual risk of recurrence (personalized treatment based on tumor characteristics) as well as the patient’s individual risk for acute or long-term toxicity (personalized treatment based on host characteristics) represents a glimpse of the future of oncology, but highlights the enormous challenge for the future as well. We will be expected to digest and act upon an enormous amount of information about each patient derived from the tumor genome and the host genome, and we will match the pathway dysregulation disclosed by interrogating the tumor genome with agents that target those specific pathways in the tumor, while being certain to choose an agent that best spares the patient the toxicities predicted by his pharmacogenomic profile. We will be highly dependent on the informatics capabilities of our electronic record systems to facilitate this. All this will, of course, be layered upon the range of information that we already incorporate into formulation of a therapeutic plan—age, existing co-morbidities, life expectancy, personal values, etc. The future is thus both exciting and challenging. There will be plenty to keep us engaged in our field, and busy!
- Schneider BP, Li L, Miller K, et al. Genetic associations with taxane-induced neuropathy by genome-wide association study (GWAS) in E5103. J Clin Oncol. 2011. 29;15:suppl 1000.
- Best T, Li D, Skol AD, et al. Variants at 6q21 implicate PRDM1 in the etiology of therapy-induced second malignancies after Hodgkin’s lymphoma. Nat Med. 2011. 17;8:941-3.