Population-based cancer pharmacosurveillance is an emerging area of interest and there is a need for proof-of-principle projects to demonstrate the feasibility and potential value of a comprehensive data system for deploying pharmacogenomics knowledge in population research. Hypotheses about clinical and molecular predictors of doxorubicin treatment choice, toxicity and response in patients with lymphoma provide a useful test case.Non-Hodgkin lymphoma (NHL) ranks 5th in cancer incidence and cancer-related deaths, and is the second fastest growing cancer. In addition, NHL is the most commonly occurring hematological cancer, and tends to affect older adults (61% were age > 60 years). For patients with diffuse large B-cell lymphoma, there is usually a compelling rationale for anthracycline or anthracenedione based combination (ABC) chemotherapy, however, in the community less than half of older lymphoma patients receive it. Even when ABC chemotherapy is prescribed to older adults, it is commonly initiated at an attenuated dose1. Older lymphoma patients frequently have adverse prognostic factors at the time of treatment decision-making. It is likely that concerns about doxorubicin-induced cardiotoxicity and myelotoxicity are partially responsible for observed low rates of ABC chemotherapy among elderly lymphoma patients in clinical practice and particularly among those with a history of cardiac disease or anemia. One such factor that is likely prominent, but difficult to measure from retrospective data sets is physician/patient perceived concern for frailty , or functional disability in the elderly. The net impact of dose-reduction to avoid treatment toxicities is not known. It is unknown, for example, whether the risk of cardiac toxicity is greater than the increased risk of death from lymphoma in patients who don't receive adequate ABC chemotherapy because of asymptomatic cardiac disease or asymptomatic reductions in cardiac ejection fraction. Provocatively, we found no significant improvement in survival among responders to CHOP chemotherapy age 75 and over who received 6 or more cycles of CHOP compared to those who stopped chemotherapy early, i.e. who received less than 6 cycles. In contrast, among younger chemotherapy responders (age 60-74), early cessation was associated with worse survival2.Furthermore, dose-reductions to avoid neutropenic events (FN) would seem unjustifiable if the result is an increase in deaths from lymphoma, especially since there are known effective methods for preventing FN. In the Oncology Practice Patterns Study (OPPS), we studied the cycle-specific incidence of FN and distribution of first FN cases by cycle. The cycle in which the first hospitalization occurred was similar for older and younger patients (p=0.71)3. It is quite clear that FN risk is concentrated in the first cycle. FN prediction models that require information based on abosulte neutrophil count (ANC) response to cycle 1 therapy are thus not apt. The balance of lower toxicity vs. possible reduced survival benefit from dose-reduction is unknown. The incorporation of pharmacogenomics into the drug utilization process holds great promise in the ability to individualize drug therapy. Such findings may serve to identify patients who will receive antitumor benefit from doxorubicin without experiencing severe toxicity. Substantial progress has been made in identifying important single nucleotide polymorphisms (SNPs) in genes thought to be involved in the doxorubicin metabolic pathway. Doxorubicin has a narrow therapeutic index (i.e. the dose prescribed is close to the dose likely to produce toxicity in most individuals). Substantial interpatient differences in drug disposition are known to exist, are unrelated to body surface area (casting doubt on the use of body surface area as a key parameter for dose calculation), and may result in differences in clinical response and toxicity. The doxorubicin pathway consists of detoxification and DNA repair mechanisms. Polymorphisms in these genes may result in diminished repair, reduced inactivation, or decreased activity of the anthracycline compound. In addition, polymorphisms in transporter genes may result in increased penetration of the blood-brain barrier and increased neurocognitive effects. Knowledge about clinical and molecular predictors of doxorubicin treatment toxicity and response will provide new approaches to the clinical management of lymphoma patients in community practice.
