Genomic prediction tools developed using phenotypes from disease progression models

Genomic prediction tools developed using phenotypes from disease progression models

The lead researcher, Manish Sharma, is currently supported by an institutional K12 grant that is funded by the NIH (PI: F. Olopade). This grant supports my salary, protects 75% of my time for research, and has additional research funds to partially support a research assistant. I have applied for a K23 career development award from NIH that would fund my salary for an additional 3-4 years and provide research funding to support a research assistant and the expenses associated with the proposed clinical trial. Dr. Ratain and Dr. Cox are co-principal investigators on a Pharmacogenomics of Anticancer Agents Research (PAAR) grant that will support the some of the computational and whole genome interrogation efforts proposed.

None

Kidney cancer is the twelfth most common cancer in the world, with 338,000 new cases diagnosed in 2012. Although many new drugs have recently become available for kidney cancer, we do not have any reliable way of predicting which patients are most likely to benefit from these drugs.

There is growing evidence that genetic factors (our DNA) impact the intended and unintended effects of drugs, explaining some of the variability between patients. There are published studies suggesting that changes in the DNA of specific genes can help to predict who is most likely to benefit from some of the new drugs for kidney cancer. We believe that additional links between genetics and drug effect can be achieved by two strategies: 1) developing better measures of drug effect; and 2) using all of the common changes in the DNA of patients (instead of one or two specific changes) to develop tools that predict drug effect.

Data from patients on two studies recently sponsored by GlaxoSmithKline (GSK) will be used in the proposed research. The drugs used in these studies are pazopanib and sunitinib. These studies were selected because two important items were collected in these studies: 1) DNA from a blood sample (on most patients); and 2) images from CT scans before and during therapy. The main goal of the research is to develop tools that can predict drug effect in individual patients based on common changes in their DNA. First, a computer program will be applied to the images from CT scans to measure the volume of tumors before and during therapy. Second, these volume measurements will be used to develop a mathematical model that describes the growth of cancer over time and the effect that a drug has to slow down or reverse this growth. Third, a computer algorithm will be used to identify the best tools that can predict the drug effect based on common changes in DNA. Finally, a clinical trial will be conducted to confirm that the tools work to predict drug effect in patients with kidney cancer who are being treated with these drugs in routine clinical practice.

Upon completion of this project, we will present our work at major national and/or international meetings that are attended by cancer and drug experts. We will also publish our work in widely read medical journals so that our results can positively impact the way physicians treat patients with kidney cancer. The successful completion of this project would not only potentially improve outcomes for patients with kidney cancer, but would also establish a new method that could be used to potentially improve outcomes for patients with other cancers and other diseases.

[{ "PostingID": 461, "Title": "GSK-VEG105192", "Description": "Medicine: pazopanib, Condition: Carcinoma, Renal Cell, Phase: 3, GSK Clinical Study ID: VEG105192, Sponsor: GSK." },{ "PostingID": 463, "Title": "GSK-VEG108844", "Description": "Medicine: pazopanib, Condition: Carcinoma, Renal Cell, Phase: 3, GSK Clinical Study ID: VEG108844, Sponsor: GSK." }]

The publication citation will be added after the research is published.