OBJECTIVES: Major innovations in radiotherapy (RT) delivery (3D conformal RT, intensity modulated RT) now permit RT dose escalation to be tested as a means of improving control of localized prostate cancer. With delivery innovations, severe toxicity occurs rarely, but significant bladder and rectal toxicity is common. The investigators have studied a cohort of 98 patients who received dose-escalated 3D-CRT and have obtained evidence of dosimetric factors underlying rectal toxicity. They posit that the late radiation toxicity disease state has significant genetic determinants. These determinants are neither understood nor accounted for in selection of treatment, and the investigators propose to study the above well-characterized cohort, who are clinically well from a disease control perspective, given that comprehensive dosimetric and outcome information is available on all.

HYPOTHESIS: Comprehensive analysis of the genomic, proteomic, and dosimetric characteristics of patients with prostate cancer will provide a novel understanding of the genetic predisposition of cellular and tissue responses to irradiation which result in the clinical occurrence of significant chronic rectal and bladder injury after 3D-CRT.

APPROACH: For a thorough understanding of the molecular processes underlying tissue responses to radiation damage, the investigators propose a genomic analysis. Their working hypothesis is that rectal and bladder toxicity will be correlated to patient genetics as measured by single nucleotide polymorphisms (SNPs) in a select group of genes. The criteria for selecting SNPs will be based on evidence for the various genes implicated or demonstrated in DNA repair pathways and radiation-induced tissue damage. In addition, since radiation injury can be considered a phenotypic tissue response to RT in genetically predisposed individuals, the investigators will examine the protein profile "signature" in sera from patients who have suffered significant toxicity. Analysis of these data will be unique in that the investigators will use both statistically based bioinformatics approaches, and biophysical modeling. This proposal represents the most comprehensive exploration to date of the concept of a molecular "signature" of normal tissue radiation injury of which the applicants are aware.

SPECIFIC AIM 1:

• To create a database of SNPs among long-term prostate cancer survivors who have been treated with 3D-CRT, and for whom detailed toxicity, quality of life and dosimetric information are available;
• To assess the feasibility of using bioinformatics approaches with this database to examine the possibility that specific SNPs, in the context of known dosimetric factors, can predict occurrence of late rectal or bladder toxicity;
• To perform a proteomic analysis of serum samples from a cohort of 25 patients experiencing higher levels of ongoing late toxicity and 50 patients with no ongoing late toxicity.

SPECIFIC AIM 2:

• To utilize biophysical approaches to modelling clinical complication data, using the concept of generalized mean biological effective dose (GMBED) and the critical volume model.

SUMMARY: This proposal will be the first comprehensive retrospective analysis of genomic, proteomic and dosimetric determinants of late radiation injury in prostate cancer. The investigators have assembled a broad array of expertise from the CCI Polyomx program (the only high throughput SNP facility in western Canada), the NCI/FDA Clinical Proteomics Program at Bethesda, MD, and the unique biophysical modeling capability of CBIAR researchers to reach these goals.