With a projected 55,170 deaths in the United States in 2006, colorectal cancer is second only to lung cancer as the most deadly. Patients with operable, but somewhat advanced, disease (stages II-III) are either observed only or treated with adjuvant therapy to reduce the chances of recurrence after surgery. Although chemotherapy is standard in stage II-III rectal and stage III colon cancer patients, the benefits are restricted to a relatively small fraction of patients. Adjuvant treatment of stage II colon cancer patients is particularly controversial due to the very small benefit in this group. Unfortunately, the prognostic power of simple clinicopathologic data is insufficient to identify high-risk stage II colon cancer patients--or low-risk rectal or stage III colon cancer patients. This proposal describes development of an individualized test based on the status of multiple molecular markers in the patient's tumor to provide an accurate prediction of the risk of recurrence. The molecular markers were carefully selected to provide broad coverage over a range of biological pathways with established prognostic potential. They will be analyzed by immunohistochemistry (IHC) in >600 stage II colorectal cancer patients. Pathologists will score up to 22 markers, as will an automated digital image analysis system for maximal objectivity. Statistical pattern recognition methods will then be used on the data to select a subset of the markers and clinicopathologic features that, when combined with an optimized machine learning algorithm, will produce an accurate model that predicts recurrence. Measuring proteins directly in tumor sections with IHC is arguably one the most relevant strategies available to assess patient prognosis and is widely accepted in the pathology community. Unlike IHC, competing methods like gene expression cannot directly assess protein levels or localization, and it is very difficult or impossible to distinguish normal from tumor cells in the homogenized tissue mixtures that are assayed. Furthermore, unlike most previous related studies, the informatic analyses proposed here include robust, powerful methods to discover complex interactions between markers--maximizing performance--while avoiding overfitting--maximizing the potential for generalizability. The proposed application of advanced quantitative scoring (for objective analysis) and cutting-edge informatics (for complex marker interaction elucidation) to the tried and true IHC protein staining technology is a highly innovative approach to the development of cancer prognostics. The ultimate product will consist of a panel of assays for the relevant molecular markers and the associated prognostic algorithm that predicts recurrence. Colorectal cancer patients and their oncologists will be able to use the results to guide their treatment decisions, resulting in positive impacts on life expectancy and quality of life. With the help of their oncologist, patients with newly diagnosed stage I-III colorectal cancer must decide whether to endure the potentially serious side effects of chemotherapy after surgery in the hopes of preventing a recurrence. However, the benefit of chemotherapy is restricted to a relatively small fraction of these patients, and the predictive ability of current clinical guidelines is relatively crude. The goal of this study is to develop an individualized test that will more accurately predict risk of recurrence based on advanced mathematical analysis of the status of multiple molecular markers within the patient's tumor. It should lead to improvements in both survival and quality of life for colorectal cancer patients. ? ? ?

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Small Business Innovation Research Grants (SBIR) - Phase II (R44)
Project #
2R44CA117001-02
Application #
7272249
Study Section
Special Emphasis Panel (ZRG1-ONC-L (12))
Program Officer
Tricoli, James
Project Start
2007-08-01
Project End
2009-07-31
Budget Start
2007-08-01
Budget End
2008-07-31
Support Year
2
Fiscal Year
2007
Total Cost
$882,531
Indirect Cost
Name
Prediction Sciences, LLC
Department
Type
DUNS #
364233523
City
La Jolla
State
CA
Country
United States
Zip Code
92037
Krajewska, Maryla; Smith, Layton H; Rong, Juan et al. (2009) Image analysis algorithms for immunohistochemical assessment of cell death events and fibrosis in tissue sections. J Histochem Cytochem 57:649-63