Simple New Test Aids in Fight Against Breast Cancer
The researchers developed what they have tentatively named the Pedigree Assessment Tool (PAT) by analyzing the results obtained from 3,906 women without a personal medical history of breast cancer who completed a questionnaire as part of a routine mammogram. The responses given by those who participated in the survey were then evaluated by accepted biostatistical methods and compared to the results predicted by 2 well-established models (the Frank and Gail models) that have been used to determine the risk of breast cancer occurring within 5 years and in a lifetime.
The PAT correctly predicted a high probability (= 10%) of the presence of BCRA mutations in 82 patients with a known family history of breast cancer. After factoring in differences in ethnicity, environment, and other related factors the PAT demonstrated an overall reliability of 96%, a sensitivity (the number of patients that actually had one or more BRCA mutations) of 100%, and a specificity (correctly predicting those without the BRCA mutations) of 93%.
In the “Discussion” section of their report the authors conceded that potential limitations of the study include a failure to predict susceptibility due to high penetrance genes other than BRCA1 and BRCA2; and the use of the Frank model as the “gold standard” against which the PAT and Gail model were judged. However the authors are optimistic that the PAT, most likely in combination with other risk assessment strategies, will be a valuable improvement in the efforts to reduce breast cancer deaths by quickly and reliably identifying those at highest risk.
“The PAT is a simple and accurate tool for identifying women at risk for the hereditary breast cancer syndromes that can be employed as part of a comprehensive breast cancer risk-screening strategy in the primary care setting … We believe that the concept of a comprehensive breast cancer risk-screening strategy applied to large populations at the primary care level warrants further investigation, and that such a strategy could be effectively employed by combining the Gail model with a tool like the PAT.”
Discussion
This study reports on a new technique for predicting who will be at an increased risk for development of breast cancer by reliably determining which women are most likely to be carrying one of the two mutations in the BCRA gene that have been implicated as strongly indicative of future cancers. While this will be a welcome addition to current prediction and screening models it is important to realize that this study, like any other medical test, is based on the principle of statistical inference (using data obtained from a small sample to make accurate predictions about a much larger population). The classic example of statistical inference is the “failed battery” example.
A manufacturer has an assembly line that constructs automobile batteries. The manufacturer tests these batteries on a purely random basis and has determined that 0.5% of all batteries will fail when tested. It can be inferred from the previous results that, out of 1000 batteries produced, 5 will not perform adequately when tested. Notice that while 5 batteries will be defective, the testing doesn’t predict which 5 will fail.
In medical research a test that detects all cases of a certain disease is said to be 100% sensitive for that disease and a test that detects only those cases where the disease is actually present (no false positive events) is said to be 100% specific. I can personally assure you that there has been no test in all the history of medicine that has been both 100% sensitive and 100% specific.
The point of the above discussion is to remind the reader that there will always be some margin of uncertainty involved in any aspect of medical testing and/or diagnosis and the reader is encouraged to discuss such margins of errors with their health care provider.
Online Tools
The National Cancer Institute maintains an interactive risk assessment tool based on the Frank Model as part of its Breast Cancer Risk Assessment program.
References
Bondy, ML and Newman, LA. Assessing Breast Cancer Risk: Evolution of the Gail Model. Journal of the National Cancer Institute 2006; Vol. 98, No. 17, 1172-1173.
Hoskins, KF, Zwaagstra, A, and Ranz, M. Validation of a tool for identifying women at high risk for hereditary breast cancer in population-based screening. Cancer (In Press).
Disclaimer
The information presented in this article and its included links is of an informational nature only and is not intended as a recommendation of any changes in the reader’s health care program. Before making any changes in diet, medications, or other treatments the reader is strongly advised to consult with their health care provider.
