BRCA1 and BRCA2 Pathways and the Risk of Cancers Other Than Breast or Ovarian

Methods

Many PubMed searches were conducted. To search for possible effects between BRCA mutations and sporadic cancers, individual cancers were entered so the search read "BRCA1 and melanoma" and then "BRCA2 and melanoma." Initially, the term "colon cancer and BRCA" was entered. Then all possible variants were entered such as "colorectal cancer and BRCA1." An additional strategy was "hand searching" by consulting appropriate references in literature that seemed pertinent. A few searches of EMBASE were also done. Additional searches were conducted for incidence of cancers in high-risk groups, for second cancers after diagnosis of primary breast or ovarian cancer at young ages, and for second cancers after diagnosis of male breast cancer.

More than 30 observational studies going back 20 years and including case histories from many thousands of mutation carriers or possible mutation carriers in high-risk populations were collected and intensively reviewed. Collection of data representing a large number of mutation carriers was achieved by including several types of studies as follows. Population studies were used that tabulated cancer histories for known mutation carriers and for their first-degree relatives. Large populations enriched in mutation carriers, such as those eligible for BRCA gene mutation testing, were also included. A third type of study measured incidences of second cancers after breast cancer in young women. The likely number of BRCA mutations was then estimated from data in the publication or from mutation incidence tables available at www.myriadgenetics.com . I also included studies that measured whether the frequency of BRCA gene mutations was elevated in patients presenting with certain cancers other than breast or ovarian.

The literature results were used directly as reported by the original authors. This enabled inclusion of data that were 10 to 20 years old in addition to the most recent studies. The goals for data analysis were to use as little additional statistical manipulation as possible such that it would allow anyone to reproduce the results. The summary method of meta-analysis using the StatsDirect Computer Program was used for statistical calculations. StatsDirect calculates stratum weights as the inverse of the variance for the summary statistic (Y) supplied. The pooled estimate of Y is calculated as a weighted mean, ie, the sum of weighted Y for each stratum divided by the sum of the weights. Meta-analyses based on approximate relative risk calculations were also run as a double-check using available data in the original publications.

Studies were excluded in whole or in part if they did not present relative risks and confidence intervals or provide enough data to calculate missing information. Because of the close interactions of Fallopian tubes and ovaries and because data show that Fallopian tubes are highly susceptible in BRCA mutation carriers, Fallopian tube cancer was considered to be a part of the breast/ovarian cancer syndrome.

Data from the studies included sample size, population, method, control population, method of expressing results (as relative risks, percentages, odds ratios, etc), and 95% confidence interval.

I found no evidence for misdiagnosis of cancer, and most studies reported histologic verification in a high percentage of cases. It was necessary to make the assumption that cancers were diagnosed correctly.

© 2005  Medscape

Cite this: BRCA1 and BRCA2 Pathways and the Risk of Cancers Other Than Breast or Ovarian - Medscape - Jun 29, 2005.