Correspondence

Microarray Analysis and Tumor Classification

N Engl J Med 2006; 355:960August 31, 2006

Article

To the Editor:

Quackenbush (June 8 issue)1 provides a comprehensive review of microarray technology within the framework of clinical application and also addresses some limitations of molecular analyses. A consideration not mentioned in the review is that the process of acquiring tissues for molecular analyses may result in important changes in gene expression, independent of disease biology. For example, we found that significant alterations in transcription factors occur simply as a result of surgical excision of the prostate.2 Those interpreting the results of molecular analyses of surgical samples should recognize that such alterations occur rapidly and may be associated with the method of tissue acquisition. Such changes in gene expression may be mistaken for markers of disease severity or response to treatment. These considerations are crucial for designing and analyzing molecular correlates of clinical studies, particularly for comparing samples obtained by disparate methods of collection or evaluating tissue samples obtained from different surgical centers, where the postsurgical processing methods may not be standardized. Optimal annotation of the samples should include methodologic factors (e.g., the duration of ischemia) that could alter gene-expression signatures.

Daniel W. Lin, M.D.
University of Washington, Seattle, WA 98195
dlin@u.washington.edu

Peter S. Nelson, M.D.
Fred Hutchinson Cancer Research Center, Seattle, WA 98109

2 References

1. 1

   Quackenbush J. Microarray analysis and tumor classification. N Engl J Med 2006;354:2463-2472
   Full Text | Web of Science | Medline

2. 2

   Lin DW, Coleman IM, Hawley S, et al. Influence of surgical manipulation on prostate gene expression: implications for molecular correlates of treatment effects and disease prognosis. J Clin Oncol (in press).
   

Author/Editor Response

Lin and Nelson point out an important fact: that transcriptional profiling with the use of microarray analysis or other techniques is sensitive to the manner in which the samples are collected and handled and that surgical manipulation can influence what we observe as gene-expression patterns. However, this point brings up a more fundamental issue. Although we discuss gene expression as if it were occurring in vivo, we are really measuring the representation of RNA in ex vivo tissues. This measurement is influenced by alterations in gene expression that are due to a variety of factors, including surgical manipulation, as well as to factors such as RNA degradation and the presence of contaminating tissue. Microarray projects should begin with an experimental design that minimizes the potential biases and confounding factors and includes protocols for the collection of high-quality samples. Successful projects involve close collaboration among surgeons, pathologists, clinical coordinators, bench biologists, and statistical and bioinformatics scientists to ensure the high quality of the samples, the data, and the analysis.

John Quackenbush, Ph.D.
Dana–Farber Cancer Institute, Boston, MA 02115
johnq@jimmy.harvard.edu

Citing Articles (1)

Citing Articles

1. 1

   Kai Tan, Jesper Tegner, Timothy Ravasi. (2008) Integrated approaches to uncovering transcription regulatory networks in mammalian cells. Genomics 91:3, 219-231
   CrossRef