Gene expression and lung cancer prognosis

While there have been significant advances in the use of gene expression profiling to assess a cancer prognosis, a Mayo Clinic review and analysis of existing lung cancer studies shows that this technology has not yet surpassed the accuracy of conventional methods used to assess survival in lung cancer patients.

The interest in and the knowledge of gene expression profiling in medical science has exploded since the completion of the human genome project in 2003. Researchers caution that the science of gene expression profiling, in which scientists examine the genetic signature of a cell, is in its infancy, particularly in lung cancer.

"Growing evidence suggests that gene-based prediction is not stable and little is known about the prediction power of a gene expression profile as compared to well-known clinical and pathologic predictors," according to Ping Yang, M.D., Ph.D., the corresponding author of the study that appears in the November issue of Cancer Epidemiology, Biomarkers and Prevention (CEBP). The study's first author is Zhifu Sun, M.D., a research associate with the Department of Health Sciences Research at Mayo Clinic.

Dr. Yang, a researcher with Mayo Clinic's Department of Health Sciences, said that while gene expression profiling has been successfully used to classify various tumors and assess tumor stage, metastasis and patient survival rates, the evidence suggests that gene-based prediction for lung cancer is not yet entirely dependable. However, some results have been promising: gene profiling has reliably predicted patient survival for lung adenocarcinoma almost as well as established predictors.

The results of conventional methods that factor in age, gender, stage, cell type and tumor grade outweigh the predictive advantage of a gene expression profile. "Any new technique that does not significantly outperform less expensive and easily conducted approaches is less likely to be useful in clinical practice," the authors wrote.

Few studies have compared conventional methods of lung cancer prediction with gene profiling. It remains to be seen whether gene expression profiling of lung cancer cases can replace or augment the existing assessment tools and, furthermore, whether it can lead to improved patient care.

In terms of problems associated with gene expression profiling in lung cancer research, the authors found:

• The accuracy of gene expression-based outcome prediction varies greatly among studies.
• Most studies lacked independent validation.
• Clinical outcome prediction between gene expression profiles and pathological features overlap significantly.
• Current analytical algorithms favor genes at high expression or genes highly differentially expressed, most of which are related to tumor differentiation and may not correlate with clinical outcomes; conversely, genes expressed at low levels or in a subtle difference are often overlooked, which may be quite relevant biologically to clinical questions.

The authors of the study recommend that medical scientists engaged in gene expression profiling should:

• Clearly define a study aim. The main focus in microarray studies should explore the molecular explanations for varied clinical outcomes given a group of patients with similar clinical and pathological characteristics.
• Lay out and compare alternative study designs
• Carefully select samples in terms of size, quality and unambiguous clinical outcomes
• Understand the limitations of DNA microarray
• Provide clinically relevant interpretation from the study results and address the value added in practice