Lung cancer--potential biomarkers

http://www.ncbi.nlm.nih.gov/pubmed/21206385

J Thorac Oncol. 2011 Jan 4. [Epub ahead of print]
Pharmacogenetic and Germline Prognostic Markers of Lung Cancer.
Horgan AM, Yang B, Azad AK, Amir E, John T, Cescon DW, Wheatley-Price P, Hung RJ, Shepherd FA, Liu G.

Department of Medical Oncology, Princess Margaret Hospital, Toronto, Ontario, Canada.
Abstract
INTRODUCTION: Lung cancer is the leading global cause of cancer-related mortality. Interindividual variability in treatment response and cancer outcomes has focused attention on genetic polymorphisms as prognostic markers. We evaluated the overall contribution of candidate polymorphism association studies to our current understanding of the genetic predictors of lung cancer outcomes.

METHODS: We examined the results of 90 studies that evaluated associations between genetic polymorphisms and lung cancer outcomes published between January 1990 and May 2009.

RESULTS: A total of 170 genetic variations in 90 studies were identified. Overall survival was a primary outcome in 81% of the studies and toxicity in 19%. Candidate polymorphisms in the DNA repair/synthesis pathway were the most frequently studied. Strong evidence in large-scale confirmatory studies of any single polymorphism was lacking. Polymorphisms of EGFR, XRCC1, and ERCC1 were associated with pharmacogenetic outcomes, whereas polymorphisms of MDM2, p53, and GSTM1 were associated with prognostic outcomes. All remaining polymorphisms had results lacking or failing replication testing. Heterogeneity in study populations, incomplete reporting of important population or study characteristics, inadequate power, and inconsistencies in methodology were common.

CONCLUSIONS: Although the quality of existing studies involving the candidate polymorphism approach is highly variable, a small set of candidate polymorphisms was identified as potential biomarkers of clinical or pharmacogenetic outcome and would benefit from further replication testing. Newer approaches including haplotype tagging, pathway, genome-wide association, and combination methods with validative approaches may facilitate a more accurate prediction of lung cancer outcomes by genetic variation.