RNA signature predicted lung cancer in asymptomatic smokers

The results of a small pilot study indicated that transcriptional changes in peripheral blood predicted which smokers would eventually develop lung cancer.

Thomas Zander, MD, of the Center for Integrated Oncology at University Clinic Cologne, Germany, presented the results at the 2008 ASCO Annual meeting in Chicago.

Zander and colleagues concluded that the prediction of prevalent lung cancer is possible using a lung cancer–specific gene expression. The test produced a sensitivity of 90%, specificity of 85% and accuracy of 88%.

The researchers examined peripheral blood from 13 patients with prevalent lung cancer and 11 control patients and extracted blood-based gene expression profile specific to patients with lung cancer. They validated the expression profile against a second, independent cohort of 22 patients with prevalent lung cancer and 15 control patients.

The researchers then examined the blood of smokers that had been collected in the European Prospective Investigation into Cancer and Nutrition study. Those patients (n=12) then developed cancer two years after collection.

The researchers compared the blood samples with those of healthy people, defined as individuals who had not experienced cancer within a median of 8.6 years. The signature predicted the future onset of lung cancer and produced a sensitivity of 75%, a specificity of 85% and an accuracy of 80%.

When applying the test in the incident lung cancer cohort, the researchers predicted the clinical manifestation of lung cancer (P<.05). The researchers found similar results (OR=17; 95% CI, 2.2-121) when using identical test criteria on data generated on a different technical platform. – by Jason Harris

Zander T. #1509. Presented at: 2008 ASCO Annual Meeting; May 30-June 2, 2008; Chicago.

These are incredibly small sample sizes; this really is a pilot study. The gene expression signature identified is interesting in that it identified genes associated with homeostasis and gene function instead of the genes we more typically think about in terms of cancer. However, that is a function of looking at peripheral blood mononuclear cells. Prospective validation using larger numbers is important, and the major question is: Can we couple this to other detection methods to try to actually come up with a system that will result in not only early detection but also a reduction in mortality?

– Gary Gordon, MD

Vice President for Global Oncology Development,
Abbott Laboratories, Abbott Park, Ill.