Serum Test IDs Lung Cancer in Never Smokers

April 21, 2011
Anne Landry
Anne Landry

A team of researchers has used mass spectrometry to identify a novel six-biomarker serum test that effectively identified lung cancer in never smokers, and which may have other important diagnostic applications in lung cancer.

A team of researchers has used mass spectrometry to identify a novel six-biomarker serum test that effectively identified lung cancer in never smokers, and which may have other important diagnostic applications in lung cancer.

The team is from Celera, located in Alameda, California. Lead investigator Charles E. Birse, PhD, associate director of product development at Celera, presented the findings at the 102nd annual meeting of the American Association for Cancer Research, held April 2 through 6 in Orlando, Florida (abstract 2813).

About 20% of lung cancer cases worldwide occur in people who have never smoked tobacco, and the proportion of such cases is expected to increase in the future, as smoking-cessation initiatives help to reduce lung cancer deaths among smokers, Dr. Birse said. Adding to the potential complexity of its management, lung cancer in never smokers appears to be a different disease, compared with lung cancer in smokers, based on clinical, epidemiological, and molecular variations reported to date. (For example, lung cancer in never smokers occurs disproportionately among women and is commonly adenocarcinoma, with bronchoalveolar features.)

To develop the biomarker panel used in the study reported at AACR, Dr. Birse and colleagues drew on data from an earlier biomarker study in which they had used proteomic analyses with mass spectrometry to identify candidate biomarkers differentially expressed in lung tumor tissues and lung cancer cell lines, focusing on the cell surface and secreted proteins. More than 500 candidate markers were identified from the earlier study. The investigators prioritized the markers and included a number of previously characterized lung cancer markers, to optimize the biomarker panel.

Nine biomarkers were then validated by immunoassay of serum collected from smokers with non–small-cell lung cancer (NSCLC) and from an appropriately matched control group. The nine biomarkers were identified from testing of more than 600 serum specimens from four independent case-controlled smoker studies. The samples were randomly divided into a training set and a testing set, each with NSCLC cases and controls, to develop a regression-based algorithm for lung cancer detection. Ultimately, a global six-biomarker regression model was developed on the training set and tested on the validation set.

The six-marker regression model was effective at identifying smoking-associated cancer cases (AUC = 0.87; 269 cases, 360 controls). All stages of cancer and all of the major histological cell types were distinguished.

When the six-marker model was applied in an independent validation study to a never smoker group of 40 subjects with lung cancer (50% with adenocarcinomas) and 40 age- and gender-matched controls, it again showed strong sensitivity and specificity in discriminating the malignant cases (AUC = 0.91; 85% sensitivity and 83% specificity).

Commenting on the study, Dr. Birse emphasized that, while effective imaging modalities are central to earlier detection and better outcomes in lung cancer, improvements in diagnosis and management in this setting are clearly needed. For example, he said, while preliminary data from the National Lung Screening Trial (NLST) show that CT scanning reduces lung cancer mortality by 20% compared with conventional chest X-rays, “concerns still remain regarding the low specificity of CT scanning [about 75%] and the morbidity associated with the necessary diagnostic follow-up, be it through needle biopsy, bronchoscopy, or surgery.”

Although the current findings need to be confirmed by larger studies in never smokers, Dr. Birse noted that biomarkers for NSCLC detection among both smokers and nonsmokers “may serve as a valuable adjunctive tool to CT scanning-either pre-CT, to serve as a filter identifying the highest-risk individuals, with positive cases being confirmed by follow-up imaging, or post-CT, after a nodule has been identified in the lung, to act as a rapid, noninvasive method to confirm the malignant status of the nodule.”