DUBLINThe shift from histologic classification to molecular analysis of bronchial epithelial cells is opening up new prospects for detecting the molecular signature of lung cancer before the disease becomes clinically evident, James L. Mulshine, MD, of the National Cancer Institute, said at the 8th World Conference on Lung Cancer.
Our target is often palliative, but wed like to start thinking more about cure and more about returning patients with lung cancer to health so that they can resume their normal course of employment and activities, Dr. Mulshine said. With early detection and management of preinvasive cancer, I think that these are achievable goals at some point in the future.
Dr. Mulshine said that the area at risk of cancer encompasses all cells of the upper aerodigestive epithelium that are chronically washed by inhaled carcinogens. Although tobacco combustion products lead to many types of cellular injury, a clinically significant lung cancer arises only from transformation events in the bronchial epithelial cells. The challenge, therefore, is to distinguish between background cellular injuries and the specific epithelial injuries that will culminate in cancer.
In this regard, Dr. Mulshine noted, sputum morphology by itself is not informative, since molecular changes precede morphological changes. Currently, we have a number of candidate markers that we know relate to some aspect of lung cancer biology, but the important thing we have to do is determine which of these markers relate to the progression to cancer, he said.
He cautioned that the presence of an oncogene mutation or a chromosomal deletion does not necessarily mean that cancer will develop, and stressed that the identification of early detection markers can only be accomplished through clinical validation.
With hundreds of millions of people at high risk for lung cancer worldwide, Dr. Mulshine said, early detection of molecular markers has broad implications. However, he pointed out, such detection will only be clinically meaningful if it can be done cheaply and efficiently. To this end, a number of research groups are developing automated devices that will allow high-through-put, gene-based diagnostic assays.
Stopping Preclinical Disease
The prospect of routine early detection raises the obvious question of whether it will be possible to arrest preclinical lung cancer (see figure).
One of the important directions that we think will be useful will be to use the same route of administration for delivering chemoprevention drugs as was used by the original carcinogen, Dr. Mulshine proposed.
With state-of-the-art aerosolized delivery systems, he said, some 40% to 70% of the inhaled dose can be delivered to the target tissues. This technology would be relevant for whatever compounds we think are useful in blocking the early stages of carcinogenesis, he said.
He further suggested that the same molecular markers applied in the detection of early epithelial transformation should prove useful as intermediate endpoints in evaluating the efficacy of various chemoprevention agents.
One of the more promising molecular markers of lung cancer is hnRNP A2B1, a protein involved in the packaging and transport of messenger RNA. In an eight-year screening study conducted at Johns Hopkins University, annual collection of sputum samples from 5,000 healthy smokers showed that the hnRNP A2B1 antigen was the single best predictor of future development of lung cancer.
Johns Hopkins researcher Melvyn S. Tockman, MD, PhD, told the Lung Cancer Conference that regulated, granular overexpression of hnRNP A2B1 occurs in proliferating lung tissue, whereas dysregulated, uncontrolled, diffuse overexpression of this protein is a characteristic of neoplastic lung epithelium.
In fact, he suggested, if the progression from regulated to dysregulated overexpression of hnRNP A2B1 signals transformation to preneoplastic epithelium, then monitoring the expression of this antigen in sputum samples could markedly boost the accuracy of preclinical lung cancer detection.
Two major trials are currently underway to prospectively compare the utility of hnRNP A2B1 overexpression with that of routine sputum morphology in forecasting the advent of lung cancer.
Dr. Tockman reported first-year results from the 11-center Lung Cancer Early Detection Working Group study of nearly 600 patients with resected stage I non-small-cell lung cancer (NSCLC).
The researchers found that hnRNP A2B1 expression was highest in patients who went on to develop a second primary lung cancer; expression was intermediate in those who eventually experienced a recurrence, and it was lowest in those who remained healthy. We found 13 cases that developed a second primary lung cancer and 16 recurrences, and out of these, none of the cytology specimens showed frank malignancy, he said.
The sensitivity of hnRNP A2B1 staining in detecting a second primary tumor was 77%, and specificity was 82%. This translates into a positive predictive value of 67%, he said. That means that if the sputum stains positive, an individual has a 67% risk of developing lung cancer within one year, instead of a background risk of 2.2%.
Similar findings have recently come to light in the Hopkins study of Yunnan Tin Company miners who are at high risk of primary lung cancer by virtue of their exposure to tobacco and underground radon and arsenic. Dr. Tockman noted that hnRNP A2B1 expression was strikingly higher among miners with primary lung cancer than in a matched cohort without cancer.
Here the sensitivity of hnRNP A2B1 staining for primary lung cancer was 82%, specificity 65%, and positive predictive value 68%, as compared with a one-year background risk of 0.9%.