Serum tumor markers CEA, CA 50, TATI, and NSE in lung cancer screening.

BACKGROUND

There are no effective means for screening for lung cancer, so the authors assessed the utility of four lung cancer tumor makers for screening.

METHODS

A case-control study, nested in a cohort study based on the linkage of records of health survey examinees with Finnish Cancer Registry records, was used to test the validity of tumor markers carcinoembryonic antigen (CEA), tumor-associated trypsin inhibitor (TATI), neuron-specific enolase (NSE), and CA 50 in lung cancer screening. Ten years after health examinations, record linkage indicated that 187 men had lung cancer; 344 control subjects, matched for age, sex, and municipality were drawn from the same records.

RESULTS

The data allowed assessment of the sensitivity of the marker assays at a 95% specificity level, which was highest for CEA (17% at a concentration level of 5.3 micrograms/l). Logistic discrimination analysis indicated that of the other markers, only TATI, when used in combination, improved the discriminatory power of CEA. CEA and TATI levels correlated significantly with smoking. They also showed a significant gradient toward increasing risk of lung cancer from the lowest to the highest quintiles of marker levels (for CEA, crude relative risk between the highest and lowest quintiles, 8.6). The gradient also was evident in the subgroup whose cancer had been diagnosed more than 5 years after serum specimen collection. The trend persisted, although relative risk was halved after adjustment for smoking.

CONCLUSIONS

The markers do not seem to be useful tools for lung cancer screening. However, CEA and TATI levels seem to give information on cancer risk long before the clinical cancer stage, as the quintile-based analyses of marker levels indicate.