Amid slings and arrows, ELCAP forges ahead

In almost eight years of screening a high-risk population for lung cancer, Dr. Claudia I. Henschke has learned a thing or two. One is how to survive her critics.

Henschke presented new preliminary data and revised parameters regarding the Early Lung Cancer Action Project (ELCAP) during a special lung cancer screening session at the European Congress of Radiology.

Three experts who shared the dais with Henschke did not share her enthusiasm for nonrandomized screening trials. Although cautiously applauding her work, they all agreed that only randomized trials can determine if CT screening for lung cancer saves lives.

Henschke's methods include screening everyone with low-dose CT at baseline to determine lung cancer distribution by size and stage. In contrast to most trials, ELCAP randomizes patients after the initial screening to either treatment or no treatment.

One of the most important conclusions of ELCAP, according to Henschke, is that baseline results are very different from annual repeat findings. Cancers found at baseline are typically slower growing than those found on annual repeat exams.

Because of this finding, Henschke has reconsidered the definition of a positive screening result, which leads to further workup. She has also redefined the workup algorithm. Rather than work up every nodule found at baseline, ELCAP will ignore those less than 5 mm in diameter in the first year.

"This will reduce the positive results of screening to about 11% to 15%. And there's no increase in false negatives as well," she said.

ELCAP's initial results from 1000 screens reported in 1999 and those from nearly 2000 subsequent screens have shown that no lesions below 5 mm found at baseline turned malignant in the first year of follow-up.

Henschke has never included calcified nodules in the definition of a positive result, but she's now suggesting a size criteria change. As more data come in, researchers will use other characteristics to further reduce the number of positive results.

Henschke is also revising parameters based on new lesion morphology data that have come to light because of CT technology advances. Besides the commonly known solid nodules, ELCAP has identified nonsolid nodules, in which the only solid components are the vessels within them, and part-solid nodules. The part-solid nodules have three times the probability of malignancy compared with the solid or nonsolid nodules, she said.

Henschke has devised a new protocol based on morphology data from more than 10,000 screens (including the international ELCAP). Any solid or part-solid noncalcified lesion equal to or greater than 5 mm in diameter and any nonsolid nodule greater than 10 mm should be considered a positive result and requires further workup at baseline.

Henschke stressed that these changes apply only to baseline. For annual repeat exams, any growing noncalcified nodule is considered a positive finding and requires workup. That protocol remains unchanged.

ELCAP is starting randomized treatment trials for certain classes of lesions. The preliminary results in terms of reduction of deaths from lung cancer look promising, Henschke said.

Low-dose CT screening for lung cancer is very cost-effective, she said, costing less than $2500 per life year saved. This was shown with the initial approach in 1993 and continues to be validated. Reasons for its cost-effectiveness include a well-identified population subset and the relative low cost of a CT exam ($300 in ELCAP). Treating early-stage cancer costs about half as much as treating late-stage cancer.