Good follow-up in lung trials may speed cancer diagnosis

January 22, 2008

Comparison of two lung cancer screening studies finds better outcomes with well-planned workup

The lack of an effective, detailed workup algorithm for positive results in multicenter lung cancer screening trials results in a lower cancer yield from invasive procedures and later diagnoses for participants, according to researchers from the New York Early Lung Cancer Action Program.

Dr. David Yankelevitz, an investigator in the NY-ELCAP, and colleagues compared workup algorithms and results for the New York trial and the Lung Screening Study (LSS), which is the pilot study of the National Lung Screening Trial.

Participants in the two studies were at similar risk for lung cancer, had the same initial baseline test, and underwent an annual repeat test. In the NY-ELCAP, 92% of invasive procedures prompted by the baseline study were positive. Invasive procedures prompted by the annual repeat exam were also positive in 92% of cases. By comparison, in LSS participants, 57% of invasive procedures prompted by the baseline study were positive and 44% of those prompted by the annual repeat exam were positive.

At baseline in both studies, cancers were detected at larger sizes, as expected. The size of cancers detected at annual repeat differed markedly, however, as 55% were smaller than 10 mm in the NY-ELCAP study, whereas only 13% were smaller than 10 mm for LSS. The researchers also noted that the percentage of cancers detected at stage 1 was higher in the New York study relative to the LSS.

The better outcomes in the NY-ELCAP study were most likely due to differences in the workup algorithms, according to Yankelevitz, a professor of radiology at Weill Cornell Medical College.

"This study shows the benefit of workup algorithms in screening protocols. Screening is not just about the initial test. It includes the baseline screening test all the way through diagnostics, including pathology," he said in an interview with Diagnostic Imaging.

Dr. Denise R. Aberle, chief of thoracic imaging at the University of California, Los Angeles and principal investigator of the NLST, questions the validity of the comparative study.

"The LSS feasibility study was just that-a pilot trial designed to determine the feasibility of randomizing participants to receive either CT or chest x-ray screening, in preparation for the much larger ongoing randomized National Lung Screening Trial," she said in an e-mail to Diagnostic Imaging after the RSNA meeting. "The pilot was not designed or powered to address the efficacy of CT screening with respect to any endpoint and did not purport to analyze the relationships between screening result and recommended follow-up."

Many methodological factors in the studies could have influenced the outcomes reported-differences in the risk profiles of participants based on varying eligibility criteria; timing of follow-up screens and interventions; and the degree to which all interventions, complications, and outcomes were faithfully recorded-according to Aberle.

"All of these differences make cross-study comparisons unreliable," she said.

A variety of guidelines for diagnostic follow-up of positive screens have been developed by investigator groups, Aberle said. None of the published algorithms have been validated in head-to-head comparisons with alternative approaches. All are predicated on current best practices, the availability of and institutional preferences for specific services at screening institutions, and underlying assumptions about the biological behaviors of lung cancers.

"For all published studies in which the purported objectives were to assess some aspect of CT screening performance, algorithms have been provided as guidelines. Radiologists are not held to specific algorithms, but are encouraged to use clinical experience to optimize the care of participants with positive screens," Aberle said.