Length and lead time biases in radiologic screening for lung cancer.

摘要

OBJECTIVE: Our goal was to investigate whether the length and lead-time biases of radiologic screening for lung cancers vary according to the histologic type of the tumor. METHODS: We analyzed the survival rates and radiographs of 119 cases of adenocarcinomas-large-cell carcinomas (ALC) and 50 peripheral squamous cell carcinomas (PSQ) detected in 205,401 screened individuals. RESULTS: The sensitivity of screening and 5-year survival rates were 84.0% and 48.2% for ALC, and 52.0% and 18.5% for PSQ, respectively. The corrected length bias was 4.3% for ALC and 4. 6% for PSQ. Stage III-IV ALC was often identified on 1-year-old films, but stage III-IV PSQ was not. Half of stage I ALC presented 2 or more years before detection, while half of stage I PSQ appeared within 1 year before detection. The survival rate of nonresected cases with stage I ALC was decreased 4 years after detection, while that of nonresected cases with stage I PSQ was decreased just after detection. The period of stage I ALC and PSQ was at least 6 years and 1 year, respectively. CONCLUSIONS: Slowly growing ALC had high sensitivity in radiologic screening and a high rate in 5-year survival, but had long lead time and delay in detection. PSQ grew rapidly resulting in low sensitivity in radiologic screening, and short lead time and survival. In both types, the magnitude of corrected length bias was not remarkable. The survival of ALC should be carefully evaluated because of the long lead time. Copyright Copyright 1999 S. Karger AG, Basel