Lung cancer mortality decreased by 26% in high-risk men and by as much as 61% in women who had screening chest CT scans, a large randomized European study called NELSON showed.
The trial showed a consistent reduction in lung cancer mortality of 24%-26% in screened versus unscreened men after 8-10 years of follow-up, whereas screening in the small group of women enrolled in the trial led to relative risk reductions of 40%-60%.
A difference in lung cancer mortality favoring the screened group became apparent within the first few years and persisted throughout the study, Harry J. de Koning, MD, PhD, of Erasmus Medical Center in Rotterdam, The Netherlands, reported here at the World Conference on Lung Cancer.
The study confirmed the benefits of lung cancer screening in high-risk individuals, first demonstrated in the landmark National Lung Screening Trial (NLST), and showed even larger reductions in lung cancer mortality.
“These results are more favorable than the NLST results and suggest gender differences,” said de Koning. “Volume CT lung cancer screening of high-risk former and current smokers results in low referral rates and a very substantial reduction in lung cancer mortality in both genders.”
The findings should have an immediate impact on existing screening programs and demonstration projects, said invited discussant John Fields, PhD, of the University of Liverpool in England. Current screening protocols should be reevaluated, as should existing recruitment strategies for men and women. Additionally, clinical trials should be integrated into lung cancer screening programs and include protocol-specified, personalized postsurgical interventions.
“There is now conclusive evidence for implementation of lung cancer screening based on two large randomized controlled trials,” said Fields.
Published in 2011, the NLST provided the first — and until now, the only — evidence that screening high-risk individuals can reduce the risk of dying of lung cancer. Involving more than 50,000 participants (59% men), the trial showed 20% reduction in lung cancer mortality in the trial participants randomized to annual screening with low-dose chest CT for 3 years.
Women appeared to derive more benefit from screening than men (relative risk reduction of 27% versus 8%), although a post-hoc analysis failed to show a significant difference in the benefit by sex (P=0.08).
De Koning said the sex difference is consistent with the natural history of lung cancer by histology and that CT screening in women might have increased the lead time for detection more so than in men.
Now Comes NELSON
Investigators in The Netherlands and Belgium designed NELSON as another randomized trial to investigate the potential lung cancer mortality benefits of screening with CT imaging in current and former smokers. Study participants were recruited by questionnaires from population registries. Eligibility included age 50-74, smoking history >10 cigarettes a day for >30 years or >15 cigarettes a day for >25 years, and smoking cessation within the past 10 years.
The randomized population consisted of 15,792 participants, 84% of whom were men. The screened and unscreened groups were balanced with respect to age, sex ratio, smoking history, and smoking cessation. Participants assigned to screening had chest CT imaging during the first and second years of the trial then at 4 and 6.5 years. Follow-up continued to year 10. Compliance with screening was 95.6% in year 1, 92.3% in year 2, 87.6% in year 5, and 66.8% with the last screen.
The trial protocol included centralized reading of CT image, monitoring of lung nodule volume and nodule doubling time, centralized adjudication of causes of death, and follow-up through national registries. Initially planned as a trial for high-risk men, the trial had statistical power to detect a 25% reduction in lung cancer mortality 10 years after randomization.
A total of 27,053 screening CT exams produced 2,503 (9.3%) indeterminate test results, 598 (2.2%) positive results, and 243 (0.9%) lung cancer detections. Comparing positive CT results with confirmed lung cancer detection, de Koning said the screening had a 41% positive predictive value.
Participants who had indeterminate screening results were told that an abnormality of uncertain significance had been identified and that they should return in 3 to 4 months for a follow-up CT scan to determine whether the abnormality had changed.
During follow-up to year 10, the screened group had about 400 lung cancers diagnosed versus about 350 in the control arm. Half the cancers in the screening arm were stage Ia at diagnosis, and 65%-70% were stages Ia-II. In contrast, about 70% of cancers in the control arm were stage III/IV at diagnosis.
De Koning reported that 214 patients in the control arm died of lung cancer, as compared with 147 in the screened group. The lung cancer mortality rate ratio for men in the screened versus unscreened arms was 0.75 at 8 years (95% CI 0.59-0.95, P=0.015), 0.76 at 9 years (95% CI 0.60-0.95, P=0.012), and 0.74 at 10 years (95% CI 0.60-0.91, P=0.003). Corresponding rates for women were 0.39 (95% CI 0.18-0.78, P=0.0037), 0.47 (95% CI 0.25-0.84, P=0.0069), and 0.61 (95% CI 0.35-1.04, P=0.0543).
The overall rate reduction for men was 0.74 (95% CI 0.60-0.91).
Roche Diagnostics and Perceptronix provided support for laboratory studies, and Siemens Germany provided workstations for uniform nodule assessment software. De Koning and co-investigators did not report any relationships with industry.
Primary Source
World Conference on Lung Cancer
