All‐cause mortality versus cancer‐specific mortality as outcome in cancer screening trials: A review and modeling study

Abstract

Background All‐cause mortality has been suggested as an end‐point in cancer screening trials in order to avoid biases in attributing the cause of death. The aim of this study was to investigate which sample size and follow‐up is needed to find a significant reduction in all‐cause mortality. Methods A literature review was conducted to identify previous studies that modeled the effect of screening on all‐cause mortality. Microsimulation modeling was used to simulate breast cancer, lung cancer, and colorectal cancer screening trials. Model outputs were: cancer‐specific deaths, all‐cause deaths, and life‐years gained per year of follow‐up. Results There were large differences between the evaluated cancers. For lung cancer, when 40 000 high‐risk people are randomized to each arm, a significant reduction in all‐cause mortality could be expected between 11 and 13 years of follow‐up. For breast cancer, a significant reduction could be found between 16 and 26 years of follow‐up for a sample size of over 300 000 women in each arm. For colorectal cancer, 600 000 persons in each arm were required to be followed for 15‐20 years. Our systematic literature review identified seven papers, which showed highly similar results to our estimates. Conclusion Cancer screening trials are able to demonstrate a significant reduction in all‐cause mortality due to screening, but require very large sample sizes. Depending on the cancer, 40 000‐600 000 participants per arm are needed to demonstrate a significant reduction. The reduction in all‐cause mortality can only be detected between specific years of follow‐up, more limited than the timeframe to detect a reduction in cancer‐specific mortality.

Cost-effectiveness analysis of PSA-based mass screening: Evidence from a randomised controlled trial combined with register data

“…Our analysis likely provides a useful building block in that its findings could be input for testing existing cost-effectiveness models’ sensitivity to new information, which has been shown to be useful elsewhere [46]. In addition, similar analyses from comparable trials of organised screening could provide illuminating corroboration or contradiction of the findings presented here, because we are not aware of any other published analyses of mortality from causes other than prostate cancer among diagnosed men in trials of PSA mass screening for prostate cancer [47].…”

What is the balance of benefits and harms for lung cancer screening with low-dose computed tomography?

“…29,30 As lung cancer causes a minority of deaths, even among smokers who meet eligibility criteria for the National Lung Screening Trial, a relative reduction in lung cancer mortality of 20% would only equate to an all-cause mortality reduction of at most 0.8%. 14 Participants in studies, and indeed, screening programmes are likely to represent healthier cohorts within the populations at risk of developing cancer, with important consequences for outcomes. 31 A comparison of National Lung Screening Trial participants, compared to those who would be eligible for lung cancer screening in the US, suggests that a 'healthy volunteer' effect was evident 32 while NELSON participants have been considered broadly representative of the wider population.…”

“…Arguably this is an unrealistic benchmark since it has been estimated that it would require 80,000 patients to be randomised and followed up for over a decade to demonstrate all-cause mortality benefits based on improved lung cancer outcomes alone. 14 The NLST and NELSON studies are summarised in table 1.…”

“… 29 , 30 As lung cancer causes a minority of deaths, even among smokers who meet eligibility criteria for the National Lung Screening Trial, a relative reduction in lung cancer mortality of 20% would only equate to an all-cause mortality reduction of at most 0.8%. 14 …”

Author response to “lack of benefit from low dose computed tomography in screening for lung cancer”

“…Furthermore, studies may not be sufficiently powered to detect a possible favorable difference in all-cause mortality. According to Heijnsdijk et al [7], a significant reduction in all-cause mortality could be expected between 11 and 13 years of follow-up for a sample size of minimal 40,000 high risk participants in each arm. Studies included in our meta-analysis have sample sizes far below 80,000 and some of trials less than 11 years of follow up.…”