Statistics from Altmetric.com
We read with interest the paper by Black and colleagues1 which outlines the current status of CT screening, and value the authors’ cautious interpretation of the relatively few well conducted studies regarding this controversial topic. However, when using reduction in lung cancer mortality as proof of screening efficacy with CT due to early intervention, one has to be careful in interpreting the calculation of the potential reduction in mortality. This is because the denominator used for calculating disease-specific mortality is also affected and is thus biased by the proportion of early cancers detected, especially when overdiagnosis is likely to be encountered.
We would like to highlight the recent ELCAP study to illustrate this.2 Screening of 27 456 participants led to the detection of 74 early lung cancers which translated to an annual incidence of 269/100 000 persons at risk (100 000/27 456×74). The reported cure rate was 80% and mortality was 20%. Although we acknowledge that the study included participants from several countries, Centers for Disease Control and Prevention have reported annual lung cancer mortality of 83.3/100 000 men and 53.7/100 000 women.2 3 Assuming equal gender distribution, lung cancer mortality of 68.5/100 000 is obtained. When this figure is compared with the ELCAP study, an overdiagnosis of 200/100 000 persons could be implied by CT screening alone. Considering the generally quoted dismal cure rate of 15% for 69/100 000 persons, overdiagnosis and overtreatment of such a magnitude would automatically result in a higher cure rate of 78.5% (69×15% + 200×100%) and 21.5% mortality.
It would therefore appear premature to associate the effectiveness of lung screening with a higher cure rate or reduction in mortality. Instead, a significant reduction in annual lung cancer mortality following the start of any screening method will be the proof of clinical significance. In our opinion, it should decrease lung cancer mortality statistics year after year.
Competing interests: None.