To the Editors:
We read with great interest the recent article by Calabro et al. 1 showing the close relationship between obstructive lung disease and lung cancer. In particular, we are interested in their suggestion that spirometry might play a very useful role in improving patient selection for targeted lung cancer screening.
Recently, we and others have shown that between 50 and 80% of patients diagnosed with lung cancer had pre-existing chronic obstructive pulmonary disease (COPD) 2–4 consistent with the findings of Calabro et al. 1. In our series of 446 lung cancer patients, we reported that 60–65% had pre-bronchodilator Global Initiative for Obstructive Lung Disease (GOLD) 1+ spirometric criteria for COPD 2 consistent with other published studies 3, 4. This compared with a prevalence of only 15% in a randomly selected population of smokers matched for pack-yrs, sex, ethnicity and smoking exposure 2. This suggests that the association between COPD and lung cancer is strongly independent of smoking exposure dose. Indeed, the risk of lung cancer conferred by a reduced forced expiratory volume in 1 s (FEV1) has been shown to be greater than that conferred by smoking dose (pack-yrs) 5. When chest CT scan evidence of emphysema is included in the COPD (spirometric) criteria, the proportion of those with lung cancer showing signs of airway disease (small airways disease and/or emphysema) may even be higher than 80% 3. Of interest, the study by Calabro et al. 1 confirmed that even mild reductions in FEV1 (<90% predicted) are associated with lung cancer. In our series of nonscreened lung cancer cases 2, 80% had FEV1 <90% pred (unpublished data).
We have previously suggested that reduced FEV1 is a biomarker of susceptibility to smoking for both COPD and lung cancer 6. Here we differ with the views of Calabro et al. 1 who consider FEV1 to be a surrogate marker of exposure to carcinogens from smoking. Although smoking exposure is an important pre-requisite for most lung cancer, the contribution of smoking dose to the variance in FEV1 is modest and much less than that from genetic factors 7–9. This would accord with data from epidemiological studies of chronic heavy smokers where the distribution of FEV1 is bi- or tri-modal rather than unimodal as it is in light and never-smokers 6. In combination, these observations lead us to suggest that FEV1 is a very useful marker of “susceptibility” rather than exposure. This susceptibility of the lungs to smoking is probably due to the combined effects of exaggerated inflammation and/or aberrant repair 10. The former being supported by the close correlation between elevated inflammatory markers (e.g. C-reactive protein or interleukin-6) and FEV1 decline in prospective studies 10.
The basis of the COPD–lung cancer association may be due to overlapping pathogenetic pathways, where exaggerated inflammation and/or lung remodelling leads to COPD (with release of matrix metalloproteinases and growth factors) and subsequent malignant transformation through epithelial mesenchymal transition 10. Recent genome-wide association studies show that some of the genes implicated in lung cancer are also implicated in COPD 11. Specifically, we and others reported in 2008 that the chromosome 15q25 locus, initially associated with lung cancer, is also associated with COPD 12. One important clinical application of the COPD–lung cancer association highlighted by Calabro et al. 1 is the use of spirometry to better target CT screening for the early detection of lung cancer 1. Such an approach has already been reported by Bechtel et al. 13 where including spirometric criteria for CT screening eligibility resulted in a lung cancer pick up rate of 6%, much greater than ≈1.0% reported by Calabro et al. 1 and others 14 where selection is based on age and smoking history alone. Given the constraints of cost and harms from nontargeted screening with CT, we believe maximising specificity over sensitivity is a much more desirable approach to lung cancer screening. By targeting screening for lung cancer to those smokers with evidence of airway obstruction and/or lung parenchymal damage (emphysema), lung cancer pick up can be substantially improved 13. We advocate the use of other biomarkers of lung risk, such as gene-based lung cancer risk assessment 15, to further improve pick up rates (specificity) for those undergoing screening for lung cancer. In this regard, we strongly agree with Calabro et al. 1 in advocating the use of available risk biomarkers (spirometry and genes) to better target lung cancer screening to those smokers at greatest risk.
Statement of interest
None declared.
- © ERS