It has been nearly 70 years since Doll first described the association between cigarette
smoking and lung cancer (1), and later others followed by discovering the association
of smoking with chronic obstructive pulmonary disease (COPD). Some smokers suffer
the unfortunate “double whammy” of having preexisting COPD that increases the risk
of developing lung cancer and are diagnosed with both. More recently, screening high-risk
current and former smokers for lung cancer using radiation low-dose computed tomography
(CT) has been added to the mix (2, 3). As screening is implemented globally, and given
the increased risk of developing lung cancer in those suffering from COPD, there is
a temptation to advocate screening all eligible patients with COPD with the hope of
enriching the screened population with more patients likely to develop an early, curable
cancer. To test that hypothesis in those eligible for screening, it would be important
to know the prevalence and severity of COPD, the presence of respiratory symptoms,
and the prevalence of other comorbid conditions to better understand their effect
on screening outcomes. The study by Ruparel and colleagues (pp. 869–878) published
in this issue of AnnalsATS goes some way to providing that information (4).
The study design was clever. Nearly 8,000 patients eligible for lung cancer screening
by age, smoking history, or meeting certain thresholds for developing lung cancer
using validated lung cancer risk models were invited to participate in a “lung health
check,” rather than just being offered a chest CT screen (4). In addition to a detailed
medical history, they all had spirometry. Perhaps one of the most important aspects
of this study was that it was undertaken in underserved communities, as evidenced
by the fact that more than half had left school before the age of 15 years and there
was a higher proportion than expected of current heavy smokers. These communities
are where lung cancer screening is likely to have the most impact. The findings were
eye opening, with more than half of the participants having spirometric evidence of
COPD. Surprisingly, two-thirds of those did not have a documented prior history of
the disease. Not surprisingly, the odds of having respiratory symptoms increased significantly
as airflow obstruction worsened, and the presence of comorbidities was significantly
higher in those with known versus undiagnosed COPD.
Now comes the tricky part. How do we convert the knowledge gained from this study
into our practice of screening for lung cancer? We know that the risk of lung cancer
increases in a linear relationship as the severity of airflow limitation worsens,
which means those at greatest risk of lung cancer generally have the highest prevalence
of COPD (5, 6). This observation underpins an important limitation of using a risk-based
(i.e., age and pack-year history) approach to selecting smokers for screening. Increasing
age and smoking history not only increases the likelihood of developing lung cancer
but also increases the likelihood of getting COPD. This is important because COPD
is associated with a number of other smoking-related diseases, such as coronary artery
disease, and is a strong marker of reduced life expectancy among smokers (7–9). This
translates to reduced gains from screening. Indeed a post hoc analysis of the NLST
(National Lung Screening trial) reported that for those with COPD (the presence of
airflow limitation on the basis of spirometry), the reduction in lung cancer mortality
associated with the CT arm relative to the chest radiograph arm was about 15%, in
contrast to those with normal lung function, where the reduction in lung cancer mortality
was 28%, nearly twofold greater (6). In a further post hoc analysis, this reduced
benefit for those with COPD was linked to lower surgical rates for early-stage cancer
and greater deaths from non–lung cancer causes (6). In the same analysis, there was
no screening benefit in those with Global Initiative for Chronic Obstructive Lung
Disease 3 or 4 severity (6). These findings suggest that underlying (and unrecognized)
COPD may contribute to both undertreatment, where surgery is inappropriately withheld,
and overtreatment, where surgery is undertaken but no survival benefit is achieved.
Another feature of having preexisting COPD in the context of lung cancer is that it
is associated with more aggressive forms of lung cancer (10). This means more aggressive
biology reflected in both histology (more small-cell, squamous-cell, and non–small-cell
subtypes) and shorter volume-doubling times (10, 11). This suggests that lung cancers
developing in those with advanced COPD may be less amenable to early diagnosis by
screening and confer only modest long-term survival after surgery (6).
If reducing lung cancer mortality and prolonging survival are the key determinants
of successful screening, then it is likely that screening those at intermediate risk
where life expectancy is greater (than for those at high risk) may be a more cost-efficient
way to undertake screening. Although the routine use of spirometry in the setting
of lung cancer screening may identify those with significant airflow limitation and
greatest risk of lung cancer, it may have a greater utility in identifying who has
severely limited life expectancy and who gains little from screening for lung cancer
(e.g., Global Initiative for Chronic Obstructive Lung Disease 3–4 COPD) (6). In other
words selection using a solely “risk-based” approach, with the aim to enrich the number
of lung cancers identified through screening, may not improve the risk–benefit balance,
particularly where the benefit ignores long-term survival. Recently, Cheung and colleagues
developed a life-years–gained model, compared it to traditional risk-based screening,
and found that using their approach would maximize the benefits of screening by including
patients who have both a high risk of developing lung cancer and a long life expectancy
(12).
Although all of these machinations seem like an exercise built for ivory tower academics,
the reality on the ground is that <10% of those eligible for screening in the United
States have been screened (13–15). Although spirometry is likely to be helpful in
assessing a patient’s risk of developing lung cancer and may provide information regarding
the risk/benefit ratio from screening, one has to question whether it is feasible
to add spirometry to a visit that already includes a shared decision-making component
and a smoking cessation consultation for current smokers. Still, the concept of a
“lung health check” is intriguing, as it may provide a forum for checking a patient
for both COPD and lung cancer and present an opportunity for smoking cessation counseling
to those with newly diagnosed disease. Following this cohort is critical, as it will
add to our knowledge of how to interpret the effect COPD has on the outcomes of lung
cancer screening and whether upfront spirometry should become a routine part of the
screening visit. Stay tuned.
Supplementary Material
Supplements
Author disclosures