Importance of research design in biomarkers for lung cancer

What is it about?

Biomarkers may be blood, breath and other tests that can be used to find people that may be at high risk of lung cancer or have cancer at the time of the test. CT scanning is an excellent way to detect lung cancer early but is only recommended for people who are at high risk of developing lung cancer. Many people that have scans, as in all screening programmes, do not have lung cancer. Biomarkers may be an important way to improve this situation by selecting more people for CT scans who have cancer and fewer that do not. In this issue of the ERJ two biomarkers intended for this use are reported but both cannot currently be recommended mainly because of the way the research was designed. In the ECLS study, the Early CDT-Lung blood test was used to select current or former smokers for screening but only 10% of the participants who had the blood test were positive and only these people had CT scans. The researchers were not able to say whether the reduction in late-stage lung cancer they found was due to the blood test or the CT scan. It is likely that if all of the people who had a blood test had been scanned, regardless of the result, that many more people would have been detected with early-stage lung cancer. Further work is needed to confirm the place of this biomarker. The second biomarker, again a blood test, was developed and its ability to detect lung cancer was tested in a group of people with lung cancer (cases) and in those who did not have lung cancer (controls). The tests seemed to work well but the cases and controls were different in other ways such as age and smoking habit, two very important risk factors for developing lung cancer. Although promising, this biomarker needs to be tested in the population it might be used in - i.e. people who are at risk of the disease.

Featured Image

Photo by CDC on Unsplash

Photo by CDC on Unsplash

Why is it important?

Low radiation dose computed tomography is an effective method of screening people at high risk of developing lung cancer. Participants are selected either by age and tobacco smoking criteria or multivariable models. Using biomarkers to select people for lung cancer screening may be able to identify those at high risk who do not meet conventional criteria while avoiding further investigation in those without malignant disease. In this issue of the ERJ two biomarkers intended for use in selecting people for CT screening are reported. Both studies are unable to conclude that the biomarker is suitable for this intended use because of issues with the study design. The EarlyCDT-Lung blood test was used to select people at high risk of lung cancer for a regime of LDCT screening but only those with positive tests (10%) had LDCT, which is itself highly sensitive for early-stage lung cancer and can detect cancers 4-6 years before they present. Thus, it is impossible to calculate the contribution of the blood test to the detection of early-stage disease because 90% of the participants were excluded from LDCT. Gaga and colleagues report on the development and validation of a new blood biomarker; however here the issue is that the case and control groups are different and hence discrimination is likely over-estimated. Study design should allow the added contribution of the biomarker to be identified and quantified, in direct comparison to the gold-standard method and in the population in which it is to be used.

Perspectives