Lung cancer fulfils many of the requirements of a condition that is suitable for population screening:
- It is an important health problem: Lung cancer is the second most common cancer diagnosis in each gender after breast/prostate cancer, and the most common cause of death from cancer.
- It has well defined risk factors: Lung cancer is more common in the 55-75 year age group and in smokers
- The natural history of the disease is well understood: For non-small cell lung cancers, the disease is often localised before spreading to lymph nodes then distantly.
- Treatment of earlier stages of disease is available and has better outcomes: Survival after treatment for stage I disease is about 60-70% at 5 years versus < 5% for stage IV disease, therefore earlier detection should lead to higher cure rates. Treatment for early stages is surgery or radiotherapy with or without adjuvant chemotherapy.
- Tests are available: Sputum cytology, Chest x-ray, CT scan and PET scans are all available screening tests and most are available (CT and PET at some expense)
- Tests are acceptable: Most patients are happy to undergo sputum cytology or CXR; CT and PET have higher radiation dose and are less acceptable.
Arguments against a screening policy include:
- There is a high false negative rate: About 95% of suspicious nodules found on CXR or CT are benign. This can cause considerable angst for the patient as well as place them at risk from diagnostic tests.
- There is a high overdiagnosis rate: Screening often detects higher numbers of malignancies, but the mortality rate in a population remains the same, suggesting that some of those detected cancers would not have killed the patient
- There are risks with diagnostic tests: Confirmation of cancer may require bronchoscopy, CT guided biopsy, or thoracotomy. There are risks associated with these procedures, and in 95% of cases the outcome will be benign
- Screening is expensive: The high cost of diagnostic tests, and low number of cancers detected compared to false positives, means that there is considerable expense associated with screening. The single trial (NLSP) that compared CT versus CXR and showed a survival benefit requires a 130,000-260,000 US$ investment per quality of life year saved (compared to $47000 for colorectal and $12000 for breast cancer).
- Screening is with ionising radiation which may increase the number of lung cancers, particularly in smokers, by a small but significant amount on the population level.
- Smoking cessation provides a higher reduction in mortality than screening.
Importantly, only one study has compared CXR screening to standard care. The remainder have compared intense screening to less intense screening. Most of these studies detected higher numbers of low stage cancers in the more intensely screened groups; however, survival was not impacted. Meta-analysis of these trials demonstrated no significant impact of screening with x-rays. The single study to compared annual CXR for 3 years with standard care demonstrated no difference in the rates of lung cancer diagnosis after 15 years follow up; only 20% of cancers were detected through screening. Chest x-ray is therefore an inadequate screening tool.
CT has been evaluated in a single randomised controlled trial, the National Lung Screening Trial. This enrolled over 50,000 'high risk' patients (30 pack year history or more, current smoker or quit within 15 years) to annual CT versus chest x-ray for three years. The trial was stopped early after a 20% reduction in lung cancer mortality was found on an interim analysis. About 95% of nodules on either CT or CXR were benign after evaluation, usually serial imaging. 320 people require annual screening to prevent one lung cancer death.
There are numerous other ongoing studies evaluating the use of CT scanning in lung cancer screening.
Some cancer guidelines recommend screening in high risk individuals. I would discuss the possibility of CT screening with high risk patients, including pros and cons:
- Reduce lung cancer mortality by 20% on a population level; but likely to assist only 1 in every 320 high risk patients who are screened
- Regular imaging tests (inconvenience, ionising radiation)
- Not guaranteed to improve a patient's particular problem
- High false negative rates and anxiety
- Invasive diagnostic tests
I would not recommend introducing a population based lung cancer screening program at this time due to the high cost, low yield and potential for complications during diagnostic work-up.