Screening for lung cancer: The evolving challenge

Cancer of the lung is the leading cause of cancer death in both women and men in the United States. In the year 2012, the estimate is for 226,160 new cases of lung cancer in the U.S., and 164,770 deaths; the five-year survival rate for lung cancer is 16 percent. This sobering outlook is due primarily to the fact that most patients have advanced disease at the time of presentation.

Nonsmall cell lung cancer (NSCLC) accounts for approximately 85 percent of lung cancer diagnoses. For this subset of patients, surgery is often curative if presentation is early (stages I and II). Unfortunately, only approximately 30 percent of diagnoses fall within this early-stage category.

Nearly a quarter of the U.S. population actively smokes, and at the same time contributes to the downstream effects of secondhand smoke. While prevention remains the most important strategy to stem the epidemic of lung cancer, until that goal is realized there remains an urgent need for new and improved means of early diagnosis.

Effective screening

Screening as a means to achieve early diagnosis has always sounded like a good idea but until recently, no screening test had been shown to impact deaths from lung cancer. Prior studies, including the Mayo Lung Project and the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial, utilized chest radiograph (CXR), with or without sputum cytology. Results showed that CXR screening does not result in lives saved from lung cancer.

Computerized tomography (CT) has been in clinical use since the mid-1970s, but it wasn't until technical advances allowing for rapid scanning at low dose that screening became feasible. Single-arm studies of CT screening reported detection of more cancers, more early-stage cancers and with a high resectability rate.

The Mayo CT study enrolled more than 1,500 participants; the baseline nodule detection rate was 51 percent and, after 5 annual scans, 74 percent of the participants had one or more nodules. The vast majority of nodules were found to be benign. There were a total of 68 lung cancers; 79 percent of the incident (not present on the baseline scan) NSCLC were stage I. However, due to the biases inherent to screening, a suggested survival improvement was insufficient proof of effectiveness.

National Lung Screening Trial

The large randomized controlled trial, the National Lung Screening Trial (NLST), has now shown that screening with low-dose CT scanning, compared to CXR, results in fewer deaths from lung cancer. Results were published in August 2011 issue of The New England Journal of Medicine. The NLST compared two methods of detecting lung cancer. Data showed that patients screened with low-dose CT scanning had a 20 percent lower risk of dying from lung cancer compared with patients who received standard CXR.

Investigators at Mayo Clinic in Minnesota and Florida participated in the NLST. The study included 53,454 subjects considered high risk of lung cancer, ages 55 to 74, who were current or former smokers with at least a 30-pack-a-year history. Participants were randomized to either low-dose spiral CT or chest X-ray at baseline, year one, and year two, with follow-up over about six years.

Findings: CT vs. CXR

The finding on CT of a noncalcified nodule of at least 4 mm was considered a positive result, and 27 percent in the CT arm had a positive baseline screen, of which 96 percent were false positives. There were 649 prevalent cancers detected by CT and an additional 367 cancers in the CT arm diagnosed during follow-up after screening. On CXR, any noncalcified nodule visualized was considered a positive result; 9 percent were positive at baseline, of which 94 percent were false positives. In the CXR arm, there were 279 cancers detected and 525 subsequently diagnosed during follow-up post screening.

Within the CT arm, 63 percent of the lung cancers detected by CT were stage I; only 29.8 percent were stages III or IV. Among those detected by CXR, 47.6 percent were stage I, and 43.2 percent were stage III or IV. CT demonstrated the ability to shift stage at diagnosis from advanced disease to more frequent detection in early stage and potentially curative early disease.

There were 356 lung cancer deaths among those in the CT arm versus 443 deaths among those in the CXR arm — a 20 percent reduction with CT screening. In the NLST the number of high-risk participants needed to screen with CT to save one life from lung cancer was 320.

Several other smaller randomized CT screening trials are currently under way in Europe. Recent results from the Danish Lung Cancer Screening Trial (DLCST), where 4,104 subjects were randomized to CT versus no screening, showed that, although a higher proportion of early-stage disease was detected by CT screens, there was no difference in either the diagnosis of advanced-stage disease or mortality after five annual screenings.

Performing the CT is technically simple; the difficulty lies in interpretation and in limiting unintended consequences. CT screening presents issues of false positive scans, resection of benign nodules, overdiagnosis, harmful effects of radiation and cost. For the most part, now that low-dose rapid-scanning CT has been shown to reduce mortality, the problems of CT screening are recognized as manageable, with the possible exception of cost.

Uncertainties of lung CT screening

With the favorable findings on computerized tomography (CT) from the National Lung Screening Trial (NLST) published in 2011, there is a growing urgency to address the uncertainties and apparent limits of CT screening, including:

High rates of false positive scans. A noncalcified nodule of any size will be found in approximately 50 percent of those undergoing CT screening and 98 percent of those will be false positives. Tests to distinguish malignant nodules from benign are imperfect and could subject patients to unnecessary procedures. In single-arm CT screening studies, 15 to 25 percent of the surgical procedures were performed for benign histology. Researchers at Mayo Clinic are developing imaging techniques such as computer-aided nodule assessment and risk yield (CANARY) that may help predict malignant behavior among CT abnormalities, and, by so doing, limit the number of thoracic procedures performed for benign disease.

Overdiagnosis. This term refers to cancers identified and treated that, if otherwise had gone undetected, would not have impacted long-term morbidity or mortality. Many cases of overdiagnosis in lung cancer screening have been attributed to ground-glass opacities (GGOs), which have a long tumor-doubling time. In the Mayo CT screening study, 27 percent of the screen-detected NSCLC exhibited doubling times of > 400 days, suggesting these may have been overdiagnosis cancers. The findings of the Danish Lung Cancer Screening Trial also have raised the possibility of overdiagnosis. Awareness of this issue, it is hoped, will help reduce unnecessary surgery for these indolent cancers.

Radiation risks. There is concern over the risks of radiation exposure associated with serial CT imaging and its role in the development of future lung or other cancers. Comparatively, a single, low-dose CT scan delivers about 1.5 millisieverts (mSv) of radiation, ambient radiation exposure associated with inhabiting planet Earth is estimated at 3 mSv a year, while a standard diagnostic CT of the chest delivers 7 to 8 mSv.

Cost-effectiveness. In this era of intense focus on health care cost containment, questions about the cost-effectiveness of CT screening for lung cancer will continue to arise. The NLST found that the number of high-risk patients needed to screen with CT to save one life from lung cancer was 320. While this number compares favorably to those associated with mammography and colonoscopy, the actual cost per quality-adjusted life-year (QALY) — a standardized measure of cost-effectiveness — is projected to be severalfold higher for CT screening for lung cancer. Estimates of the cost of CT screening have varied from as low as $19,000 to about $169,000 per life-year saved. Further cost information results from the NLST are eagerly anticipated and will be critical to shaping the acceptance of CT screening by third-party payers.

Importantly, it should not be forgotten that smoking cessation programs are a cost-effective preventive measure that has far-reaching downstream effects on reducing deaths from lung cancer.

Next steps

Widespread implementation of screening programs largely depends upon resolution of these uncertainties, particularly the reimbursement issues. Mayo Clinic has developed a program that would include screening patients who fit the NLST criteria for enrollment and others who are at equal or higher risk. Risk is determined on the basis of a diagnosis of chronic obstructive pulmonary disease or family history as supported by the National Comprehensive Cancer Network guidelines.

Greater precision is needed in clarifying indications for screening and for determining who among those with CT abnormalities have lung cancer. This level of clinical detail is likely to be gained by the application and refinement of novel tests of blood, urine, sputum, mucosa or breath analysis that are currently in development.

For more information

The National Lung Screening Trial Research Team. Reduced lung-cancer mortality with low-dose computed tomographic screening. The New England Journal of Medicine. 2011;365:395.