Skip to Main Content

An official website of the United States government

Introduction

The National Lung Screening Trial (NLST) was a randomized controlled clinical trial of screening tests for lung cancer. Approximately 54,000 participants were enrolled between August 2002 and April 2004. Participants were randomly assigned to two study arms in equal proportions. One arm received low-dose helical computed tomography (CT), while the other received single-view chest radiography. Participants were offered three exams (T0, T1, and T2) at one-year intervals, with the first (T0) performed soon after entry. The goal of the study was to assess whether low-dose CT screening reduces lung cancer mortality relative to chest radiography among high-risk individuals. Active follow-up data were collected on cancer diagnoses and deaths that occurred through December 31, 2009. Extended follow-up data were collected for deaths through December 31, 2015. Median follow-up time for active follow-up was 6.5 years. The ClinicalTrials.gov number for NLST is NCT00047385.

NLST was a collaborative effort of the National Cancer Institute's Division of Cancer Prevention (DCP) and Division of Cancer Treatment and Diagnosis (DCTD). DCP funded and administered the Lung Screening Study network (LSS, 10 centers, ~35,000 participants) while DCTD funded and administered the American College of Radiology Imaging Network (ACRIN, 23 centers, ~19,000 participants). The study protocols and data collection forms used by LSS and ACRIN were harmonized. A trial-wide database of harmonized data elements was built during the trial for reporting to the Data and Safety Monitoring Board (DSMB). Some additional data elements were collected only by ACRIN or by LSS, as described in the final section of this document.

Recruitment and Enrollment

The NLST Screening Centers recruited potential participants and evaluated their eligibility. Eligibility criteria included the following: age 55-74 years, 30 or more pack-years of cigarette smoking history, and at most 15 years since quitting for former smokers. Exclusion criteria were: history of lung cancer, history of non-lung cancer within the past 5 years, chest CT in the past 18 months, recent medical problems (pneumonia or other acute respiratory infection, unexplained weight loss, coughing up blood), participation in other studies of cancer screening or prevention, removal of any portion of the lungs, inability to lie flat with arms raised over head, home oxygen supplementation, and metallic implants in the chest or back.

Individuals who were ruled eligible and signed an informed consent form were randomized to one of the two study arms. Randomization was stratified by screening center, gender, and age. The date of randomization was the official date of entry into the trial.

Near the time of randomization, participants completed questionnaires on demographics and medical history. Data collected included smoking history, gender, age, race, ethnicity, height, weight, education, marital status, prior cancer diagnosis, prior respiratory or cardiovascular disease, occupational history, family history of lung cancer, history of alcohol consumption, and passive smoke exposure. The T0 screening exam was either performed on the date of randomization or was scheduled for soon after randomization.

Screening for Lung Cancer

Radiologists at the screening centers reviewed the images obtained at each of the three annual screening exams to check for signs of lung cancer. An isolation read and a comparison read, when feasible, were done.

In the isolation read, the radiologist reviewed the images from the screening exam without reference to any historical images. The radiologist recorded information about all visible abnormalities and assigned a preliminary screening result. A positive screening result (suspicious for lung cancer) was assigned if any non-calcified nodules or masses >= 4 mm in diameter were noted or if any other abnormalities were judged suspicious for lung cancer by the radiologist. Three types of negative screening results were possible: clinically significant abnormalities not suspicious for lung cancer, minor abnormalities not suspicious for lung cancer, and no significant abnormalities. Inadequate exams (due to poor image quality) were also noted, but participants with inadequate exams were invited to be scanned again, a practice that yielded adequate exams for nearly all such participants.

In the comparison read, the radiologist compared the images from the most recent exam to images from any previous exams (mostly prior NLST screening exams). Information was recorded about changes to currently observed abnormalities that were also visible in earlier images, and any abnormalities that were missed on the isolation read but noticed after comparison were noted. The radiologist then assigned a final screening result based on the comparison read. For the T0 exam, the result categories were the same as for the isolation read (positive, 3 types of negative, and inadequate). However, for the T1 and T2 exams, there were two types of positive results, as radiologists were required to specify for each positive exam whether abnormalities showed a significant change relative to prior images or not. This requirement had not yet taken effect for T1 exams performed early in the trial, when a single positive category was used. Also, at the T2 exam, if the abnormalities met the criteria for a positive exam but had not changed significantly relative to the previous two exams, the radiologist had the option to assign a result of "Negative - minor abnormalities not suspicious for lung cancer."

The official screening result, reported to participants and their health care providers, arose from the comparison read. Recommendations for diagnostic follow-up were based on this result.

Diagnostic Evaluation for Lung Cancer

Letters reporting the results of the screening exam were sent to the participants and their health care providers within four weeks of the exam. Participants with a result of positive or negative with other clinically significant abnormalities were strongly encouraged to receive a diagnostic evaluation for lung cancer or other suspected condition. The reading radiologist's recommendations for diagnostic procedures were included in the results letter, but the participant's primary physician had full discretion to decide which procedures to perform. The trial did not require a specific diagnostic protocol.

For participants with positive screens, data on their diagnostic evaluation were collected through medical record abstraction. Abstractors reviewed the participants' medical records and recorded information about diagnostic and staging procedures on data collection forms. If lung cancer was diagnosed, further information was recorded about cancer characteristics, initial treatment, and cancer progression.

Lung Cancer Data

Data were collected on all lung cancer diagnoses received by participants during the course of the trial. For cancers not detected by a positive screening exam, several mechanisms were used to ensure complete ascertainment of lung cancer diagnoses. These included (1) study update questionnaires completed by all participants (once per year for LSS, twice per year for ACRIN); (2) direct report to NLST Screening Centers by participants, relatives, or physicians; and (3) death certificates collected by NLST Screening Centers. Medical records were obtained for participants with reported lung cancer. Data were collected on diagnostic and staging procedures, cancer characteristics, initial treatment, and cancer progression. Data on cancer stage were originally recorded according to the AJCC 6 manual, but AJCC 7 data are also available from a re-review of medical records after the conclusion of data collection.

Non-lung Cancer Data

Data were also collected on all cancers of organ sites other than lung that were diagnosed during the trial. The data were obtained through medical record abstraction and were limited to date of diagnosis and ICD-O-3 code. NLST Screening Centers learned about non-lung cancers by the same mechanisms used for lung cancer ascertainment.

Participant Withdrawal

The vast majority (96.6%) of participants were followed until death or study termination, whichever came first. Participant withdrawal from the trial was rare and occurred for several reasons. Some participants refused further participation and communicated this to the NLST Screening Center. A few of these even withdrew consent to participate in the NLST, and all their data were removed from the trial database. Other participants could not be located by the NLST Screening Centers for an extended period of time, so further attempts at follow-up were discontinued by the center. Others stopped participation due to medical problems or cognitive impairment.

Mortality Data

Active Follow-up

The primary trial end point was death due to lung cancer. To assess this end point, data were collected on all deaths (from any cause) that were known to have occurred during the trial. The main ways in which trial personnel learned of participants' deaths were (1) study update questionnaires, (2) reports from relatives, and (3) National Death Index (NDI) Plus searches. After notification, NLST Screening Centers attempted to obtain death certificates for deaths that occurred before December 31, 2009. All information on the death certificate was coded and recorded in the trial database. The underlying cause of death based on the death certificate was derived according to rules established by the National Center for Health Statistics. For deaths that may have been caused by lung cancer, an Endpoint Verification Process (EVP) was conducted to review medical records and determine whether the cause was lung cancer. The EVP cause of death was considered authoritative and was used in statistical analyses of the primary end point.

Extended Follow-up

Additional follow-up beyond the trial endpoint of December 31, 2009, is available. Extended follow-up is available through December 31, 2015.

After the active follow-up period (Dec. 31, 2009), participants were followed only passively through linkages with the NDI. Linkages were performed using probabilistic linkage methods. All except one center was able to be linked. The personally identifiable information NLST had available for linkage included Social Security Number, full name (for some screening centers), date of birth, and sex.

Mortality follow-up was through the end of 2015 for centers with NDI linkage (comprising 97.8% of trial participants) and through the end of 2009 for the one center without NDI linkage.

Results of the Study

The final analysis was conducted in October 2010 and showed a 20% reduction in lung cancer mortality in the low-dose helical CT arm relative to the chest radiography arm. See the main findings section of this Web site for more details.

Additional Data / Materials

Images

The CT images from the screening exams were transferred from the NLST Screening Centers to centralized repositories. CT images may be obtained for research as described on the Images page.

Pathology images of lung tissue were captured as part of the tissue microarray (TMA) construction process. For LSS, these images may be viewed as described on the Images page.

Some of the chest radiograph images from ACRIN sites are available. They may be obtained by completing a data access request per the ACRIN Data Sharing Policy. More information about chest radiograph images may be requested by selecting "Contact Us" at the ECOG-ACRIN website. By completing the form and choosing the subject of "Data Sharing", the pertinent ECOG-ACRIN staff members (including the Data Sharing Administrator at Biostatistics Center) will be notified of the request and will respond to the investigator within 48 working hours.

Biospecimens

A subset of ACRIN Screening Centers participated in biospecimen collection (Biomarker Sites). At these sites, participants were invited to contribute samples of blood, urine, and sputum at each of the three annual screening visits.

Tissue microarrays (TMAs) were constructed from resected tissue from lung cancer participants in LSS and ACRIN. They contain tumor tissue, normal lung tissue, and, in some cases, tissue from lymph nodes and metastatic sites. Separate tissue cores for later extraction of DNA or RNA were also obtained when additional tissue was available.

The ACRIN biospecimen resources are available through the NLST-ACRIN Biorepository Application Process, including the ACRIN tissue microarrays and DNA cores. More information about ACRIN biospecimens may be requested by selecting "Contact Us" at the ECOG-ACRIN website. By completing the form and choosing the subject of "Data Sharing", the pertinent ECOG-ACRIN staff members (including the Data Sharing Administrator at Biostatistics Center) will be notified of the request and will respond to the investigator within 48 working hours. The LSS tissue microarrays and DNA cores may be made available through the PLCO Etiologic and Early Marker Studies (EEMS) process as described on the PLCO Biospecimens portion of the CDAS website.

LSS Contamination Assessment Data

Each April from 2004 through 2010, a sample of 400-500 LSS participants beyond their baseline year was selected to receive a questionnaire (HAQ) about certain medical procedures performed in the previous 12 months. The primary purpose was to assess whether participants in each screening arm were also receiving the other screening arm's test outside the trial during the T0-T2 study years. Information about continuation of screening after the trial's three annual screens was also obtained by the HAQ. The HAQ data are available for download from this Web site.

Additional ACRIN Data

ACRIN collected several types of data that were not collected by LSS. Availability of additional ACRIN data can be requested by selecting "Contact Us" at the ECOG-ACRIN website. By completing the form and choosing the subject of "Data Sharing", the pertinent ECOG-ACRIN staff members (including the Data Sharing Administrator at Biostatistics Center) will be notified of the request and will respond to the investigator within 48 working hours.