Skip to main content
Premium Trial:

Request an Annual Quote

JAMA Study Finds Protein Biomarkers Could Help Select Patients for Lung Cancer Screening


NEW YORK (360Dx) – A global consortium of scientists has identified a panel of protein biomarkers that could help better assess patients' risk of developing lung cancer.

In a study published last week in JAMA Oncology, the researchers from the Integrative Analysis of Lung Cancer Etiology and Risk (INTEGRAL) Consortium for Early Detection of Lung Cancer developed and validated a four-protein panel for identifying patients at high risk of being diagnosed with lung cancer within one year.

The researchers are exploring whether the panel could be useful in selecting patients for lung cancer screening via low-dose computed tomography (LDCT).

LDCT has in recent years become accepted as an approach to early detection of lung cancer. This has stemmed largely from the results of the National Lung Screening Trial, a 53,454-patient study run by the National Cancer Institute and the American College of Radiology from 2002 to 2010 in which annual screening with LDCT demonstrated a 20 percent reduction in mortality compared to screening with standard chest X-rays.

These results led organizations including the US Preventive Services Task Force to recommend annual LDCT screening for individuals at high risk for lung cancer. The procedure is also widely covered by insurers, including Medicare.

The NLST defined high-risk patients as those 55 to 80 years old who have smoked at least a pack of cigarettes per day for at least 30 years, or 30 pack-years, and who currently smoke or have quit within the past 15 years. Under these criteria, around 7 million people in the US are eligible for annual LDCT screening. However, as Samir Hanash, professor in the department of clinical cancer prevention at the University of Texas MD Anderson Cancer Center and senior author on the JAMA Oncology study, noted, these guidelines miss a large proportion of patients who will go on to develop lung cancer.

"The majority of lung cancer patients do not meet current screening criteria, particularly in terms of smoking history," he said.

Given this, researchers have been exploring various approaches to further refine LDCT eligibility criteria, looking at, for instance, whether different cutoffs for age and/or smoking history could better select patients at high risk for lung cancer. The INTEGRAL study aimed to address the challenge through the use of circulating protein biomarkers, and, the authors noted, represents the first such effort in which "a blood-based biomarker score was developed using one cohort and externally validated using prediagnostic samples from other independent cohorts."

Hanash and his colleagues approached the question by selecting five protein markers — surfactant protein B (Pro-SFTPB), cancer antigen 125 (CA125), cytokeratin-19 fragment (CYFRA 21-1), carcinoembryonic antigen (CEA), and human epididymis protein 4 (HE4) — that in past studies have been shown to be useful either in predicting lung cancer risk or for working up and diagnosing the disease.

They then measured levels of these markers in samples from 108 ever-smokers (patients with any past history of smoking) who were diagnosed with lung cancer within one year after blood collection and 216 smoking-matched controls, both taken from the US Carotene and Retinol Efficacy Trial (CARET) cohort. Using those measures, they developed a biomarker-based score incorporating CA125, CEA, CYRFRA 21-2, and Pro-SFTPB to identify patients at elevated risk of developing lung cancer within a year.

They then validated this score in a separate cohort of 63 ever-smokers diagnosed with lung cancer within a year of blood collection and 90 smoking matched controls, both taken from the European Prospective Investigation into Cancer and Nutrition (EPIC) study and the Northern Sweden Health and Disease Study (NSHDS). Comparing a risk model that combined the biomarkers with patient smoking histories to a conventional smoking history-based model, they found that the model incorporating the protein markers outperformed the standard approach in identifying patients who would go on to be diagnosed with lung cancer within a year.

The biomarker-based approach achieved an area under the curve of .83 compared to .73 for the smoking-based model. At a set specificity of 83 percent (with a specificity of 1.00 meaning that none of the healthy controls were deemed eligible for screening), the biomarker-based test had a sensitivity of 63 percent, compared to 43 percent for the smoking-based model (with a sensitivity of 100 percent meaning that all lung cancer cases were deemed eligible for screening). At a set sensitivity of 42 percent, the biomarker-based model had a specificity of 95 percent versus 86 percent for the smoking model.

The researchers also found that addition of the biomarker information could significantly change a patient's risk status as determined by their smoking history. For instance, the smoking model predicted a 60-year-old man with 30 pack-years of smoking history to have a one-year risk of lung cancer of 0.37 percent. Adding the biomarker score, the man's one-year risk shifted either to .07 percent (assuming a biomarker score in the lowest risk quartile) or to 1.56 percent (assuming a biomarker score in the highest risk quartile). Use of the biomarker approach boosted median risk in the lung cancer cases cohort from 0.27 percent to 0.45 percent while lowering it among controls from 0.12 percent to 0.04 percent.

The study authors noted that inclusion of additional markers could boost the model's performance, but Hanash said that he and his colleagues would likely not be adding new circulating markers like proteins due to the need to keep the cost of any such test relatively modest.

"The purpose here is to identify subjects at risk who may not be at high risk based on their smoking history," he said. "Therefore, it's a wider population that would be eligible for the test, and, therefore, the test has to be cost effective. Incorporating additional markers for screening purposes has to be weighed against cost."

Hanash said that the researchers are now planning to further validate the test in samples from the NCI's Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening trial. He added that MD Anderson has also launched a study of LDCT-based lung cancer screening that has enrolled more than 9,000 subjects to date.