November 2013, Vol 2, No 7

← Back to Issue

Gene Expression Tests Using Bronchial, Nasal Epithelium Under Development for Early Detection of Lung Cancer


Diagnosing lung cancer by swabbing a patient’s nose may be possible in the not too distant future. Changes in nasal gene expression in patients with lung cancer have been found to correlate with changes in gene expression in the bronchus, opening the door to the possibility of nasal gene expression as an early diagnostic biomarker of lung cancer, said Avrum Spira, MD, MSc, chief, Division of Computational Biomedicine, and director of Translational Bioinformatics, Boston University School of Medicine.

“What’s in the nose may be a very good surrogate for what’s going on deeper within your lung, “ he said. He delivered his remarks at the second annual Global Biomarkers Consortium conference.

Using Biomarkers for Early Detection

Because the discovery of molecular biomarkers has not yet resulted in cures with the use of targeted therapies, efforts should be directed toward early detection and screening. Currently, fewer than 1 in 5 patients with lung cancer is diagnosed at stage 1, a potentially curable stage, and a similarly small proportion is alive 5 years after their diagnosis, he said.

The vast majority of lung cancers occur in persons with a history of tobacco smoke exposure, but only 10% to 15% of smokers will develop lung cancer during their lifetime. The challenge, said Spira, is how to identify the high-risk subset.

The National Lung Screening Trial (NLST) was a landmark study that found that annual lung CT scans in asymptomatic smokers reduced lung cancer mortality by 20%. By the criteria used in NLST (age >50 years, >30 pack-years of smoking), about 9 million individuals in the United States would be eligible for screening. But only half of lung cancers occur in this group, he said, creating a need for molecular biomarkers to personalize screening decisions.

“I would argue that the second and more urgent need right now is developing biomarkers that can distinguish a benign lesion from a malignancy that’s found on routine CT screening,” said Spira. “CT screening is incredibly sensitive but has a very high false-positive rate…leading to a lot of unnecessary invasive procedures.”

Numerous molecular biomarkers are in development. “In the next 3 to 5 years, a number of them are going to be emerging in the clinic to guide both screening and diagnostic decisions in this setting,” he said.

A bronchial airway gene expression biomarker will be available as a CLIA (Clinical Laboratory Improvement Amendments) test in early 2014. “We’re in the process of extending that same type of biomarker into the nasal epithelium so that we have less invasive sampling and the ability to go into the screening setting,” he said.

Field of Injury Paradigm

The bronchial biomarkers are based on the “field of injury” paradigm that smoking alters epithelial gene expression through the respiratory tract and that variability in airway epithelial cell genomic response to and damage from smoking are linked to lung cancer.

The sensitivity of bronchoscopy for diagnosing lung cancer is only 50% to 60%, and even lower in the early stages of lung cancer. Many bronchoscopies are therefore nondiagnostic, in which case gene expression information can be used to select patients for biopsy and those that can be noninvasively monitored with repeat imaging studies.

In Spira’s work as founder of Allegro Diagnostics, bronchial epithelial cells were collected at the time of bronchoscopy, allowing for a relatively pure population of epithelial cells that line that airway. RNA was extracted from the epithelial cells and processed on a microarray platform that allows for interrogation of the gene expression of all genes in the genome. An 80-gene biomarker profile was identified that could distinguish between smokers who do and do not have lung cancer.

The 80-gene expression biomarker has been validated on 2 independent cohorts. It has an 80% sensitivity and specificity for lung cancer. “More importantly, if we combine our gene expression biomarker with the routine
cytology collected at bronchoscopy, you get 95% sensitivity and 95% negative predictive value,” he said. “So we
can almost, not quite but almost, rule out lung cancer.”

Based on successful validation, the BronchoGen genomic test will be launched as a CLIA assay in 2014. A clinical validation study of smokers undergoing bronchoscopy for suspected lung cancer is ongoing. Data from the first 350 patients show an area under the curve (AUC) of 0.77 in distinguishing cancer from noncancerous nodules, and a negative predictive value of about 90%. The sensitivity of BronchoGen for nodules <3 cm and for stage I was 88%, compared with 66% for bronchoscopy, and the sensitivity for stage I and II disease was 86% compared with only 40% for bronchoscopy. When the biomarker test was combined with bronchoscopy, the sensitivities increased to 96% and 93%, respectively.

Early data suggest that airway gene expression alterations precede the development of lung cancer in high-risk smokers with dysplasia, making intervention in the preventive setting a possibility.

Further research has extended the field of injury to the upper airway. RNA obtained from nasal mucosal brushings has been used to demonstrate that the gene expression changes in the nasal epithelium reflect the changes observed in the bronchial epithelium in smokers. In a diagnostic trial, cells are being collected from the bronchus and nasal epithelium; microarray analysis is being performed to identify biomarkers in the nose in an attempt to diagnose lung cancer. From the first 100 patients, 60 genes were found to be altered in patients with lung cancer compared with those without lung cancer.

“Those genes that change in your nose when you have lung cancer change in almost the identical matter in your bronchus,” said Spira.

A nasal gene expression biomarker has since been developed. From a training set of 100 patients, a 5-gene nasal biomarker yielded the maximum AUC (0.74). The biomarker was tested on 84 independent samples, achieving an AUC of 0.71. “It’s not quite as good as what we’ve seen in the bronchial airways, but there’s still a relatively good signal-to-noise ratio for lung cancer diagnosis,” he said.

In the clinical workflow, he envisions the nasal test being used in patients who have a very small lung lesion on CT who would not undergo bronchoscopy. “A more important place for the nasal testing is as a screening tool, upstream of CT,” he said. “It might be used to decide who should get annual CTs of their chest. The most exciting potential unmet need is to stratify high-risk smokers, not simply to a CT screening protocol, but also into chemoprevention trials.”

His group is in the process of looking at micro-RNA–based biomarkers in airway epithelial cells as potentially more robust biomarkers.

Uncategorized - December 2, 2013

Redesigning Clinical Trials Necessary to Improve Odds of Finding Effective Targeted Agents in the Genomic Era

Defining optimal therapeutic efficacy in the genomic era will require that clinical trial design in oncology move from a drug-centric to a patient-centric approach. Retrofitting current knowledge into old paradigms will slow the progress in discovering effective targeted agents, said Razelle Kurzrock, MD, at the second annual Global Biomarkers Consortium [ Read More ]

Uncategorized - December 3, 2013

VA Takes Lead Role in Using Informatics to Validate and Translate Biomarkers

Clinical informatics – the use of data from the electronic medical record (EMR) to inform the development and improvement of evidence-based medicine, the dissemination and implementation of health outcomes research, and translational research initiatives – can facilitate the validation and translation of biomarkers, said Julie Lynch, PhD, RN, MBA, at [ Read More ]