|Subject: Bronchial Gene Expression Classification for the Diagnostic Evaluation of Lung Cancer|
|Document #: GENE.00051||Publish Date: 07/06/2022|
|Status: Reviewed||Last Review Date: 05/12/2022|
This document addresses gene expression classification for the diagnostic evaluation of lung cancer in individuals with suspected lung cancer following identification of pulmonary lesions on computed tomography (CT) scans. One commercially available test, the Percepta® Bronchial Genomic Classifier, evaluates the expression of 23 genes in bronchial brushing samples taken during bronchoscopy. The intention of the test is to improve the diagnostic performance of bronchoscopy and reduce the need for invasive diagnostic procedures.
Please refer to the documents indicated below for information regarding screening for lung cancer and managing pulmonary nodules:
Investigational and Not Medically Necessary:
The use of bronchial gene expression classification for the diagnostic evaluation of lung cancer in individuals with pulmonary lesions is considered investigational and not medically necessary.
The Percepta® Bronchial Genomic Classifier (Veracyte, South San Francisco, CA), which analyzes 23 genes and age, is commercially available in the United States. The diagnostic performance of this bronchial genomic classifier for evaluating pulmonary lesions was evaluated prospectively by Silvestri and colleagues (2015). Development of the classifier was previously reported by Whitney (2015). The Silvestri publication reported on two cohorts of participants, known as the Airway Epithelial Gene Expression in the Diagnosis of Lung Cancer (AEGIS)-1 and AEGIS-2 studies. Both cohorts had the same eligibility criteria. They included current and former smokers (≥ 100 lifetime cigarettes) age 21 and over who were undergoing bronchoscopy to evaluate lesions suspected of being lung cancer. Individuals were followed until a diagnosis was made or until 12 months after bronchoscopy. If diagnosis of a benign condition or no specific diagnosis was made after 12 months, individuals were considered to be cancer-free.
The gene expression analysis was conducted using epithelial cell samples collected using cytology brushes during bronchoscopy. The same classifier algorithm was used in the two cohorts. Scores were reported as test-positive or test-negative based on a pre-specified threshold value. In addition, treating physicians reported each individual’s pre-bronchoscopy likelihood of having lung cancer on a five-level scale (< 10%, 10 to 39%, 40 to 60%, 61 to 85% and > 85%).
A total of 855 participants in AEGIS-1 and 502 in AEGIS-2 met eligibility criteria and were enrolled in the study, although specimens from 155 individuals (11%) yielded insufficient or poor-quality RNA precluding measurement of the classifier. A total of 639 individuals in the AEGIS-1 and AEGIS-2 cohorts were included in the final analysis. Bronchoscopic examinations were non-diagnostic in 272 individuals (43%), including 120 of the 487 individuals (25%) in whom lung cancer was eventually diagnosed. The sensitivity of bronchoscopy for detecting lung cancer was 74% (95% confidence interval [CI], 68 to 79) in AEGIS-1 and 76% (95% CI, 71 to 81) in AEGIS-2. Of 267 with a non-diagnostic bronchoscopy, 170 (64%) underwent an invasive procedure for diagnostic purposes. Of these 170 individuals, 52 (31%) were diagnosed with benign lesions and 118 (69%) with lung cancer.
Diagnostic accuracy of the bronchial genomic classifier was similar in the two cohorts. In the AEGIS-1 cohort, the classifier accurately identified 194 of 220 individuals with lung cancer (sensitivity 88%; 95% CI, 83 to 92%) and 37 of 78 individuals without lung cancer (specificity 47%; 95% CI, 37 to 58%). In the AEGIS-2 cohort, the classifier correctly identified 237 of 267 individuals with lung cancer (sensitivity 89%; 95% CI, 84 to 92%) and 35 of 74 individuals without lung cancer (specificity 47%; 95% CI, 36 to 59%). When classifier results were combined with bronchoscopy findings, the sensitivity was 96% in AEGIS-1 and 98% in AEGIS-2, which was significantly higher than the sensitivity with bronchoscopy alone, p<0.001.
When the analysis was limited to individuals with a non-diagnostic bronchoscopy (n=272), the classifier accurately identified lung cancer in 49 of 57 individuals in AEGIS-1 (sensitivity 86%; 95% CI, 74 to 94%) and in 58 of 63 individuals in AEGIS-2 (sensitivity 92%; 95% CI, 82 to 97%). However, 13 individuals had a false negative classifier result (i.e., lung cancer and a negative classifier score). Of these, 10 individuals had a high (> 60%) pretest probability of lung cancer according to their treating physicians, and 3 had an intermediate (10-60%) pretest probability. The authors suggested that the classifier might be most useful in individuals with an intermediate pretest probability of cancer because the number of false negatives was small and this group would most likely undergo active surveillance which would likely identify any lesion growth, although active surveillance cannot be guaranteed in clinical practice.
Vachani and colleagues (2016) reported on subsequent invasive procedures obtained by participants in the AEGIS studies. The analysis included 188 individuals with a low or intermediate pretest probability of lung cancer who had an inconclusive bronchoscopy and had available data on additional procedures. Of these, 35 individuals were ultimately diagnosed with lung cancer and 153 had benign diagnoses. Of the 35 individuals who were diagnosed with lung cancer, 4 individuals had a negative classifier result, for a false negative rate of 11%. Of the 153 individuals with benign disease, 42 (27%) underwent subsequent invasive procedures. The classifier was negative (true negatives) in 21 of the 42 individuals (50%) who had invasive testing.
The potential impact of the Percepta bronchial genomic classifier on clinical decision-making was evaluated by Ferguson and colleagues (2016). This study compared board-certified pulmonologist’s recommendations from case evaluation that either included or did not include classifier findings. Physicians were presented with case vignettes based on 36 participants in the AEGIS studies. About half of the cases included classifier information and these cases were presented in random order. A total of 202 physicians participated and completed 1523 case reviews on the 36 AEGIS participants. This included 787 (52%) cases assessed based on clinical information without classifier results and 736 (48%) cases assessed with clinical information plus the classifier results. For cases with a negative classifier finding, physicians recommended invasive procedures in 57% of cases when provided only with clinical information on the case, compared with 18% when informed there was a negative classifier finding (p<0.001). For cases with positive classifier findings, the rate of recommending invasive procedures was 50% when provided only with clinical findings versus 65% when the positive classifier finding was also reported (p<0.001). As a secondary endpoint, the investigators examined the association between physician recommendations and the ultimate diagnosis. In the group with benign conditions the rate of invasive procedure recommendations was 54% when only provided with clinical information and 41% when classifier results were provided (p<0.001). For malignant cases, the rate of recommending invasive procedures was 50% with clinical information only and 64% with classifier information (p=0.003). This study was not conducted in a clinical setting, and subsequent procedures and health outcomes were not assessed.
Another analysis of the potential impact of the Percepta bronchial genomic classifier on clinical decision-making, using data from participants in the AEGIS studies, was published by Sethi and colleagues in 2022. The study included information from 37 AEGIS participants who met the following 2 criteria: 1) pre-broncoscopy risk of malignancy was up-classified by the Percepta test from high-risk to very high-risk; and 2) the initial broncoscopy was not diagnostic for cancer. A total of 101 board-certified practicing pulmonologists were presented with 10 de-identified case vignettes that were presented in 1 of 3 ways: 1) pre-post case in which the information was presented first with and then without Percepta results; 2) case without Percepta results; and 3) case with Percepta results. For the pre-post cases, physicians recommended ablative therapy 17% of the time without Percepta findings and this increased to 45% of the times after Percepta findings were revealed. Similarly, for the cases presented without Percepta results, physicians recommended ablative therapy 17% of the time, and for cases presented with Percepta results, ablative therapy was recommended 56% of the time. As with the Ferguson study, this was a survey and not a clinical study, and subsequent procedures and health outcomes were not assessed.
Lee and colleagues (2021) reported on an analysis of data from a prospective registry study of individuals with pulmonary nodules who had a nondiagnostic bronchoscopy. A total of 283 individuals in the registry had a Percepta test, with at least 1 year of follow-up after testing. Of these, 213 (75.3%) were assessed by physicians prior to bronchoscopy as having either had a low (n=39) or intermediate (n=174) risk of malignancy. The primary endpoint of the analysis was to measure down-classification using Percepta test results in this low- and intermediate-risk group. Invasive procedures were initially planned in 67 of the 213 individuals (31%). After consideration of Percepta results, 23 of the 67 (34%) were down-classified and 17 individuals underwent surveillance instead of invasive procedures. The reduction in recommendations for invasive procedure recommendations with Percepta test results was statistically significant, p<0.01. At the 1-year follow-up, 14 of the 23 individuals (61%) continued to undergo surveillance. The remaining 70 of 283 individuals (24.7%) had a high pretest risk of malignancy. The secondary endpoint of the analysis was to evaluate up-classification based on Percepta results. A total of 60 of 70 individuals (85.7%) were up-classified based on Percepta results to very high risk. Forty high-risk individuals were diagnosed with malignancy, including 37 (92.%) of those who had been upclassified and 3 for whom recommendations were unchanged. There was not a statistically significantly longer time to diagnosis in the individuals with unchanged recommendations; however, the numbers of individuals are small. The study was not randomized and did not compare health outcomes in individuals managed with and without the Percepta test.
While publications on the AEGIS studies show that, with the Percepta bronchial genomic classifier, there is the potential for individuals with a non-diagnostic bronchoscopy to avoid unnecessary invasive procedures, no prospective, randomized application of the Percepta bronchial genomic classifier has been conducted to determine clinical utility or assess outcome data. Of note, the risk of a false-negative finding has the potential to delay diagnosis, resulting in poorer health outcomes. The performance of the Percepta bronchial genomic classifier outside of the clinical trial setting remains unknown.
The National Comprehensive Cancer Network (NCCN) guideline for lung cancer screening (V1.2022) has algorithms for evaluation of different types and sizes of nodules identified by low-dose CT screening. The guideline does not mention gene expression profiling or genomic bronchial profiling for evaluating nodules identified by screening.
Lung and bronchus cancer is the third-most commonly diagnosed cancer in the United States and the most common cause of cancer mortality.
As of 2018, the incidence of new lung and bronchus cancer cases was 53.6 per 100,000 people and the mortality rate was 34.8 per 100,000 people (CDC, 2022). Lung cancer has a better prognosis when it is diagnosed at an earlier stage. For example, the 5-year survival rate of the most common form of lung cancer, non-small cell lung cancer (NSCLC), is 70-90% for stage 1 tumors (Blandin Knight, 2017). However, about 75% of individuals have advanced disease (Stage III/IV) at the time of diagnosis and a much less favorable prognosis.
Cigarette smoking is the leading risk factor for lung cancer (ACS, 2019). The risk increases as the length of time of smoking and the number of cigarettes increases. Screening current and former smokers with low-dose computed tomography (CT) was evaluated in the National Lung Screening Trial (NLST). The study included individuals who were between the ages of 55-74 with at least a 30 pack-year history of smoking and were current smokers or former smokers who had quit within the previous 15 years. The primary finding of the study was that screening with low-dose CT significantly reduced the rate of lung cancer mortality compared with screening with chest radiography (NSLT Research Team, 2011).
Screening for lung cancer is currently recommended for high-risk individuals similar to NLST participants. In 2021, the U.S. Preventive Services Task Force (USPSTF, 2021) published an updated lung cancer screening guideline (Grade B), as follows:
The USPSTF recommends annual screening for lung cancer with low-dose computed tomography (LDCT) in adults aged 50 to 80 years who have a 20 pack-year smoking history and currently smoke or have quit within the past 15 years. Screening should be discontinued once a person has not smoked for 15 years or develops a health problem that substantially limits life expectancy or the ability or willingness to have curative lung surgery.
With the increased availability of CT screening for lung cancer, an increased number of pulmonary nodules are being identified, many of which are false positives. A systematic review of evidence for the USPSTF (Jonas, 2021) stated that the false positive rate in the 27 reviewed studies rangedbetween 7.9 and 49.3 percent in baseline screens, and between 0.6 to 28.6% for subsequent screening rounds. Positive results, including false positives, require follow-up, which can include additional imaging, percutaneous needle biopsy and invasive procedures.
The NCCN has recommendations for evaluation of findings from lung cancer screening in their lung cancer screening guideline (V1. 2022). For solid nodules identified on an initial low-dose CT screen, when lesions are between 6 and 8 mm, a repeat low-dose CT in 6 months is recommended. For lesions between 8 and 15 mm, a repeat CT in 3 months is recommended (or alternatively, consideration of a PET/CT scan), and a PET/CT scan and/or Chest CT are recommended for lesions over 15 mm.
Several classifier tests have been developed to aid in the differentiation of benign and malignant pulmonary nodules. These include proteomic tests and gene expression profiling tests. The Percepta® Bronchial Genomic Classifier (Veracyte) evaluates the expression of 23 genes in bronchial brushing samples taken during bronchoscopy. Classifier test results are intended to reduce the need for invasive diagnostic procedures for nodules identified on CT scans.
Bronchoscopy: An endoscopic test that utilizes either a rigid or flexible scope, in order to visualize and collect samples (washing, brushing, biopsy, culture, etc.) from the endobronchial tubes/branches of the respiratory system.
Computed tomography (CT): An imaging technique that creates multiple cross-sectional images of the body by using special x-rays and computer enhancement to detect disease or abnormalities.
Gene expression profiling: A laboratory test that measures the activity of multiple genes at once for diagnostic or prognostic purposes. The test result is often reported as a proprietary summary score.
The following codes for treatments and procedures applicable to this document are included below for informational purposes. Inclusion or exclusion of a procedure, diagnosis or device code(s) does not constitute or imply member coverage or provider reimbursement policy. Please refer to the member's contract benefits in effect at the time of service to determine coverage or non-coverage of these services as it applies to an individual member.
When services are Investigational and Not Medically Necessary:
For the following procedure codes; or when the code describes a procedure indicated in the Position Statement section as investigational and not medically necessary.
Unlisted molecular pathology procedure [when specified as a gene expression classifier for lung cancer, e.g., Percepta Bronchial Genomic Classifier]
All diagnoses, including but not limited to
Solitary pulmonary nodule
Other nonspecific abnormal finding of lung field
Peer Reviewed Publications:
Government Agency, Medical Society, and Other Authoritative Publications:
|Websites for Additional Information|
Percepta® Bronchial Genomic Classifier
The use of specific product names is illustrative only. It is not intended to be a recommendation of one product over another, and is not intended to represent a complete listing of all products available.
Medical Policy & Technology Assessment Committee (MPTAC) review. Background/Overview and References sections updated.
MPTAC review. Background/Overview and References sections updated.
MPTAC review. Background/Overview and References sections updated.
MPTAC review. Initial document development.
Federal and State law, as well as contract language, including definitions and specific contract provisions/exclusions, take precedence over Medical Policy and must be considered first in determining eligibility for coverage. The member’s contract benefits in effect on the date that services are rendered must be used. Medical Policy, which addresses medical efficacy, should be considered before utilizing medical opinion in adjudication. Medical technology is constantly evolving, and we reserve the right to review and update Medical Policy periodically.
No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, or otherwise, without permission from the health plan.
© CPT Only - American Medical Association