CMS MolDX: Molecular Testing for Detection of Upper Gastrointestinal Metaplasia, Dysplasia, and Neoplasia Form

Summary Of Evidence

Esophageal adenocarcinoma (EAC) has become one of the most rapidly increasing cancers in Western countries.^1-3 Although the 5-year survival rate for EAC has improved from 5% in the 1970s to nearly 20%, it is still among one of the lowest survival rates for all cancers in the United States.^3,4 Prognosis is strongly related to stage at diagnosis, with the 5-year survival rate dropping to less than 3% in patients with a late stage presentation with 40% of patients also having distal metastases.^5,6 As such, there is a need to improve the ability to detect and prevent EAC at an earlier stage with options of curative therapy.

Virtually all EAC arises from BE, a pre-cancerous condition of the distal esophagus where there is metaplastic replacement of the normal stratified squamous epithelium by an intestinal-type columnar epithelium with dysplasia.⁷ Individuals with BE have an over 10-fold increased risk of developing EAC as compared to those without BE.⁸ The prevalence of BE is estimated to be 1.6% in the general population, and higher in patients with GERD.^9-11 Additional risk factors for BE include: male gender, age >50, Caucasian race, smoking history, central obesity, and family history of BE or EAC.^7,12 Importantly, the risk of BE increases additively when there is GERD plus the presence of any additional risk factor.¹³

Progression of BE to EAC often follows a prolonged course through multiple stages of dysplasia encompassing NDBE, BE with low-grade dysplasia (LGD), high grade dysplasia (HGD), intramucosal carcinoma, and finally EAC. The risk of progression of patients with NDBE to EAC is estimated to be between 0.1% and 0.6%.^8,14-17 However, once LGD develops, the risk of cancer progression increases with reported ranges between 0.6-13.4%.^14,15,18 A systematic review of over twenty studies encompassing 2700 patients by Singh et al,19 concluded an annual progression rate from BE-LGD to EAC of 0.54% (1 in 185 patients) annually, albeit with variability across the various studies.

The concept of cellular progression has led to the hypothesis that early surveillance followed by intervention will lead to a decreased incidence of EAC.^20,21 Studies have shown that patients diagnosed through surveillance programs have earlier stage tumors and better survival than those who present with cancer.^22,23 However, guidelines from major gastroenterology societies on early endoscopic surveillance vary.^7,24-26 For example, the American Society for Gastrointestinal Endoscopy (ASGE), only conditionally recommends surveillance endoscopy in patients with NDBE based on very low quality of evidence when compared to no surveillance.²⁵ In addition, the ASGE states that if endoscopy for the identification of BE is performed, it should only be in at-risk patients such as those with a family history (high risk) or those with GERD plus at least one other risk factor (moderate risk). The American College of Gastroenterology (ACG) recommends assessing men with chronic (>5 years), and/or weekly or more symptoms of GERD, and two or more risk factors for BE or EAC through a single screening endoscopy and that for patients without dysplasia, endoscopic surveillance should take place at intervals of 3-5 years.⁷

While such a diagnostic procedure is considered safe, endoscopy is an invasive procedure, requires patient sedation, and is associated with significant direct and indirect patient costs.²⁷ In addition, endoscopists often do not adhere to the Seattle protocol^28,29 and dysplasia can be patchy and not always associated with visible abnormalities leading to biopsy sampling error.²⁸ Once the biopsy is obtained, there is also variable interobserver agreement among pathologists on the grading of dysplasia,^30,31 inflammation can be difficult to distinguish from LGD, there are no formally validated morphologic features to distinguish LGD from HGD, and distinguishing HGD from intramucosal carcinoma can be difficult.^30-32 Also, despite the increasing use of endoscopy, a large percentage of EAC is diagnosed in patients without a prior BE diagnosis, and 25% of patients with NDBE (or BE with LGD) are diagnosed within 1 year of the initial endoscopy.^8,33 In fact, one retrospective study concluded that endoscopic surveillance of patients with BE was not associated with a decreased risk of death from EAC.³⁴

The progression from NDBE to EAC involves several molecular changes including DNA or sequence alterations, structural genomic changes, and epigenetic modifications. Multiple models have been proposed including a gradual transition in which mutations are accumulated, a “big bang” in which a majority of mutations are observed during the first few cell divisions establishing BE, punctuated equilibrium in which periods of stasis are punctuated by rapid periods of transformation, and massive chromosomal rearrangements occurring at a single event, known as chromothripsis.²¹ Therefore, molecular markers that can be used to identify these stages and stratify patients at increased risk for progression to EAC are of clear interest. For example, genomic instability, gains or losses in parts of chromosomes, is a hallmark of cancer development. Using DNA flow cytometry Choi et al,³⁵ recently demonstrated that biopsies from patients with NDBE demonstrated normal DNA content, in contrast to patients with HGD in which abnormal DNA was detected in 93.8% of patients with over half of those patients having concurrent or subsequent cancer diagnosed in under 3 months. A recent case-control study reported that somatic TP53 mutations were detected in 46% of biopsy samples from progressors (HGD and EAC) and 5% of non-progressors (BE with at least 5 years without progression).³⁶ In addition, the authors noted the detection of TP53 mutations in NDBE often preceded the progression to HGD or EAC.³⁶ A similar study measuring mutational load (ML) using a panel of DNA markers, loss of heterozygosity, and microsatellite instability also found that patients who have NDBE that eventually progressed to HGD or EAC had elevated ML prior to progression.³⁷ Other studies have shown that deoxyribonucleic acid (DNA) methylation is an important mechanism that defines subclasses of BE and EAC, and mediates the development of EAC.^38-41

ACG guidelines state that a swallowable, non-endoscopic capsule sponge device combined with a biomarker is an alternative to endoscopy, albeit with conditional strength of recommendation and very low quality of evidence.⁷ The AGA best practice advice statements have also been updated to consider non-endoscopic cell collection devices as an alternative.²⁴ Several assays using these devices have been developed.^41-43 The Cytosponge is a cell sampling device comprised of a compressed mesh attached to a string encapsulated within a gel. The capsule is swallowed, dissolves in the stomach and the mesh is withdrawn, collecting cells that can be processed for immunohistochemistry or molecular analyses. Initial studies examined collected tissues labeled with trefoil factor 3 (TFF3), a protein biomarker for intestinal metaplasia (IM).⁴⁴ The Cytosponge-TFF3 test has been tested in multiple clinical studies comprising the Barrett’s Esophagus Screening Trials (BEST) 1, 2 and 3. The BEST3 study was a randomized trial in which eligible patients either received standard management of their symptoms with referral to endoscopy if required, or were placed in an intervention group and offered the Cytosponge-TFF3 procedure with subsequent endoscopy if results were positive.⁴⁵ Of over 13,000 patients eligible for the study, 1654 successfully underwent the Cytosponge-TFF3 test. The results showed that the Cytosponge-TFF3 test diagnosed 10 times more cases of BE than standard of care. However, the test results required manual review by a pathologist and there was variation in the quality of endoscopies across the 24 hospitals taking part in the study, highlighting the potential for observer bias at multiple steps of the test. Economic evaluations considering the increased numbers of endoscopies performed as a result of the procedure are underway. Additional studies using the Cytosponge coupled with a panel of biomarkers including protein, methylation, and TP53 status to risk stratify patients are also in progress.⁴⁶ Moinova et al,⁴¹ describe the use of a balloon device that is swallowed, inflated, and then withdrawn back through the distal esophagus to sample the luminal epithelial surface. DNA is then extracted from the cells that are collected and tested for the presence of methylated markers CCNA¹ and VIM.⁴¹ Although the test demonstrated similar performance to a training group of cytology brush samples obtained during endoscopy with a sensitivity of 88% across all BE, dysplasias, cancers, and 92% specificity, only a very limited number of dysplastic cases were tested and sensitivity of detecting HGD and EAC were 50% and 87.5% respectively.⁴¹ Sensitivity in the detection of NDBE was 90.3%. At an incidence rate of 5% of BE in patients with GERD, these results yield a positive predictive value (PPV) of 35.7%, a negative predictive value (NPV) of 99.3%, and an accuracy of 91.5%. In an alternate study using a cytosponge-based assay for the detection of multiple methylated genes in more than 250 patients, Chettouh et al⁴⁷ identified methylated TFPI2 as a molecular marker for BE with sensitivities of 85% and 79% in pilot and validation cohorts respectively with 96% specificity.

Analysis of Evidence

Multiple medical societies recommend both screening and surveillance for BE. In most instances, only patients with GERD plus additional risk factors are recommended to be assessed by endoscopy and surveillance intervals for those with NDBE vary.^7,24-26 However, many society guidelines cite the cost of endoscopy, potential harm, and patient burden as prohibitive and recommendations are conditional based on a low quality of evidence. In addition, no studies comparing at-risk patients that receive endoscopy to those that do not have been performed. However, given the low incidence of EAC in patients with BE, such a study would require thousands of patients and years of follow-up.⁴⁸ Therefore, many studies are performed on patients already in surveillance programs and care must be taken in interpreting those results.⁴⁹

Studies have reported that fewer than 10% of patients with EAC have a prior diagnosis of BE and approximately 40% describe no history of GERD, indicating that the current practices do not identify many at-risk patients.^8,34,48,49 Therefore, alternative methodologies that are less costly and less invasive than conventional endoscopy are desired, and a number of these tests are currently in development or commercially available.

Molecular biomarkers have been recognized as a valuable source of information regarding cellular progression in upper-gastrointestinal pathology. Evidence indicates that this type of testing has the potential to determine which patients are or are not likely to have treatable conditions related to metaplasia, dysplasia, or malignancy without the need for invasive testing. Patients with increased risk who have not progressed to dysplasia will benefit from early testing as they are unlikely to have a meaningful benefit from endoscopic evaluation. This would permit management of these patients congruent with clinical guidelines without subjecting them unnecessarily to more invasive procedures. Conversely, patients with results indicating the presence of dysplasia would benefit from early preventative intervention, such as endoscopic eradication therapy. As such, well validated assays with clearly demonstrated clinical and analytical validity can further inform clinicians on whether a patient has a treatable malignant or premalignant condition are reasonable and necessary.

To provide the benefit of detecting precancerous disease amenable to early treatment a test must be able to detect a condition before the patient presents with symptoms, lead to the avoidance of unnecessary and invasive procedures, or to an intervention that yields a superior outcome, and be readily available, feasible and economical. In this case, the test would ideally be highly sensitive and highly specific. Studies must also be carried out in the relevant populations, and caution must be taken in extrapolating results of studies performed in populations already enriched for BE.

Although there are several promising molecular biomarker tests designed to further identify at-risk patients, this contractor finds that there are currently no existing tests that have demonstrated AV, CV, and clinical utility (CU) to fulfill the necessary criteria. This contractor will continue to monitor the evidence and may revise this determination based on the pertinent literature and society recommendations.

Current molecular diagnostic tests that identify individuals with upper gastrointestinal metaplasia, dysplasia, and neoplasia are non-covered by this contractor.

We expect these types of tests to meet the following criteria:

• The beneficiary is being actively managed for chronic gastroesophageal reflex disease (GERD) and/or non-dysplastic Barrett’s esophagus (NDBE); and
• They also have at least three additional risk factors for Barrett’s esophagus (BE) as described in nationally recognized guidelines.
• The beneficiary has not been previously diagnosed with dysplasia or esophageal carcinoma; and
• The beneficiary is tested no more than recommended by established national or society guidelines; and
• The test identifies patients with dysplastic disease that may benefit from endoscopic treatment or surveillance, or patients with non-dysplastic disease who may benefit from surveillance; and
• The test results  will be used in determining treatment or management of the beneficiary.
• The beneficiary is within the population for which the test was developed and validated. The laboratory providing the test is responsible for clearly indicating to treating clinicians the population and indication for test use.
• The test demonstrates analytical validity (AV) including an analytical and clinical validation for any given measured analytes, and has demonstrated equivalence or superiority for sensitivity or specificity of detecting dysplasia to other already accepted methods for the same intended use measuring the same or comparable analytes.
• Clinical validity (CV) of any analyte measured must be demonstrated in the published peer-reviewed literature, establishing a clear and significant biological/molecular basis for stratifying patients and subsequently selecting (either positively or negatively) their clinical management decision within a clearly defined population.
• The test successfully completes a Technical Assessment that will ensure that criteria set in this policy are met to establish the test as Reasonable and Necessary.