Test Description

The test is as described¹, but has been modified for clinical testing as follows: Genomic DNA is prepared from patient EDTA-blood samples. 2.6 million nucleotides of target genomic regions, representing exons, intron boundaries and non-exonic mutation containing regions in 527 genes are enriched ~500-fold from the 3.16 billion nucleotide (nt) genome of each sample. Enrichment uses hybrid capture, in which tens of thousands of oligonucleotide probes capture 8,614 genomic DNA fragments, collectively comprising 592 disease genes. Patient DNA is fragmented, denatured and incubated with the oligonucleotides. The target-oligonucleotide hybrids are isolated by magnetic capture²³. Next generation sequencing of the enriched targets is performed with Illumina HiSeq and TruSeq sequencing-by-synthesis, yielding ~3 billion nucleotides of sequence per sample, each ~125 nucleotides long. Sequences are aligned to the reference human genome uniquely, covering each target nucleotide ~150 times. Alignment uses the algorithm GSNAP²⁴²⁵, with parameters that have been optimized for clinical diagnostic use. Enrichment and sequencing are performed on multiplexed samples, which are disambiguated by molecular barcodes. ~1% of target nucleotides are not covered, while ~95% of target nucleotides have at least 16-fold sequence coverage. The majority of missed nucleotides are in high GC-content targets, are missed reproducibly, and are labelled as such. An automated bioinformatic decision tree is used to identify and genotype variations in the aligned sequences^242627282930. Variants are retained if present in at least 8 sequences of quality score >25 and in exons with at least 16-fold sequence coverage²⁴. Variants detected in >86% of reads are considered homozygous, while those present in 14-86% of reads are heterozygous. Variants are classified according to ACMG and other guidelines^{2183132333435}, using literature knowledge as well as in silicotools, such as comparison with a variety of mutation and human variation databases, PolyPhen-2 and SIFT, to determine the pathogenicity of each variant. Pathogenic variants are assembled into genotypes and reported. For diagnostic testing, where variants are of uncertain significance, further evidence is sought, using additional in silico tools, literature evidence, clinico-pathologic correlation, confirmatory family studies or functional assays, as appropriate. In general, variant interpretation is identical to that performed using conventional molecular diagnostic assays with the exceptions that clinico-pathologic interpretation and masking of non-relevant genes are routine in diagnostic use of the assay and that ~90% of variant annotation and reporting is automated, facilitating interpretation and standardization of reporting. Reporting of variants differs in carrier testing of adults and diagnostic testing of children³¹. Carrier testing reports carrier status in all genes. Diagnostic testing reports positive and negative results in genes relevant to the clinical presentation. Diagnostic testing in children does not report carrier status in genes that are not relevant to presentation³¹. In a subset of cases, further communication between the laboratory director and ordering physician is necessary to guide additional studies and assist in interpretation.

Public Health Importance

Mendelian diseases collectively affect 13 million people in the US, accounting for ~20% of infant mortality and ~18% of pediatric hospitalizations^3637383940.

Diagnostic testing in potentially affected children

Simultaneous diagnostic testing for 595 recessive childhood diseases is anticipated to have several public health impacts: 1). Extension of the prevention, diagnosis, and treatment benefits demonstrated for conventional genetic testing to hundreds of recessive diseases for which testing is not available today; 2). Reduction in time-to-diagnosis, particularly in illnesses where the differential diagnosis is broad and the conventional approach is serial univariate testing. Serial univariate testing can take over a year, delaying timely intervention or counseling. The initial turnaround time of the test will be 4 weeks. 3). Reduction in cost of diagnosis. The average cost per patient of serial univariate molecular diagnostic testing is ~$10,000 at our institution. The test is anticipated to cost ~$600. 4). Increased rate of definitive molecular diagnosis. Less than 50% of patients undergoing serial univariate molecular diagnostic testing receive a molecular diagnosis. This is anticipated to increase with test use, particularly in illnesses where the differential diagnosis is broad, such as mitochondrial myopathies or intellectual disability. Timely diagnosis of affected individuals has several potential benefits:

1. Prevention of death or markedly diminished disease severity where curative treatments are available. Quite a large number of recessive diseases have specific therapies. Neonatal diagnosis and treatment of phenylketonuria (PKU) and congenital hypothyroidism prevent severe intellectual disability. Likewise, death is prevented in certain forms of congenital adrenal hyperplasia (CAH), medium chain acyl-coA dehydrogenase deficiency (MCAD), and galactosemia (OMIM #230400).

2. Genetic counseling of patients and families about risks for relatives and in additional offspring.

3. Improvement in quality of life in disorders where treatments are ameliorative. While many recessive diseases lack curative treatments, timely diagnosis nevertheless allows specific interventions that can substantially improve quality of life. Such interventions may slow disease progression, lessen symptoms, prevent complications or improve function in affected organ systems.

4. Substantial psychosocial benefits with respect to anxiety, self-image, uncertainty and lifestyle decisions.

5. Multiplexed testing allows rule-out of differential diagnoses, decreasing unnecessary treatments.

Use of the research version of the test revealed that 27% of literature mutations are common polymorphisms or misannotated¹. Thus, it is critical to establish a clinical grade mutation database for recessive illnesses. Implementation of the test for diagnosis in affected children will, with time, improve the quality and quantity of annotated mutations, particularly for diseases that for which no molecular test is available currently.

In addition, test results have a cumulative potential to inform an understanding of disease mechanisms. In each individual with a Mendelian disorder, the specific mutations impact the age of onset, disease severity, rates of progression, distribution of affected organs, complications, pleiotropy and outcomes. Only in diseases for which molecular diagnosis is undertaken can such knowledge be accumulated. A broad understanding of genotype-phenotype relationships can enable individualized care of patients with recessive diseases. This can potentially include individualized treatment intensity and prediction of disease progression, severity and likely complications. Thus, in the long term, the test, when performed in a research setting, can allow identification of genotype-phenotype relationships that allow conveyance of individualized diagnostic information.

Initial experience with the test has revealed the existence of novel modifier mutations and pleiotropy in patients with recessive illnesses (Kingsmore et al., submitted). Only through multiplexed molecular testing can such knowledge be accumulated. A broad understanding of modifier genes can further enable individualized care of patients with recessive diseases. Thus, in the long term, the test, when performed in a research setting, can allow identification of modifier genes that allow conveyance of individualized diagnostic information.

Finally, timely molecular diagnosis can allow intervention before organ decompensation, when treatment is likely to alter outcomes. Currently, study of new therapies for rare disorders are hampered by diagnosis after organ damage and low rates of ascertainment. Timely diagnosis can permit regional referral of affected individuals for specialized treatment.

It should be noted that substantiation of the potential public health impacts in prevention, diagnosis, or treatment of recessive childhood illnesses is needed. Such assessments should include measurement of cost effectiveness including costs of follow up of ambiguous test results and counseling.

Published Reviews, Recommendations and Guidelines

Systematic evidence reviews

The emerging use of targeted sequencing of panels of genes, whole exome sequencing and whole genome sequencing for molecular diagnosis of Mendelian diseases was recently reviewed⁴⁵.

Recommendations by independent group

Currently none.

Guidelines by professional groups

The United Kingdom’s Human Genetics Commission recently reported guidance on preconception genetic testing within the framework of a population screening program¹⁴.

Evidence Overview

Analytic Validity:

The 437 genes responsible for 448 childhood recessive diseases are listed in Table 1. Using genotyping cut-offs of 14% and 86% to differentiate homozygotes and heterozygotes and >20X nucleotide coverage and >10 reads of quality >20 to call a variant, the accuracy of the test for SNP genotyping was 98.8%, analytic sensitivity was 94.9% and analytic specificity was 99.99% for 92,106 SNPs in 26 samples genotyped both by high density arrays and the test¹. The positive predictive value (PPV) of the test for SNP genotyping was 99.96% and negative predictive value was 98.5%, as ascertained by array hybridization¹. As sequence depth increased from 0.7 to 2.7GB, test sensitivity increased from 93.9% to 95.6%, whereas PPV remained ~100%. Area under the curve (AUC) of the receiver operating characteristic (ROC) of the test for 92,106 SNP genotypes in 26 samples, when compared with array hybridization, was 0.99 when the number and % reads calling a SNP was varied.

For known substitution, indel, splicing, gross deletion and regulatory alleles in 76 samples, analytic sensitivity was 100% (113 of 113 known alleles). The higher sensitivity for detection of known mutations reflected manual curation. The twenty known indels were confirmed by PCR and Sanger sequencing. Of note, substitutions, indels, splicing mutations and gross deletions account for the vast majority (96%) of annotated mutations²⁷.

Unexpectedly, 14 of 113 literature-annotated disease mutations were either incorrect or incomplete. PCR and Sanger sequencing confirmed that the 14 variants and genotypes called by the test were correct¹.

Gross deletions were detected both by perfect alignment to mutant junction reference sequences and by local decreases in normalized coverage (normalized to total sequence generated). Eleven of eleven gross deletion mutations for which boundaries had been defined were identified¹. Further analytic validation of ability to detect and genotype gross deletions, gross insertions and complex rearrangements is required.

It should be noted that the clinical version of the test will feature several improvements that are anticipated to improve analytic sensitivity and specificity. These are: 1). Increased depth of sequencing to 3 GB per sample; 2). Automation of the sequencing library preparation and target enrichment; 3). Re-design of the target enrichment oligonucleotides; 4). Change in the variant detection parameters to >16X nucleotide coverage and >6 reads of quality >25 to call a variant; 5). Further refinement of alignment parameters to prevent variant detection solely at the ends of reads; 6). Increased library size to reduce overlap redundancy; 7). Improved sequencing-by-synthesis chemistry (TruSeq); 8). Improved HiSeq instrument specification. Repetition of analytic validation is ongoing in a CLIA-compliant laboratory setting.

Clinical Validity

There are no published systematic evidence reviews of test accuracy, reliability or predictive value in a clinical setting. Experience is being garnered with the use of whole exome or whole genome sequencing for molecular diagnosis of Mendelian diseases and was recently reviewed.

Clinical Utility

There are no published systematic evidence reviews or published clinical trials. Published experience was in a research setting and was not blinded to sample diagnosis¹. Test development and assessment of analytic and clinical validity and utility are ongoing.

Links

http://www.beyondbatten.org/

http://www.ncgr.org/preventing-rare-genetic-diseases

http://hematite.ncgr.org/

www.sciencemag.org/content/331/6014/130.full

http://www.npr.org/2011/01/13/132908098/new-gene-test-screens-nearly-500-childhood-diseases

Last updated: March 18, 2011

Table 1

Competing Interests

The author has received in-kind funding from private companies (Illumina Inc., Life Technologies Inc., Roche-Nimblegen and British Airways PLC).

Background

Colorectal Cancer (CRC) screening

Screening by colonoscopy, sigmoidoscopy and fecal occult blood testing has been shown to prevent colorectal cancer (CRC) and to reduce mortality through the detection and removal of pre-cancerous lesions and through the detection of CRC in its early stages [1] [2]. Indeed, CRC incidence and mortality have been decreasing since 1985 [2] [3]. Research suggests that CRC screening may be responsible for approximately half of the declines [2]. However, uptake of CRC screening recommendations in the U.S. is not optimal. In 2008, only about 62% of men and women aged 50-75 years reported getting the most commonly recommended CRC screening tests, a percentage that varied from 49-75% among states [4].

Several tests are available to identify colorectal cancer and pre-cancerous polyps in asymptomatic individuals. Colonoscopy visually inspects the interior walls of the entire rectum and colon. Performance characteristics (such as sensitivity and specificity) of new tests are commonly evaluated in comparison with colonoscopy [1] [5]. Flexible sigmoidoscopy involves a more limited visual inspection of the distal colon and rectum. Fecal occult blood tests (FOBTs), which include conventional guaiac FOBT, high-sensitivity guaiac FOBT, and fecal immunochemical tests (FITs), chemically detect small amounts of fecal blood (which can originate from pre-cancerous and cancerous colorectal lesions). CT colonography (i.e., virtual colonoscopy) and double-contrast barium enema (DCBE) are additional tests, offering enhanced x-ray images of the interior rectum and colon to aid in detecting abnormalities.

Fecal (stool) DNA tests have been under continuous development over the past several years. These tests are designed to detect in stool samples any number of DNA markers shown to be associated with CRC. ColoSure™ is the latest example of a clinically available stool DNA test.

Clinical Scenario

The clinical scenario for fecal DNA testing in general is most often presented as colorectal cancer screening in average-risk individuals.

A technical brochure for ColoSure [6] states that:

“ColoSure is not intended to replace a colonoscopy in those patients who are willing and able to undergo the procedure. Additionally, while it may be used adjunctively or in patients noncompliant with screening recommendations, it is not a screening tool for individuals at increased risk for developing disease.”

Test Description

ColoSure™ (Laboratory Corporation of America, http://www.labcorp.com ) is currently the only commercially or clinically available fecal DNA test marketed for CRC screening in the U.S. The at-home test requires that patients collect and mail one whole stool sample. The test was developed by the Laboratory Corporation of America (LabCorp), which required licensing intellectual property from Exact Sciences Corporation ( www.exactsciences.com ). As a laboratory-developed (“home-brewed”) test, ColoSure is not subject to regulation by the U.S. Food and Drug Administration (FDA) and has not obtained FDA clearance or approval.

What is the theory behind stool DNA testing? Colorectal cancer cells, which are shed into the feces, are known to have several genetic alterations which offer an array of molecular targets for DNA-based stool testing for both pre-cancerous and cancerous lesions [7] [8]. Consequently, fecal DNA has been explored for its potential as a non-invasive CRC screening methodology.

ColoSure is a single-marker test that detects methylation of the vimentin gene. Increased DNA methylation in the promoter region of genes is an epigenetic change that is common in human cancers, including colorectal cancer [9] [10]. Vimentin is a protein characteristically expressed in cells of mesenchymal origin, such as fibroblasts, macrophages, smooth muscle cells, and endothelial cells. Studies have demonstrated that the vimentin gene is not (or rarely) methylated in normal colonic epithelial cells, but is methylated in colorectal cancer and adenomas [11] [12] [13]. Aberrant methylation of vimentin has been detected in 53-83% of colorectal cancer tissue, 50-84% of adenoma samples, and 0-11% of normal colon tissue samples [11] [12] [13] [14] [15], though one preliminary study detected methylated vimentin in 29% of normal colon tissue [16].

ColoSure requires a prescription for testing. It is currently available from two sources: LabCorp [17] and from DNA Direct’s Genomic Medicine Institutes (which only offers referrals to physicians who can prescribe the test) [18] [19].

Public Health Importance

Colorectal cancer (CRC) is currently the third leading cancer diagnosed in the United States, where the lifetime risk is approximately 5% in the general population [3] [7]. According to the U.S. Cancer Statistics, ~143,000 cases of CRC occurred in the U.S. in 2007 [20]. CRC is also the second leading cause of cancer-related deaths in the United States, with approximately 53,000 deaths occurring in 2007 [20]. Approximately half of colorectal cancers are diagnosed at a late stage, when survival is poorer [4].

The most effective way of reducing the risk of developing CRC and of reducing CRC mortality is early detection and removal of pre-cancerous or cancerous lesions. It is thought that the natural history of CRC development takes between 10 and 20 years, offering an excellent opportunity for early intervention [1] [21].

Three types of tests (colonoscopy, flexible sigmoidoscopy, and fecal occult blood tests) are currently recommended as evidence-based CRC screening options by the U.S. Preventive Services Task Force [1]. However, only a modest percentage of adults meet the recommended CRC screening guidelines [4].

Stool-based DNA tests are suggested by some experts as another option for CRC screening. However, these tests are under rapid development and research to establish analytic validity, clinical validity, and clinical utility within the general (average-risk) population is needed before any fecal DNA test can be integrated into current CRC screening strategies. We now examine these factors for the ColoSure test based on the current literature.

Published Reviews, Recommendations and Guidelines (see Table 1 below)

Important Note: The following groups considered fecal DNA testing in general, but largely based their recommendations and guidelines on published research relevant to stool DNA tests that are no longer commercially available.

Systematic evidence reviews

The Agency for Healthcare Research and Quality (AHRQ) commissioned an evidence report/technology assessment on enhancing the use and quality of CRC screening [22] [23]. They found no reliable data among the included studies concerning the trends in use or quality (evidence of misuse, overuse, or underuse) of fecal DNA testing.

A systematic evidence review was performed that guided the current recommendations on CRC screening by the U.S. Preventive Services Task Force (USPSTF) [5] [24] (see subsection below).

Recommendations by independent group

Fecal DNA testing was considered by the USPSTF in its most recent recommendation statement on CRC screening (see Table 1 below). The USPSTF found insufficient evidence to evaluate the benefits and harms of this kind of testing as a screening modality for CRC (I statement) [1].

Guidelines by professionalgroups (in order by year of publication)

A Joint Guideline was published in 2008 by the American Cancer Society, the U.S. Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology (ACS-USMSTF-ACR) [25] . Stool DNA testing in general was recommended for those aged 50 years or older, but the testing interval could not be determined. In 2009, the American College of Gastroenterology (ACG) published CRC screening guidelines, in which a weak recommendation (Grade 2B) was made for stool DNA testing every 3 years for persons 50 years of age or older [26]. The American Academy of Family Physicians (AAFP) [27] published recommendations in 2010, which deferred to the analysis and findings of the USPSTF. Also in 2010, the guidelines of the National Comprehensive Cancer Network (NCCN) [28] stated that: 1) stool DNA testing is not currently considered a first-line screening test except in specific circumstances; and, that 2) the testing interval is uncertain.

Health Plan/Payer policies

CRC screening guidelines have been issued by Kaiser Permanente [29], Aetna, Inc. [30], and United Healthcare Group [31][32], all of which describe fecal DNA screening as experimental and not recommended for use.

A summary of all mentioned recommendations and guidelines appear in Table 1 below.

Table 1. Routine Colorectal Cancer Screening Guidelines and Recommendations for average-risk adults.

CTC = CT colonoscopy; DCBE = double-contrast barium enema; FIT = fecal immunochemical test; FOBT = fecal occult blood test; FSIG = flexible sigmoidoscopy; sDNA = stool DNA.

NM = not mentioned; NR = Not Recommended. ¹ In combination with high-sensitivity FOBT every 3 years.

Evidence Overview

We performed literature searches (through PubMed and Ovid MEDLINE) that included search terms such as “vimentin”, “fecal DNA”, and “colorectal cancer”.

Analytic Validity: Test accuracy and reliability in measuring methylated vimentin (analytic sensitivity and specificity).

Summary: the analytic validity of the ColoSure test could not be determined from the identified research.

Clinical Validity: Test accuracy and reliability in detecting colorectal cancer or adenomas (clinical sensitivity and specificity; predictive value).

Table 2. Summary of the published case-control studies relevant to ColoSure that reported measures of clinical validity using fecal DNA testing in selected populations

DY = refers to a specific test for DNA integrity.

– Not measured.

¹ Refers to adenomas ≥ 1 cm.

² We calculated the numerator using data presented in the article.

³ In the study, sensitivity and specificity were calculated using optimal cutpoints based on the combined dataset (Phases 1a + 1b).

⁴ We calculated specificity using data presented in the article.

Summary: the clinical validity of methylated vimentin as a biomarker for CRC screening remains to be determined in a general (average-risk) screening population. This is re-iterated in the LabCorp technical review for the ColoSure test [6], which states that: “The detection rates for general population screening have not been determined.”

Clinical Utility : Net benefit of test in improving health outcomes

Summary: the clinical utility of ColoSure (or methylated vimentin in general) in an average-risk screening population could not be determined from the identified research.

Final important note:

Fecal DNA tests are under rapid development. Exact Sciences Corporation has developed several approaches to fecal DNA testing for colorectal cancer screening over the past few years. Previous tests were replaced sequentially with newer versions, which differed in laboratory methodology or tested for a different panel of DNA markers. The current ColoSure test is a replacement of a version of the PreGen-Plus™ test (Laboratory Corporation of America), which has been discontinued. Exact Sciences recently reported results from a validation study of its newest stool-based DNA test for colorectal cancer screening, named Cologuard™. The panel that was presented included methylated vimentin as one of the tested markers [42] [50] [51]. Exact Science is currently funding a case-only study [ NCT01260168 ] to determine the sensitivity of this new multi-marker DNA panel in CRC cases. The company is planning to pursue FDA approval for Cologuard in 2012 [50]. These developments likely mean that ColoSure will be replaced in the future by this, or other, tests.

Concluding remarks:

In order to consider integrating fecal DNA testing into current CRC screening strategies, additional research is needed to establish analytic validity, clinical validity, and clinical utility within the general (average-risk) population. The estimates of DNA marker sensitivity and specificity found from small case-control studies should not be extrapolated to make any estimates of the performance of methylated vimentin or ColoSure in the general population.

In addition, the ongoing development and refinement of stool DNA tests presents some difficulty for the integration of these tests as a CRC screening approach. Currently, only one fecal DNA test is commercially available in the U.S., a test that will likely be replaced by a newer version for which FDA approval will be sought.

Other critical matters must also be addressed, including the determination of cost-effectiveness, optimal testing intervals, and strategies for the follow-up evaluation of patients who test positive on a fecal DNA test. Moreover, the willingness of individuals from the general population to adopt fecal DNA test protocols and future screening recommendations is a vital consideration. All of these factors will be crucial in affecting the impact of fecal DNA testing on the overall CRC screening paradigm and on colorectal cancer incidence and mortality.

Links (not referenced above)

Last updated: March 14, 2011

Acknowledgments

The authors would like to thank the following individuals for invaluable input and guidance on the content of this manuscript: Dave Dotson, Ralph Coates and Katie Kolor (Office of Public Health Genomics, CDC); Lisa Richardson and Djenaba Joseph (Division of Cancer Prevention and Control, National Center for Chronic Disease Prevention and Health Promotion, CDC); and Evelyn Whitlock, Beth Webber, and their colleagues (Kaiser Permanente Center for Health Research).

Funding information

This work was funded by the Office of Public Health Genomics, Centers for Disease Control and Prevention.

Competing interests

The authors have declared that no competing interests exist.

Disclaimers

The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention (CDC).

The information provided in this manuscript does not constitute an endorsement of ColoSure(TM) or of any fecal DNA test by the CDC nor the Department of Health and Human Services (DHHS) of the U.S. government. No endorsement should be inferred.

The CDC does not offer medical advice to individuals. If you have specific concerns about your health or genetic testing, we suggest that you discuss them with your health care provider.