Response to Comments: MolDX: Repeat Germline Testing

This comment was received by multiple contractors, as such it will only be presented here once.

Thank you for the opportunity to review and comment on [MolDX Contractor]’s proposed coverage policy for Repeat Germline Testing DL38353 and DL38351 (hereafter ‘draft LCD’). Members of the Association for Molecular Pathology (AMP) and the College of American Pathologists (CAP) are submitting joint comments at this time because both organizations share the same perspective regarding this draft LCD.

The AMP is an international medical and professional association representing approximately 2,300 physicians, doctoral scientists, and medical technologists who perform or are involved with laboratory testing based on knowledge derived from molecular biology, genetics, and genomics. Membership includes professionals from the government, academic medicine, private and hospital-based clinical laboratories, and the in vitro diagnostics industry.

The CAP is the world’s largest organization of board-certified pathologists and the leading provider of laboratory accreditation and proficiency testing programs. The CAP serves patients, physicians, hospitals and healthcare systems worldwide, fostering and advocating excellence in the practice of pathology and laboratory medicine.

We commend [MolDX Contractor] for recognizing the vital importance of providing coverage to Medicare beneficiaries for germline testing and for your thoughtful approach to repeat testing of panels that contain non-duplicative test components that demonstrate clinical utility. We agree that repeated testing of an individual's genome for inherited diseases is generally not reasonable and necessary. However, improved testing methodologies and knowledge base continue to expand and improve information that can be returned from a genetic test. Therefore, we offer the following comments and recommendations for Noridian’s consideration.

1. dLCD statement: “Germline testing using gene panels that contain some genetic content that has already been tested in the same Medicare beneficiary may be considered reasonable and necessary provided that there is established clinical utility present in the remaining, non-duplicative genetic components of the test.”

Comment: Next Generation Sequencing technology has allowed for simultaneous testing of multiple genes in panel tests, greatly expanding the possibility of finding a disease-associated variant to explain predispositions to inherited cancer and germline variants that can affect treatment choices such as the use of PARP inhibitors in breast and ovarian cancer. The LCD’s current statement correctly provides for scenarios to repeat germline testing to identify inherited breast and ovarian cancer syndromes. For example, tests for inherited breast cancer used to only examine BRCA1/2, but use of a larger cancer panel has increased the diagnostic rate significantly from 2.5% (BRCA1/2 alone) to 6.3% using a guideline panel (11 genes) and 9.4% using a large cancer panel (80 genes) in a cohort of 959 patients (Beitsch et al. 2019).

Although Next Generation Sequencing (NGS) methods have become widely employed for diagnosis and treatment, rapid gains in methodology and gene data base are still being made. Modifications to protocols, techniques and instruments for NGS may provide significant improvements in detection of germline variants associated with inherited cancer today compared to earlier tests. Methods and knowledge will continue to improve, hence re-testing with newer methods or re-analysis of existing sequencing data may be needed to successfully detect the following clinically relevant variants:

a. Mosaic variants: Genetic testing is typically performed on DNA extracted from either blood or saliva samples. - Epilepsy: Most genetic causes of epilepsy are inherited in an autosomal dominant or X-linked manner.

• Tuberous sclerosis: Tuberous sclerosis complex (TSC) is caused by pathogenic variants in TSC1 or TSC2. Although 85-90% of individuals with a clinical diagnosis of TSC have a pathogenic variant identified by routine genetic testing, 10-15% do not. A significant proportion of those individuals are mosaic for a pathogenic variant that is detectable by testing of a different sample type or sequencing to a greater than typical depth of coverage (Tyburczy et al. 2015, Byers et al. 2018).
• Vascular and overgrowth disorders: CLOVES and Klippel-Trenaunay syndromes, caused by mosaic pathogenic variants in PIK3CA, and Proteus syndrome, caused by mosaic pathogenic variants in AKT1, may require NGS testing on different samples before low level mosaicism is identified (Lindhurst et al. 2011, Keppler-Noreuil et al. 2015, Luks et al. 2015). These sample types are most convenient for providers to collect and least invasive for patients. For a subset of disorders where mosaicism occurs, repeat testing with a biopsy sample from an affected tissue and /or sequencing to a greater depth of coverage is clinically indicated as mosaicism may not be well represented in blood alone.

Note, each of these disorders has clinical features that overlap other genetic syndromes diagnosable by routine testing or are themselves often diagnosed by routine testing; therefore, methods developed for detection of mosaicism would not be a first-line genetic test.

Detecting mosaicism increases diagnostic yield in several diseases including:

Although most cases are either inherited or occur de novo as heterozygous or hemizygous variant, in a study of ~1,000 cases, 1.4%-3.4% were found to be mosaic (Stosser et al. 2018, Burgess et al. 2019). There are many identified genetic causes of epilepsy and a molecular diagnosis is valuable in guiding treatment (Mei et al. 2017, Orsini et al. 2018).

b. Copy number variants: Similarly, detecting copy number variants (CNVs) has improved diagnosis in several disease types such as inherited cancer syndromes (8.3%) (Truty et al. 2019), pediatric disorders (7.7%) (Truty et al. 2019), cardiovascular disease (4.7%) (Truty et al. 2019), neurologic disorders (35%) (Truty et al. 2019), muscular dystrophy (0.8%, 7/793) (Valipakka et al. 2017), and various genetic disorders associated with CNVs detected by exome sequencing (15%, 8/54; 1.6% 12/693)(Marchuk et al. 2018, Gao et al. 2019).

Please note, these methods continue to improve. Recently reported methods attain the best result when sequence read depth is greater than typically used for routine testing (Ellingford et al. 2017, Kerkhof et al. 2017, Yao et al. 2017). Based on the limitations identified by these studies, older NGS data will be of insufficiently consistent quality to call CNVs; therefore, resequencing of previously tested genes will be the technically superior and most economical way to detect CNVs.

c. Pathogenic tandem repeat expansions: Expansions of repetitive base-pair motifs are known to cause a number of inherited neurological or muscular disorders such as Fragile X syndrome, Huntington disease, myotonic dystrophy and ataxias (Paulson 2018). Making a specific diagnosis requires molecular testing due to overlapping clinical features of these disorders. A specific diagnosis is important to patient care when specific treatments are indicated or contraindicated, as is the case with certain cerebellar motor dysfunction and ataxia disorders (Zesiewicz et al. 2018). Repetitive regions are difficult to sequence and to align to reference sequences with current methods. Technical improvements in short read sequencing quality (Dolzhenko et al. 2017) or in long read sequencing (Liu et al. 2017) generate such data. Improved informatic techniques allow such variants to be detected and characterized (Tankard et al. 2018). Such technological advances will be the result of improved methods and, therefore, require re-testing of patient samples rather than reanalysis of existing data. As expanded panel tests become more widely adopted it is imperative that coverage policies keep pace to reflect quality health care for Medicare beneficiaries.

d. Pathogenic variants in an expanded reportable range: As advanced sequencing methodologies improve they provide higher accuracy and greater reportable range for regulatory and deep-intronic gene variants and the pathogenic splicing alterations that result from them. Many labs do not report intronic variants >5bp from the exon, but pathogenic variants may exist in deeper intronic areas and have been shown to lead to alternative splicing patterns and/or disrupt normal splicing and lead to a truncated or absent protein. - GJB1, pathogenic promoter and non-coding exon variants are associated with Charcot-Marie-Tooth neuropathy X type 1 (Tomaselli et al. 2017).

• FDFT1, a pathogenic promoter variant is associated with squalene synthase deficiency (Kulshrestha et al. 2017, Coman et al. 2018).
• DMD, multiple deep intronic pathogenic variants have been associated with Becker / Duchenne muscular dystrophy (Trabelsi et al. 2014, Zaum et al. 2017).

Examples of genes with clinically relevant promoter or deep intronic variants include:

It is reasonable to conclude that a pathogenic variant in a typically untargeted region could be described in any gene associated with disease, prompting resequencing of the gene.

e. Complex variants: As sequencing technology improves, disease associated sequence variants not detected by current methods, such as reciprocal translocation breakpoint sequence, inversions, insertions, and complex chromosomal rearrangements may be detectable (Chatron et al. 2019, Schluth-Bolard et al. 2019). The sequencing methodology required would be substantially different necessitating retesting rather than reanalysis. Although such tests are likely to fall into the category of genome sequencing, it is conceivable that gene panels or gene-specific clinical tests could be offered using such methods.

Recommendation: We recommend that [MolDX Contractor] modify its coverage policy to reflect the following: A germline test is usually performed only once in a lifetime per beneficiary for inherited conditions. However, when medically reasonable and necessary, repeat testing may be allowed as follows: when additional implicated genes are included in the test; when the reportable range in genes already tested has been expanded to encompass pathogenic variants for which there is sufficient evidence for clinical testing; or when the analytic sensitivity has improved since the time of the previous test to allow the detection of difficult variants such as mosaic variants, copy number variants or triplet repeat alleles.

2. dLCD statement: "Providers should take reasonable measures to be aware of what, if any, germline testing a beneficiary has had prior to billing for germline testing so as to avoid billing Medicare for services that are not reasonable and necessary. Clinicians who order germline testing may wish to be aware of whether the test that they are ordering is covered under Medicare and may wish to verify that they are not ordering repeat germline testing.”

Comment: [MolDX Contractor] does not define what is meant by ‘reasonable measures.’ Most ordering providers will not know the reportable range of a genetic test, as that information is not likely to be in a patient’s report. Additionally, if the original test was performed by a laboratory that is no longer in business, it may be impossible to obtain details about the methodology used and analysis done. In the absence of an original clinical report and / or consultation with the laboratory, a clinician should not make assumptions about the testing performed. As more data accumulates over time, a patient may need a “variant of uncertain significance” (VOUS or VUS) to be re-evaluated, as doing so may obviate the need for additional testing if it is upgraded to likely pathogenic or pathogenic. Variant reinterpretation would consider the most up to date knowledge and require no additional technical services. Recent studies have shown utility in variant reinterpretation in inherited cancer syndromes, epilepsy, and cardiomyopathies (Aronson et al. 2012, Walsh et al. 2017, Mersch et al. 2018, SoRelle et al. 2019).

Recommendation: We recommend that [MolDX Contractor] remove this requirement from its final policy. Alternatively, we recommend modifying the LCD language to reflect the following: “Providers should take reasonable measures to be aware of what, if any, germline testing a beneficiary has had prior to billing for germline testing to avoid billing Medicare for services that are not reasonable and necessary. Any testing for which prior results are not reasonably discoverable by the beneficiary’s treating providers cannot be considered as duplicative. Providers will not be held responsible when the original report lacks information about previous testing that is needed to prevent duplicate future tests.”

We would like to thank AMP and CAP for these thoughtful and informative comments. We will address the two comments below:

Response to first comment:

Our opinion is that the exceptions presented do not fall within scope of the intent of the limitations or preclusions from coverage noted in this policy, which is intended to prevent unnecessary duplicative testing. The above illustrated rare conditions may require additional testing beyond the routine, sometimes of the same gene; however we do not believe these would be classified as testing “the same genetic content for the same information.” In each case above, either the content is different (looking at different genes or introns) or the information sought is different (different variant types such as CNVs, mosaicism, or variants only found with newer techniques). In any case, an argument could be made that the testing is non-duplicative and thus outside the scope of this policy.

We have modified the language in the policy to clarify this point, by explicitly limiting coverage to “duplicative” efforts.

Please review the associated Billing and Coding Article for this policy, as it defines a process to provide coverage where the testing is non-duplicative but cannot easily be differentiated with existing billing and coding systems.

Response to second comment:

Regarding re-interpretation, we do not agree that this falls within the scope of this policy as this LCD is specific to genetic laboratory testing.

We do agree that if prior testing reports cannot be reasonably found, or if the original test was performed but the lab cannot be contacted to consult about an existing report, a subsequent test should not be considered duplicative. Because the testing is not duplicative, it again would not fall within scope of the preclusions set by this policy. Please review the associated Billing and Coding Article, which has a provision to allow for coverage in these situations. In these exceptional situations it is expected that the provider (possibly with the assistance of the ordering physician) provide a rationale for the need for additional testing and submit for a medical review.

This comment was received by multiple contractors, as such it will only be presented here once.

The American Clinical Laboratory Association (ACLA) submits these comments on the Proposed Local Coverage Determination for MolDX: Repeat Germline Testing (DL38353).

1. ACLA is a non-profit association representing the nation’s leading clinical and anatomic pathology laboratories, including national, regional, specialty, end-stage renal disease, hospital, and nursing home laboratories. In the United States, the clinical laboratory industry employs nearly 277,000 people directly and generates over 115,000 additional jobs in supplier industries. Clinical laboratories are at the forefront of personalized medicine, driving diagnostic innovation and contributing more than $100 billion annually to the nation’s economy.

Palmetto’s MolDX program proposes that “any laboratory test that investigates the same germline genetic content, for the same genetic information, that has already been tested in the Medicare beneficiary” is not reasonable or necessary and, therefore, not covered. It proposes that a Medicare beneficiary “may only be covered for one test per lifetime” for germline tests and test panels in the range of CPT codes 81105 through 81479. The draft policy also includes: “Tests billed under the CPT code 81479 that are used for germline testing are also limited to once per lifetime.” Palmetto suggests that “[p]roviders should take reasonable measures to be aware of what if any germline testing a beneficiary has had prior to billing for germline testing so as to avoid billing Medicare for services that are not reasonable and necessary.”

A. Same germline genetic content for same genetic information

ACLA agrees that when a laboratory knows or takes reasonable measures to know whether a Medicare beneficiary already has had a germline genetic test for particular genetic information (for example, if the laboratory itself has performed the test for the beneficiary before), it would not be reasonable or necessary for a beneficiary to have the same test again. However, in many cases, it is extremely difficult for a clinical laboratory to determine if a germline genetic test for which it receives an order from a physician or non-physician practitioner (NPP) is a “repeat germline test.” A clinical laboratory oftentimes has no face-to-face interaction with a patient, whose specimen is collected elsewhere and forwarded to the laboratory. If a laboratory has not tested a particular beneficiary before, it may be able to learn from the ordering physician or NPP whether the beneficiary already has been tested. But the ordering physician or NPP may not have inquired whether the test is a “repeat germline test”, or provide that information to the laboratory, or respond to the laboratory’s request at all. The laboratory may not have a way to contact the beneficiary to ask whether it is a repeat test.

It is the ordering physician or NPP who is in a position to determine whether a beneficiary previously had a germline test for particular genetic information—and yet it is the laboratory that is not paid for its services if the physician or NPP orders a “repeat germline test.” It is both impractical and inappropriate for a laboratory to request—and wait for—information from the physician or NPP about a beneficiary’s previous germline testing prior to performing an ordered test. This is so for a number of reasons. First, some specimens begin to degrade within hours of being obtained. Second, much laboratory testing is performed overnight, when clinicians’ offices typically are closed for business. Third, the sheer volume of testing performed by many laboratories makes it impossible to obtain the information prior to every test being performed timely for a Medicare beneficiary. Waiting for a response from the ordering physician or NPP before performing a test could jeopardize the validity of the test results and delay treatment decisions by the ordering clinician, placing a Medicare beneficiary’s health at risk. Such a requirement would put laboratories in the position of having to perform tests for which they may not be paid, based on the ordering physician’s or NPP’s failure to inquire whether the test was necessary in the first place.

In light of the foregoing, ACLA believes it would be reasonable for the MolDX program to deny payment for a test that investigates the same germline genetic content, for the same genetic information, when the laboratory itself performed the test for the beneficiary previously, but we do not agree that the laboratory should be penalized for performing a test that is ordered by a physician or NPP who should be responsible for determining whether the test is medically necessary for a particular beneficiary. Unless and until the MolDX program develops a mechanism to hold an ordering clinician accountable for his or her failure to discern whether the test is needed, it should not finalize this policy as proposed and penalize the laboratory for furnishing a test pursuant to a valid order.

B. Coding Information Section

ACLA requests that the MolDX program amend the Draft LCD’s language in the Coding Information Section. That section includes this statement: “For the following CPT codes associated with germline testing, a beneficiary may only be covered for one test per lifetime…Tests billed under the CPT code 81479 that are used for germline testing are also limited to once per lifetime.” As written currently, the policy could be interpreted to preclude coverage of a test described by CPT code 81162 (BRCA) because a beneficiary previously had a test described by CPT code 81288 (MLH1), which we do not believe is the intention of the policy. Nor do we believe that the MolDX program intends to deny a Medicare beneficiary a medically necessary germline test for X when the claim was submitted under CPT code 81479, simply because the beneficiary had another germline test for Y and the claim also was submitted using this code. The code descriptor for CPT code 81479 is: “Unlisted molecular pathology procedure.” This code can describe any number of germline tests for different genetic content and different purposes.

CMS should amend the language to read: “A germline test for which a provider or supplier submits a claim using the following CPT codes may be covered for a beneficiary only if the provider or supplier submitting the claim has not tested the beneficiary with the same germline test for the same genetic information previously.” This would clarify the policy and limit the effect of the policy’s non-coverage to those instances where a laboratory itself has constructive knowledge that the beneficiary has had the same test for the same purpose before, because the laboratory itself furnished the test.

We would like to thank ACLA for their thoughtful comments.

Regarding the first comment:

We appreciate the concern and understand that clinical laboratories often do not have access to a majority of the clinical record to be able to discern of the test they are being asked to perform is duplicative. However, the function of the Medicare contractors is to provide coverage for services that are reasonable and necessary. As this policy outlines that duplicative germline testing is not reasonable and necessary, such services cannot be covered.

Reference laboratories are encouraged to work with clinicians in finding solutions to prevent duplicative testing and communicating these requirements.

Regarding the second comment:

We agree with ACLA on the intent of the policy is not to deny coverage to non-duplicative germline testing. Due to changes in Chapter 13 of the Program Integrity Manual effective January 9^th 2019, billing and coding information is no longer to be within the language of the LCD and placed in an associated Article. Unfortunately, at the time of Open and subsequent Comment periods the technology was not yet in place to further refine the language as intended for this Billing and Coding Article within the LCD, leading to an incomplete explanation of billing and coding under this policy.

Please review the associated Billing and Coding Article which addresses the issues presented above. Specifically, while 81479 falls within scope of this policy, the use of 81479 is itself not limited to once per lifetime but associated only with germline-specific Z-identifier codes.