Palmetto Lays Out Analytical Validity Requirements for Liquid Biopsy

Medicare contractor Palmetto GBA (Columbia, S.C.) outlined its coverage criteria for assessing the analytical performance of liquid biopsy tests to detect genetic variants in tumors. The requirements are limited to liquid biopsies using circulating tumor DNA (ctDNA), but can be based on any assessment technology, including next-generation sequencing. Palmetto says the coverage criteria are intended to “complement,” but not substitute for MolDX, Palmetto’s unique registration system for molecular diagnostic tests.

Central to the coverage criteria is Palmetto’s designation of ctDNA-based somatic variant detection tests as a “form of surrogate testing,” the contractor explained. Palmetto explains that the ctDNA “analyte” is the surrogate for the “tumor tissue analyte.” So, as part of analytical validation considerations, laboratories must both establish performance metrics for the surrogate analyte and demonstrate the agreement between the surrogate and the reference tumor analyte.

The lab must specify its minimum requirements in terms of, for example, cell-free DNA, library concentration, and coverage.

Pre-analytically, the requirements say the laboratory must specify its minimum requirements for reporting results, including minimum library concentration, minimum average coverage, and uniformity across all target areas. Additionally, the test must be validated for all sample types and preservatives accepted for testing.

Analytical requirements say that four general types of sequence variants may be evaluated using liquid biopsy, including single nucleotide variants, insertions and deletions (indels), copy number amplifications, and structural variants (e.g., fusions, inversions and translocations). Laboratories only need to address requirements for molecular alterations that they report out, but the full reportable range of analysis and quality control parameters established in validation studies must be clearly specified, including variant allele frequency, minimum read/molecule counts, indel length, copy number level, and test target regions.

Additionally, only variants that can be detected by both the ctDNA assay and the independent reference method should be included in the analyses, but the technology platform and/or sequencing chemistry used for the orthogonal reference method must be different than what is used for the ctDNA assay. Palmetto detailed criteria for how labs should establish a test’s positive percent agreement, analytical positive predictive value, and negative percent agreement.

Post-analytical requirements say a physician board certified in molecular genetic pathology must identify variants detected through ctDNA testing, and interpret them clinically to make therapeutic recommendations. Palmetto writes, “A Ph.D. is not a recognized Medicare provider.”

The analytical validation criteria outlined by Palmetto apply only to qualitative assessment of ctDNA performed and are not intended for the analysis of circulating tumor cells or CTC-derived tumor DNA, matched tumor-normal testing with ctDNA, or quantitative ctDNA analysis for drug response, disease monitoring, or minimal residual disease.

Reporting of “exceptions,” including incidental findings or variants below the reported limit of detection needs to “clearly state” that the test is not validated to report these results.

Interestingly, experts say this is the first time a Medicare contractor has “encouraged” reporting of somatic variants to the National Institutes of Health’s ClinVar variant clearinghouse. This is part of a trend on the part of payers. It was recently reported that Aetna is requiring that all newly contracted laboratories submit their variant interpretations to ClinVar as a condition of participation in the Aetna network for BRCA1 and BRCA2 genetic testing.

Palmetto has Jurisdiction M covering North and South Carolina, Virginia and West Virginia

Takeaway: Liquid biopsy testing continues to attract attention with more detailed coverage criteria from Palmetto.