A Validation Approach for Determining Fetal Blood Groups Non-Invasively by High-Sensitive Next-Generation Sequencing

Abstract

Introduction: For pregnant women with a history of fetal and neonatal alloimmune thrombocytopenia (FNAIT) or hemolytic disease of the fetus and newborn (HDFN), prenatal intervention in subsequent pregnancies may be necessary to prevent complications for the fetus. A non-invasive prenatal diagnostic procedure (NIPD) is recommended for fetal blood group genotyping. RT-PCR is used for fetal RHD determination as a reliable screening method with high sensitivity and specificity. For other antigens with variants involving single-base substitutions, droplet digital PCR (ddPCR) and next-generation sequencing (NGS) are recommended to reduce the risk of false-negative results. Only NGS offers the possibility of determining the cell-free fetal DNA (cffDNA) fraction in maternal plasma by sequencing additional gene fragments in parallel, but no standard exists for assay validation. Material and Methods: A custom-made primer panel was designed to target the common platelet and red cell antigens involved in fetal red cell and platelet incompatibilities, as well as additional anonymous single-nucleotide polymorphism (SNP) targets for use as an internal control. Amplicon-based NGS was carried out using semiconductor sequencing. For HPA-1a (HPA1A, ITGB3) and K (KEL01.01, KEL) assay validation, the limit of detection (LOD) and limit of quantification (LOQ) were estimated, as were false-positive antithetic alleles, linearity, and inter-assay variation, using cell-free DNA (cfDNA) extracted from the blood samples of healthy blood donors. An additional analysis was performed using 23 diagnostic samples from 21 pregnant women. Results: Regression analysis of dilution series using HPA-1a- and K-positive cell-free plasma samples in antigen-negative donor plasma showed that recovery is definitely feasible up to an HPA1A and KEL01.01 allele frequency of 1%. Base calls of false-positive antithetic alleles were detected with a maximum of 0.25% using 21 healthy blood donors. The LOD was estimated to be 0.2057% (mean + 3 SD) for HPA1A with a LOQ of 0.6298% (mean + 10 SD). For KEL01.01, the LOD was 0.1706% (mean + 3 SD) and the LOQ was 0.5314% (mean + 10 SD). The analysis of 15 of 21 cases with diagnostic samples from pregnant women with neonatal blood available for confirmatory testing resulted in 100% concordant results. The fetal fraction of these samples was calculated with a median of 11.03% (95% CI: 8.89, 13.20). Conclusions: NGS for non-invasive fetal blood group genotyping is an accurate and reliable method. In-house validation of the used assays can be performed using healthy donors to determine the LOD, LOQ and sensitivity. The threshold for paternally inherited fetal HPA1A and KEL01.01 alleles could be set at 1% (i.e., 2% fetal fraction) to obtain reliable test results. Internal controls for assessing the fetal fraction are essential to avoid false-negative test results.