High-throughput fetal fraction amplification increases analytical performance of noninvasive prenatal screening

Abstract

Purpose: The percentage of a maternal cell-free DNA (cfDNA) sample that is fetal-derived (the fetal fraction; FF) is a key driver of the sensitivity and specificity of noninvasive prenatal screening (NIPS). On certain NIPS platforms, >20% of women with high body mass index (and >5% overall) receive a test failure due to low FF (<4%).

Methods: A scalable fetal fraction amplification (FFA) technology was analytically validated on 1264 samples undergoing whole-genome sequencing (WGS)-based NIPS. All samples were tested with and without FFA.

Results: Zero samples had FF < 4% when screened with FFA, whereas 1 in 25 of these same patients had FF < 4% without FFA. The average increase in FF was 3.9-fold for samples with low FF (2.3-fold overall) and 99.8% had higher FF with FFA. For all abnormalities screened on NIPS, z-scores increased 2.2-fold on average in positive samples and remained unchanged in negative samples, powering an increase in NIPS sensitivity and specificity.

Conclusion: FFA transforms low-FF samples into high-FF samples. By combining FFA with WGS-based NIPS, a single round of NIPS can provide nearly all women with confident results about the broad range of potential fetal chromosomal abnormalities across the genome.

Keywords: analytical validation; body mass index; cell-free DNA; fetal fraction; noninvasive prenatal screening.

Fig. 1. Fetal fraction amplification (FFA) technology increases fetal fraction (FF) across all body mass index (BMI) levels.

For 2401 patients who indicated BMI values on the test requisition form, the FF levels measured without FFA (gray circles) and with FFA (purple triangles) are plotted as a function of patients’ BMI values (vertical axis). The top panel plots the histogram of samples with or without FFA. The dotted line is at 4% FF, the threshold below which the American College of Medical Genetics and Genomics (ACMG) considers a sample to have low FF. Three samples are highlighted as illustrative examples that had low FF before FFA but normal FF after FFA. NIPS noninvasive prenatal screening.

Fig. 2. Fetal fraction amplification (FFA) increases FF for 99.8% of samples tested and most appreciably for low-FF samples.

The fold-change difference in FF as a result of applying FFA is plotted for individual samples as a function of the original FF without FFA. The dashed line indicates no change in FF, and samples in the purple-shaded region had increased FF with FFA.

Fig. 3. Fetal fraction amplification (FFA) improves detection of fetal chromosome abnormalities by amplifying the signal of aneuploid regions while maintaining background noise.

(a) Schematic of the change in median depth per autosome as a result of FFA. The extent of the deviation from background is itself a measure of FF and is indicated as FF_positive. (b) The increase in FF_positive without FFA (gray circles) and with FFA (purple triangles) is shown for aneuploid samples with the indicated chromosome anomalies. (c, d) z-scores without FFA (gray) and with FFA (purple) for the same samples as in (b) are stratified by their screening results and summarized either as population distributions (c) or as individual samples (d). For visual clarity in (c), the distribution of screen-negative samples (NEG.; dashed line) has been scaled to be of comparable height as the screen-positive distributions to the right (solid lines). The vertical solid line indicates the z-score cutoff between screen-negative (left) and screen-positive (right) results. For SCAs, only female fetus pregnancies are shown (i.e., MX and TX) because a z-score is used to identify chrX aneuploidies, whereas a two-dimensional analysis that does not use z-scores (not shown) is required for identification of XXY and XYY (FF_positive increased in all XXY and XYY pregnancies tested with FFA). NIPS noninvasive prenatal screening, RAA rare autosomal aneuploidies, SCA sex chromosome aneuploidies.

Fig. 4. Fetal fraction amplification (FFA) increases the analytical accuracy of fetal sex calling.

For both standard noninvasive prenatal screening (NIPS) (top) and Prequel with FFA (bottom), the distribution of FF_chrY values is shown for samples called as female (green) or male (blue). Solid lines indicate raw data, and the dashed lines show best-fit traces for the female (Gaussian) and male (beta) populations. Only euploid samples are included in the analysis. The orange arrow depicts one sample tested on both platforms, called female in standard NIPS and male with FFA (the fetus was confirmed to be male). After minimizing the number of estimated miscalls on each platform (see “Materials and methods”), analytical miscalls are predicted to drop 318-fold with FFA.