Most of what genome-wide screening flags is not about the baby
In the Dutch national programme, half the extra findings were the placenta and a quarter were the mother.
The upgrade sold on top of standard cfDNA screening is usually called genome-wide. The word does a lot of work.
Here is what the current US specialist guidance says the resolution actually is: The resolution of genome-wide cfDNA screening is similar to the 5-10-Mb resolution of a standard karyotype, enabling the detection of RATs in addition to the three common autosomal trisomies.
And: Of note, genome-wide cfDNA screening will not detect the majority of microdeletion and microduplication syndromes, which are typically less than 5 Mb.
Genome-wide does not mean everything. It means roughly the resolution of a chromosome picture from the 1970s, read from placental DNA in your bloodstream.
What it flags, in a whole country. The Netherlands offered this nationally with complete follow-up. Among 110,739 pregnancies that chose the genome-wide option, 402 women received an extra finding. In the researchers' words: This implies that 1 in every 275 women opting for GW-NIPT received an abnormal result indicative of an additional finding.
Of those, 196 were flags for a rare trisomy. Here is the number nobody quotes.
Only fifteen of the 196 came from the fetus at all. Nine of those were harmful. The rest were mostly the placenta.
Take all 358 extra findings whose origin was eventually proved, and this is where they came from.
| Where the finding actually came from | Number | Share |
|---|---|---|
| The baby | 79 | 22.1% |
| The placenta only | 189 | 52.8% |
| The mother | 90 | 25.1% |
More than half of what genome-wide screening flags beyond the common trisomies is a placental finding, and a quarter is a finding about you rather than about your baby.
The pooled evidence agrees. A 2026 meta-analysis of 16 studies and 681,633 pregnancies found rare trisomy flags in 0.2% of cases, and states that 80% were false positives
, with about 35% of them associated with an adverse pregnancy outcome, mostly involving chromosomes 2, 4, 11, 16 and 22.
A placental finding is not simply an error. It is usually a real biological finding that carries a genuinely raised risk to the pregnancy. In the Dutch cohort, comparing the 189 placental cases with the general obstetric population: pre-eclampsia in 8.5% (16 of 189) against 0.5%, and birth weight below the 2.3rd percentile in 13.6% (24 of 177) against 2.5%. Those are large increases from small starting numbers. What has not been shown is that knowing this in advance improves the outcome. No trial has tested it.
What women thought afterwards. All 402 women were surveyed and 227 replied. Ninety-two percent said the news was unexpected, a shock, or hard to believe. Eighty-five percent said it caused a lot of worry. Then the question that matters most: would you choose to be told again? Among women whose finding was in the baby, 86% said yes. Among women whose finding turned out to be confined to the placenta, which is the majority of all findings, 49.2% said yes. That is a coin flip.
The quarter that is about you. Sometimes an unusual pattern in your blood reflects your own health, including a cancer nobody suspected. Across 231,896 Dutch pregnancies, 0.02% of results were assessed as suspicious for a maternal cancer. Genome-wide testing produced these at 0.03% and targeted testing at 0.005%, so choosing genome-wide is also choosing a higher chance of being told this. Of those referred for cancer screening, 16 of 48 turned out to have cancer. Where the pattern involved several different chromosomes, it was 16 of 23.
You may have seen a much larger figure. A 2024 study reported cancer in 52 of 107 people, about half. That is not a positive predictive value and it should never be used as one. Everyone in it had already been referred to a national cancer screening protocol as suspicious. The authors say so themselves: All participants were referred with concern for malignancy. Our cohort is not representative of the general pregnant population.
What that study does show usefully is which pattern matters: among 49 participants with gains and losses across three or more chromosomes, 47 had cancer. And more than half of those with cancer had no symptoms they had recognised.
Now the other direction, which is just as important. Of 65 Dutch women who were diagnosed with cancer during pregnancy and had also had prenatal cfDNA screening, only 16 had a suspicious result. About three quarters were missed. Eleven of those cancers were blood cancers, and the ten of them that had genome-wide testing were all flagged, whatever the stage. Only five solid tumours were, four of them already advanced. A US series found the same shape: no abnormal result in any stage 0 or stage I solid cancer, and 7 of 11 at stage IV.
"What would we do differently if it were the placenta?"
"Can I have the standard test instead of the genome-wide one?"