Biology professor Erin Lindquist was in the middle of class when she got the phone call telling her that a prenatal test had returned abnormal results. Lindquist, who was 36 and pregnant with her second child, had had noninvasive prenatal testing (NIPT), which looked for abnormal chromosomes in her blood. Since about 10 percent of DNA in maternal blood comes from the fetus, this test is used to detect fetal genetic abnormalities. The results were “very devastating,” says Lindquist: an extra copy of chromosome 13 and a missing copy of chromosome 18.
As recommended, she immediately followed up with amniocentesis, a more invasive test. The amniocentesis showed that the fetus was normal: the NIPT results had been a false positive. Her son Jaxson was born healthy on October 4, 2012.
Less than 48 hours after Jaxson’s birth, Lindquist was diagnosed with cancer, eventually identified as an advanced form of a rare cervical cancer that had metastasized to her bones. The abnormal chromosomes in the prenatal test had been hers, not her son’s. Hers was the first case, published in a 2013 paper in Prenatal Diagnosis, in which a mother’s cancer was visible in prenatal test results.
On Monday, the Journal of the American Medical Association published a paper showing that noninvasive prenatal tests have pointed to several more cancer cases. The authors looked at prenatal testing data from eight mothers with cancer diagnoses (including Lindquist) to see if signs of the cancer were visible in those test results.
In three of the eight women, the prenatal test results led directly to a medical workup that caught the cancer (in the other five women, including Lindquist, other symptoms led to the diagnosis). The paper also demonstrates that in principle, a “liquid biopsy” test could catch cancer before other symptoms appear. Of the companies that offer the prenatal tests, including Sequenom, Natera, Illumina, and Ariosa, at least Sequenom and Natera are developing cancer tests as well.
“We need to do a better job up front to communicate with patients that we might find out something about their own health as well,” says Diana Bianchi of Tufts Medical Center, the first author of the paper, who is on an advisory panel for Illumina. Women who take the prenatal test are usually not told this.
It’s also important to note that “this test was not developed as a screen for cancer,” says Bianchi. “It doesn’t detect cancer.”
A result like Lindquist’s, turning up chromosomal abnormalities even though the fetus is normal, can be explained by a number of things: the most common cause is abnormal chromosomes in part of the placenta, but it can also be a twin that died naturally in the womb, or abnormal sex chromosomes in the mother. “After we go through all those, cancer is sort of down at the bottom of the list,” says Bianchi.
The authors of the study looked through the records of Illumina, which offers prenatal testing under the name Verifi. Out of some 3,700 women with abnormal results, they found 10 women whose physicians had reported to Illumina that they had received a cancer diagnosis either before or soon after their babies were born. They asked permission from eight of the women to look at their chromosomes. (The other two were deemed too ill to contact.)
Five of the eight women, including Lindquist, had multiple instances of an entire chromosome either missing or duplicated. It’s rare for a fetus to be viable with this profile, so the authors suggest that if the test shows multiple extra or missing chromosomes, maternal cancer should be considered as a possible cause. Cancer cells can have entire missing or duplicated chromosomes because they mutate much more than normal (the noncancerous cells in the patient’s body still have normal chromosomes).
For months before her diagnosis, Lindquist had been experiencing symptoms including a growth in her cervix, pelvic pain, and bleeding, but since her Pap smears were normal, these issues were blamed on a polyp and then fibroids. “Gynecologists and genetic counselors should know” that prenatal tests can show whether a woman should be checked for cancer, Lindquist says. “It certainly would’ve helped me out if I had been diagnosed early.”
In the past year, Lindquist has undergone radiation and chemotherapy for new metastases in her brain, liver, kidney, and pancreas. When I asked what her prognosis is, she laughed and said, “I don’t get one; I don’t ask.” Her cancer is so rare that there’s not enough data for predictions.
To assess the value of screening for cancer before symptoms are present, it’s important to look at the false-positive rate for cancer—in other words, the number of times the test tells women they may have cancer when it turns out they don’t. This paper cannot address that, because it lacks data on how many women with multiple chromosomal abnormalities in their NIPT results didn’t have cancer.
“What does the clinician do next?” says Bianchi. Is it valid to inform a patient that she has cancer but no one knows where? Is it valid to order expensive, stressful tests? In a study published in JAMA Oncology in June, doctors went on to do full-body MRIs of three mothers to see if they could find visible cancer. They found a tumor in all three. “But there’s no evidence that that needs to get done,” Bianchi says.
In a comment in Nature in June, Bianchi stressed the need to develop medical recommendations and obtain consent from mothers to find out about their own health before noninvasive prenatal tests are done. She says she’s already heard of cases where genetic counselors have told women who get false positives for fetal abnormalities that they could have cancer.