I presume this article was published to inform obstetricians about the performance of the Panorama noninvasive prenatal screening test (NIPT) sold by Natera. As a practicing obstetrician, I have been unable to take away meaningful data from this study to explain to my patients about the clinical performance of this test. The conclusions drawn have no basis in the data presented, which are incomplete, and meaningful inferences cannot be drawn.

There is no breakdown of the numbers or criteria of high-risk and low-risk populations. Without these numbers how could the positive predictive value for the tests be calculated?

The study started out with 31,030 samples received. Two hundred eighty samples were excluded for not meeting the study criteria. A total of 30,750 samples met the study criteria, but 2011 sample were excluded for not meeting the quality control metrics, neither of which are explained.

Of 28,759 samples tested, 507 were reported as high risk for the 4 aneuploidies. A follow-up analysis included only 17,885, excluding 10,854. Of 507 high-risk reports, we have no confirmation or follow-up for 151 high-risk reports.

Of the 17,855 cases included in the follow-up analysis, 356 high-risk reports were issued. Fifty-seven (16%) were lost to follow-up, and 22 (6.2%) had terminations with no karyotyping confirmation. This amounts to a massive failure of genetic counseling after the NIPT results. This fact was reported in the Boston Globe by Beth Daley on Dec. 14, 2014. Thirty-six (10.1%) had spontaneous miscarriages with no karyotype confirmation. Nineteen patients with a high-risk report had no karyotyping done when the NIPT was prompted by a soft sonographic marker and a birth defect. This adds up to 134 patients (37.6%) of 356 high-risk reports who essentially had no follow-up confirmation with karyotyping.

From the data presented, all we can be sure of is that of 507 total and 356 high-risk reports with follow-up, only 184 (36% and 52%, respectively) true positive results were confirmed. Thirty-eight false-positive results (10.7%) were reported. The true value of the test cannot be evaluated when 134 of 356 (37.6%) were never followed up. How can one arrive at the true-positive and false-positive rates with so much data missing?

The authors report only 2 false-negative results, which were voluntarily reported. I do not believe voluntary reporting is an accepted means of evaluating the false-negative rate of a test. The presumption by the company that all true-negative results would have been reported to the company is mistaken. Given the incomplete data collection, how did the authors arrive at the conclusion that the positive predictive value was 87% in the population younger than 35 years and 83% in the population older than 35 years when these populations were never enumerated and categorically specified?