Bianchi et al 1 have confirmed recent prospective studies [2][3][4] showing that massively parallel sequencing of cell-free DNA in maternal plasma is predictive of common nonmosaic autosomal trisomies in high-risk pregnancies. However, their estimates of sensitivity and specificity could lead to false conclusions.Sensitivity and specificity of a test are estimated from the overlapping distributions of affected and unaffected individuals using a specified cutoff. For massively parallel sequencing, sensitivity is the proportion of trisomy pregnancies with z-score above the cutoff, and specificity is the proportion of euploid pregnancies with lower scores.In Bianchi et al, 1 using a zϭ2.5 cutoff, the sensitivity and estimated specificity are: Down syndrome 100% (90/90) and 98.5% (404/410); Edwards syndrome 97.4% (37/38) and 99.4% (460/463); Patau syndrome 81.2% (13/ 16) and 100% (485/485). Using zϭ4.0, the rates are: Down syndrome 98.9% (89/90) and 410/410 (100%); Edwards syndrome 92.1% (35/38) and 100% (463/463); Patau syndrome 68.8% (11/ 16) and 100% (485/485).However, instead of using either set of rates, the authors overstated performance, because tests with z-scoresϭ2.5-4.0 were designated "unclassified" and excluded. No reason was given for this exclusion, which, unfortunately may give the impression that a positive result (zϾ4.0) is fully diagnostic. In fact, a study of this size cannot exclude the possibility of false-positive results being relatively frequent, particularly when many types of aneuploidy are evaluated.A result in the 2.5-4.0 range is associated with about a fivefold increased risk of aneuploidy: 3.5% (5/144) of trisomies had results in this range compared with 0.7% (9/1358) without the trisomy being evaluated. Women with results in such a range would need to receive information about their individual risk given their massively parallel sequencing result, as is done for conventional aneuploidy screening.At this early stage of development, confirmatory invasive testing is needed for women with a positive massively parallel sequencing test. 5 Risks are also very high for those in the 2.5-4.0 range, and they, too, might be offered invasive testing. When sufficient data become available, it will be possible to estimate individual post-massively parallel sequencing risks for all women and a more optimal cutoff for offering invasive testing then could be defined.