The potential of multivariate discriminant analysis in the antenatal detection of neural tube defects.

A method is described which improves the ability to distinguish between normal and neural tube-defective fetuses in utero. Early second trimester amniotic fluid is classified by stepwise discriminant analysis based upon biochemical analysis for alpha-fetoprotein and eight other easily assayed variables; classification is more reliable than using alpha-fetoprotein and gestational age alone.