Many questions remain about the feasibility of using fetal cells from maternal blood for prenatal diagnosis. Although recently there has been more focus on clinically relevant methods, many studies have been performed using blood drawn after invasive procedures, and over a wide range of gestational ages. For methods to be applicable to clinical use, more work is needed on isolating cells early in pregnancy, when termination is still an option for parents who are found to have an affected pregnancy. It is generally agreed that fetal nucleated erythrocytes are the most efficacious cell type for prenatal diagnosis, but it has not yet been shown definitively whether there is an ideal gestational age for sampling, whether ABO incompatibility might limit availability of fetal cells, or whether the number of cells present might be different in normal versus abnormal pregnancies. PCR has been shown to be a powerful tool in allowing amplification and identification of very small amounts of fetal DNA. However, this is limited to cases in which a specific and unique gene from the father is sought. This means that there is the potential to diagnose many paternally inherited autosomal dominant diseases and some autosomal recessive diseases, in which the parents have different and identifiable mutations. However, when parents are both carriers of the same autosomal recessive mutations, or when the disease is X linked, PCR will not aid in prenatal diagnosis. Cytogenetic analysis of fetal cells by FISH after cell sorting is another potentially useful method of prenatal diagnosis, but requires relatively pure samples of fetal cells or an independent marker that allows easy microscopic identification. The latter might be accomplished by identifying fetal cells through their expression of embryonic hemoglobins or because they contain HLA-G mRNA. In addition, current techniques of cell sorting must be improved so that a higher percentage of fetal cells can be isolated. Currently, the best cell sorting techniques usually produce a maximum purity of 10% fetal cells. Commonly, in normal pregnancies, fewer than 0.1% of the cells isolated after sorting are fetal in origin. Improving the concentration and quantity of fetal cells will improve the accuracy of FISH. Methods such as immunophenotyping that allow the selective identification of fetal cells by microscopy, and can be used in conjunction with FISH, may be extremely valuable because they may allow the genetic analysis of only the few fetal cells within a background preponderance of maternal cells. Although the retrieval of fetal cells from maternal blood is an attractive concept, it must be clearly stated that presently it is only in the investigational phase because of the low sensitivity and specificity. There is no current application for these methods in clinical practice. It remains to be determined whether testing maternal blood for fetal cells or DNA will be used as a screening tool, similar to the maternal serum screening currently in use, or whether the accuracy can be improved to a level such that the techniques can be used diagnostically. Although there are many questions that remain unanswered at this time, the outlook for noninvasive prenatal genetic testing in the future is optimistic.