The medical records of 55 fetuses with sonographically diagnosed mild ventriculomegaly (MVM) were reviewed to assess prognosis. Fetuses were divided into two groups based on the presence or absence of sonographically detected associated fetal anomalies: 13 had no other anomalies detected (isolated MVM), and 42 had concomitant neural axis and visceral anomalies (nonisolated MVM). Mortality was 83% among fetuses with nonisolated MVM and 38% among fetuses with isolated MVM (P less than .005). If terminated pregnancies are excluded, only one of nine (11%) fetuses with isolated MVM died, compared with nine of 16 (56%) fetuses with nonisolated MVM (P less than .005). There are 15 living children: Nine (60%) are developmentally normal at 6-30 months of follow-up (six had isolated MVM), three (20%) are or are likely to be abnormal, and we were unable to follow up three (20%). Fetal anomalies were missed in 11 of 30 (37%) fetuses with detailed follow-up. However, this would have changed the classification from isolated to nonisolated MVM in only one case. Thus, in 54 of 55 cases (or 29 of 30 cases with detailed follow-up), fetuses were accurately classified as having isolated or nonisolated MVM. The authors conclude that sonographically isolated MVM is associated with a significantly better prognosis than nonisolated MVM, and fetuses can be classified accurately based on prenatal sonograms.