(1062) Transcerebellar Diameter as a Predictor of Gestational Age: A Retrospective Cohort Study

Objective:
We evaluated the predictive accuracy of transcerebellar diameter (TCD)-based gestational age (GA) estimation against the Hadlock equation for well-dated pregnancies and created nomograms to estimate GA based on TCD.  

Study Design:
This retrospective cohort study utilized data from a multi-site single academic institution between March 2022 and May 2024. We identified patients scanned between 14w0d-38w6d. Only well-dated scans were included (crown-rump length < 12 weeks, last menstrual period consistent with ultrasound < 20 weeks, or in vitro fertilization dating). Multifetal gestations and fetal anomalies were excluded. Maternal diabetes and hypertension were analyzed as sub-groups. For patients with multiple scans during the study period, only the final sonogram was included.

Descriptive statistical analysis examined the trend between TCD measurements and actual GA. A regression analysis was performed to create a nomogram that predicts GA based on TCD alone. Estimated GA based on TCD was compared to actual GA with Pearson’s correlation analysis. GA estimated using Hadlock’s equation was compared to actual GA using Pearson’s correlation. Pearson’s analysis was repeated for sub-groups of maternal hypertension and diabetes.

Results:
220 scans were analyzed after using our selection criteria. Concordance between actual GA and estimated GA using TCD was high (r = 0.98, p < 0.001, mean error = 0 weeks, mean absolute error = 0.62 weeks). Concordance between actual GA and estimated GA using Hadlock’s equation was also high (r=0.98, p < 0.001, mean error = 0.34 weeks, mean absolute error = 0.62 weeks). Outcomes were similar for sub-groups of maternal hypertension (n = 168, r = 0.97, p <0.001) and maternal diabetes (n = 42, r = 0.92, p < 0.001).

Conclusion:
Transcerebellar diameter is a highly accurate and reliable predictor of gestational age, as demonstrated by strong correlation with actual gestational age, low mean absolute error of 0.62 weeks, and lack of systemic bias (mean error of 0). TCD alone is comparable with Hadlock’s equation, which utilizes several measurements when predicting gestational age.