Fetal thoracic teratomas: mediastinal or pericardial?

Background: Mediastinal and pericardial teratomas have overlapping imaging features that may make accurate prenatal diagnosis challenging.

Objective: To identify prenatal imaging features that may aid in distinguishing between mediastinal and pericardial teratomas.

Materials And Methods: Prenatally diagnosed pericardial and mediastinal teratomas evaluated at our fetal center from 1995 to 2020 were included in this Institutional Review Board-approved study. Lesion volume was calculated using prospectively reported ultrasound (US) measurements and the formula of a prolate ellipsoid, which was then normalized to head circumference. Prenatal US and magnetic resonance imaging (MRI) studies were anonymized with two fetal imagers reviewing the US studies and two different fetal imagers reviewing the MRI studies. These experienced reviewers scored location of the mass in the craniocaudal axis and in the transverse axis. MRI reviewers also scored the presence of inferior cardiac compression by the lesion and whether there was identifiable thymic tissue. Reviewer disagreements were resolved by consensus review.

Results: Eleven pericardial teratomas and 10 mediastinal teratomas were identified. All cases underwent detailed fetal anatomic US and fetal echocardiogram and 10/11 (91%) pericardial teratomas and 8/10 (80%) mediastinal teratomas underwent fetal MRI. Median volume was higher for mediastinal teratomas compared to pericardial teratomas (42.5 mL [interquartile range (IQR) 15.9 - 67.2 mL] vs. 8.1 mL [IQR 7.7 - 27.7 mL], P=0.01) and median volume/head circumference was also statistically higher in mediastinal teratomas (1.33 [IQR 0.78 - 2.61] vs. 0.43 [IQR 0.38 - 1.10], P=0.01). Logistic regression analysis demonstrated a statistical difference between teratoma types with respect to location in the craniocaudal axis by both modalities with mediastinal teratomas more commonly located in the upper and upper-middle thorax compared to pericardial teratomas, which were more commonly found in the middle thorax (US, P=0.03; MRI, P=0.04). Logistic regression analysis also demonstrated a statistical difference between teratoma types with respect to position along the transverse axis by both modalities with mediastinal teratomas more commonly located midline or left paramedian and pericardial teratomas more often right paramedian in location (US, P<0.01; MRI, P=0.02). Inferior cardiac compression observed by MRI was associated more commonly with mediastinal teratomas compared to pericardial teratomas (87.5% [7/8] vs. 10% [1/10], P<0.01). Identifiable thymus by MRI was more commonly observed in cases of pericardial teratomas, however, this difference was not statistically significant (P=0.32).

Conclusion: Mediastinal teratomas are associated with larger lesion size and inferior cardiac compression when compared to pericardial teratomas. These features combined with lesion location in the craniocaudal and transverse axes may allow for more accurate prenatal diagnosis and optimal perinatal and surgical management.