OBJECTIVES/HYPOTHESIS: To determine if ultrasound could differentiate between thyroglossal duct cysts (TGDC) and midline dermoid cysts (DC). STUDY DESIGN: Cohort study. METHODS: A search of pathology reports yielded 91 patients with TGDC or midline DC. Ultrasound images were presented to a radiologist blinded to pathology who evaluated the following: 1) depth of lesion from skin, 2) maximum diameter, 3) dimensions, 4) midline location, 5) distance from base of tongue, 6) tract, 7) wall regularity, 8) wall thickness, 9) margin definition, 10) heterogeneity, 11) internal septae, 12) solid components, 13) intralesional Doppler flow, and 14) posterior enhancement. The predictive power of these variables was evaluated in a multiple logistic regression model. RESULTS: There were 53 TGDC and 38 DC. TGDC were significantly more likely than DC to have the following features: 1) smaller distance from base of tongue, 2) tract, 3) irregular wall, 4) ill-defined margin, 5) internal septae, 6) solid components, and 7) intralesional Doppler flow. Three clinically reliable ultrasound variables were independently able to discriminate between TGDC and DC. A predictive model was fashioned whereby each variable was scored as 0 or 1, with a total score calculated (septae + irregular wall + solid components = TGDC [or SIST] score). We propose a scoring system whereby 0 = suggestive of DC; 1 = suggestive of TGDC; and ≥2 = highly suggestive of TGDC. CONCLUSIONS: It may be possible to differentiate between TGDC and midline DC preoperatively using ultrasound.
OBJECTIVES/HYPOTHESIS: To determine if ultrasound could differentiate between thyroglossal duct cysts (TGDC) and midline dermoid cysts (DC). STUDY DESIGN: Cohort study. METHODS: A search of pathology reports yielded 91 patients with TGDC or midline DC. Ultrasound images were presented to a radiologist blinded to pathology who evaluated the following: 1) depth of lesion from skin, 2) maximum diameter, 3) dimensions, 4) midline location, 5) distance from base of tongue, 6) tract, 7) wall regularity, 8) wall thickness, 9) margin definition, 10) heterogeneity, 11) internal septae, 12) solid components, 13) intralesional Doppler flow, and 14) posterior enhancement. The predictive power of these variables was evaluated in a multiple logistic regression model. RESULTS: There were 53 TGDC and 38 DC. TGDC were significantly more likely than DC to have the following features: 1) smaller distance from base of tongue, 2) tract, 3) irregular wall, 4) ill-defined margin, 5) internal septae, 6) solid components, and 7) intralesional Doppler flow. Three clinically reliable ultrasound variables were independently able to discriminate between TGDC and DC. A predictive model was fashioned whereby each variable was scored as 0 or 1, with a total score calculated (septae + irregular wall + solid components = TGDC [or SIST] score). We propose a scoring system whereby 0 = suggestive of DC; 1 = suggestive of TGDC; and ≥2 = highly suggestive of TGDC. CONCLUSIONS: It may be possible to differentiate between TGDC and midline DC preoperatively using ultrasound.