105 Congenital Nasal Masses: CT and MR Imaging Features in 16 Cases 1 2 A. James Barkovich • The imaging studies of 16 children with pathologically proved nasal encephaloceles Pierre Vandermarck 1 (eight), nasal dermal sinuses/nasal dermoids (seven), and nasal cerebral heterotopias, MichaelS. B. Edwards2 more commonly known as nasal gliomas (one), were retrospectively reviewed and Philip H. Cogen2 compared with normal control subjects to define the normal anatomy and analyze deformities caused by these lesions. Nasal encephaloceles were always identified as complex masses of mixed soft tissue and CSF intensity that were contiguous with intracranial structures. The nasal glioma appeared as a mixed-intensity mass that, on the basis of the CT scan, appeared to be continuous with intracranial structures. Nasal dermal sinuses could only be identified as they coursed through the skin and subcuta­ neous soft tissue. They could not be identified when intraosseous. Moreover, on CT and, particularly, on MR, a number of potential diagnostic pitfalls were encountered. The most important of these was the normal fat deposition that occurs within bone during normal maturation and during aeration of the frontal sinuses and nasal bones. These fatty changes can easily be mistaken for fatty tumors if they are not recognized as normal anatomic changes. Interestingly, the classic plain film findings for congenital nasal masses were present only in the encephaloceles and nasal glioma; dermoids and dermal sinuses showed none of the classic plain film findings. In the six patients who had both CT and MR, the masses were easily identified and characterized by each imaging method. Congenital nasal masses are well characterized by both CT and MR . It is important to understand the normal changes in the anatomy of the nasofrontal region in the pediatric age group to avoid false-positive diagnoses in this region. AJNR 12:105-116, January/February 1991 ; AJR 156: March 1991 The congenital midline nasal masses are rare anomalies; their rate of occurrence is approximately one in every 20,000-40,000 live births [1 , 2] . The most common of these masses are dermoid/epidermoid tumors (almost always associated with dermal sinus tracts). nasal cerebral heterotopias (nasal gliomas), and nasal enceph­ aloceles [1-6]. Less common masses include simple inclusion cysts, hemangiomas, and aberrant ethmoidal sinuses [2]. When pediatric patients present with a midline nasal mass or a nasal dimple, an imaging study is usually obtained to better characterize and delineate the disease Received February 22, 1990; revision requested and to plan for appropriate therapeutic intervention. We have found , however, that June 14, 1990; revision received July 15, 1990; accepted July 24, 1990. the complex anatomy and often subtle findings make interpretation of these images ' Department of Radiology, Neuroradiology Sec­ difficult. In this report, we define the normal embryology and anatomy of the tion, University of California, San Francisco, 505 frontonasal region in infants and children , and then present the CT and MR features Parnassus Ave. , San Francisco, CA 94143-0628. of 16 patients who presented with congenital midline nasal masses. Address reprint requests to A. J. Barkovich. 2 Department of Neurological Surgery, Division of Pediatric Neurosurgery, University of California, Materials and Methods San Francisco, CA 94143. 0195-6108/91 /1201-0105 The imaging studies of 16 children with pathologically proved nasal encephaloceles (eight), © American Society of Neuroradiology nasal dermoidsfdermal sinuses (seven), and nasal gliomas (one) were retrospectively reviewed TABLE 1: Patient Data 0 O'l Crista Mass Patient Imaging FC Size IOD Associated Age Sex Diagnosis Presentation Size No. Method (mm) Bifid Location Size (mm) MR CT (mm) Anomalies (mm) 3 mo F CT DS dermoid Nasal mass 5 4 (+) Glabella & in- 6 x 4 - Low atten. 15 tracranial 4 x 4 2 12 mo M MR Dermoid Nasal mass 0 (closed) 7 (-) Left lateral 10 X 10 Short, T1 , T2 - 18 nasal bridge 3 2 yr F CT and DS epider- Nasal mass 3 4 (-) Glabella 3 X 4 Long T1, T2 Low atten. 22 MR maid 4 2 yr F CT and DS Nasal dimple 2 4 (-) - - - 18 MR 5 3 yr M CT and DS dermoid Nasal dimple 6 4 (-) Glabella & 3 x 5 Short T1 , T2 Low atten. 16 MR and mass nasal bridge 6 3 yr M MR DS Nasal dimple 0 (closed) 3 (-) - - - 13 7 12 yr F CT and DS Nasal dimple 7 5 (-) - - - - 23 Colloid cyst MR 8 9 mo F CT Nasal Nasal mass Poorly de- 7 (-) Left glabella 10 X 13 - Soft-tissue 12 glioma fined Atten. 9 1 wk F CT and NE Nasal mass 15 Not (-) Glabella & 30 X 50 Heterogeneous - 16 L temporal (lJ )> MR seen ethmoids soft tissue & arach- JJ CSF noid cyst :::'\ 0 10 6 wk F MR NF Nasal mass 5 2 (-) Low frontal 20 X 20 Heterogeneous - 12 Subfrontal < soft tissue & cyst 0 CSF I 11 3 mo F CT and NF and NE Nasal mass 14 3 (-) Glabella & 20 X 35 Heterogeneous - 20 Bilateral m MR ethmoids soft tissue temporal --l arach- )> noid ! cysts 12 4 mo F MR NE Meningitis, 14 Not - Glabella & 15 X 30 Heterogeneous - 28 ACC nasal seen ethmoids soft tissue & mass CSF 13 6 mo M CT NE Nasal mass 20 3 (-) Ethmoids & 25 X 20 - Hetero- 20 ACC glabella ge- neous soft tis- sue 14 6 mo M CT NO Proptosis, 8 4 (-) Bilateral or- 10 X 15 - Hetero- . R. parietal orbital bits mass ge- schizen- masses each or- neous cephaly ~ '-z bit mostly :0 fluid 15 10 mo F MR NF and NE Nasal mass 15 3 (+) Glabellar & 8 X 15 Mixed short & - 26 ACC, li- '-"' low frontal intermediate poma ::J '"c T1 , T2 ~'" 16 24 yr M MR NE Nasal mass 20 Not - Glabellar 30 X 40 Heterogeneous - 25 - "T1 seen soft tissue & g-(1) CSF c ~'" Note.-FC = foramen cecum, OS = dermal sinus, NE = nasoethmoidal encephalocele, NF = nasofrontal encephalocele, NO = nasoorbital encephalocele, 100 = interorbital distance, ACC = agenesis of the ~ corpus callosum . ~ • We were unable to measure 100 due to poorly defined medial orbital walls. AJNR :12 , January/February 1991 CT AND MR OF CONGENITAL NASAL MASSES 107 (Table 1). The eight patients (three males, five females) with enceph­ cecum , and glabella. The size of the foramen cecum was measured aloceles all presented with nasal (si x patients), low frontal (one), or on axial scans as the di stance from the most anterior aspect of the orbital (one) masses. They ranged in age from 1 week to 24 years base of the crista galli to the most posterior aspect of the frontal old at the time of their imaging study (mean , 3.5 years; median, 4 bones. The greatest width and any abnormalities in the shape of the months). Four patients had MR scans only, two had CT scans only, crista galli were recorded . The interorbital distance (defined as the and two had both MR and CT scans. The three patients (one male , narrowest distance between the medi al walls of the orbits at the level two females) with nasal dermal sinuses presented with nasal dimples; of the cribiform plate) was measured. Any abnormal soft ti ssue or the four patients (two males, two females) with associated dermoids bony masses were noted. presented with nasal masses. These seven patients ranged in age Measurements of both the foramen cecum and the cri sta galli have from 3 months to 12 years at the time of their imaging study (mean , a large potential for error because the value obtained may , theoreti­ 2.6 years; median , 3 years). Four patients had CT and MR , two had cally, vary depending on the orientation of the imaging plane to the only MR, and one had only CT. The patient with the nasal glioma skull base. Fortunately, axial CT images at our in stitution are obtained presented with a nasal mass and was 9 months old at the time of routinely in a plane parallel to th e canthomeatal line, resulting in a the study; she was studied only by CT. (usually) con sistent orientation. Further possibility for error exists in Eleven of the 12 MR scans were obtained at 1 .5 T using a 256 x the fact that the size of the foramen cecum and the width of the 256 matrix. Axial 3-mm (1-mm gap) spin-echo (S E) 600/20/2 (TR /TE/ crista vary at different axial level s. Fortunately, the phenomenon of excitations) MR images were obtained in all these patients. Axial SE volume averaging results in vi sualization of the narrowest diameter 2500f20,70f2 images with a 3-mm slice thickness were obtained in of a foramen and the largest diameter of an y bone structure within a two patients and a 5-mm slice thickness in nine patients. Sagittal SE slice; therefore, we feel that the measurements we have recorded 600/20 images with a 3- or 4-mm slice thickness were obtained in are accurate and comparable. Indeed, upon scanning a skul l phantom 10 patients (one did not have a sagittal imaging sequence) and coronal five times at slightly different scan angles, results varied by less than SE 600/20 images with a 4- or 5-mm sl ice thickness were obtained 20%. Moreover, all varied by less than 25% from the measurement in 11 patients. A single patient (case 11 ) was scanned at 0.35 T. obtained directly from the skull. Axial SE 2000/40,80 and coronal SE 500/30 images were obtained by using a 7 -mm slice thickness with a 3-mm gap and a 256 x 128 matrix. CT scans were obtained in 1 0 patients.