H. Page McAdams, MDWanda M. Kirejczyk, MD2

Melissa L. Rosado-de-Christenson, Col, USAF, MC

Shigeru Matsumoto, MD

Index terms:Bronchi, cyst, 60.1441, 67.1441,

67.3156,Lung, cysts, 60.1441Mediastinum, cysts, 67.1441,

67.3156Thorax, CT, 60.12111, 60.12112Thorax, MR, 60.121411

Radiology 2000; 217:441–446

Abbreviation:CSF 5 cerebrospinal fluid

1 From the Department of Radiology,Box 3808, Duke University MedicalCenter, Durham, NC 27710 (H.P.M.);the Department of Radiologic Pa-thology, Armed Forces Institute ofPathology, Washington, DC (W.M.K.,M.L.R.d.C.); the Department of Radi-ology and Nuclear Medicine, Uni-formed Services University of the HealthSciences, Bethesda, MD (M.L.R.d.C.);and the Department of Radiology,Jikei University of Medicine, Tokyo, Ja-pan (S.M.). From the 1995 RSNA sci-entific assembly. Received December20, 1999; revision requested January25, 2000; revision received February24; accepted February 22. Addresscorrespondence to H.P.M.

The opinions and assertions con-tained herein are solely those of theauthors and are not to be construed asofficial or as representing those of theDepartments of the Air Force or De-fense.

Current address:2 Department of Radiology, New Brit-ain General Hospital, Conn.© RSNA, 2000

Author contributions:Guarantors of integrity of entire study,H.P.M., M.L.R.d.C.; study conceptsand design, H.P.M., M.L.R.d.C.; defi-nition of intellectual content, H.P.M.,M.L.R.d.C.; literature research, H.P.M.,W.M.K.; clinical studies, all authors;data acquisition and analysis, all au-thors; manuscript preparation, H.P.M.,W.M.K., M.L.R.d.C.; manuscript editingand review, H.P.M., M.L.R.d.C.

Bronchogenic Cyst: ImagingFeatures with Clinical andHistopathologic Correlation1

PURPOSE: To characterize the imaging features of bronchogenic cysts.

MATERIALS AND METHODS: The computed tomographic (CT) and/or magneticresonance (MR) or ultrasonographic images in 68 histopathologically proved casesof bronchogenic cyst in 38 male and 30 female patients, aged newborn to 72 years(mean, 22 years), were retrospectively reviewed.

RESULTS: There were 58 mediastinal and 10 extramediastinal cysts. At CT (n 5 62),60 cysts were sharply marginated with smooth (n 5 35) or lobulated (n 5 25)borders. Twenty-five cysts were of water attenuation, 25 were of soft-tissue atten-uation, two were air filled, two had an air-fluid level, and two had dependent milkof calcium. On T1-weighted MR images (n 5 23), 18 cysts were hyperintense andfive were isointense to cerebrospinal fluid. On T2-weighted MR images (n 5 18), 17cysts were isointense or hyperintense to cerebrospinal fluid. Of the 25 soft-tissue–attenuation lesions at CT, 11 appeared cystic because of internal homogeneity, lackof internal enhancement, mural enhancement, and characteristic location. Fourteenappeared solid based on morphology and attenuation. MR imaging of nine of thelatter showed marked hyperintensity on T2-weighted images.

CONCLUSION: CT of bronchogenic cysts typically shows sharply marginated me-diastinal masses of soft-tissue or water attenuation. Most appear cystic. A minorityappear solid and can be confused with other lesions; MR imaging can be useful forelucidating the cystic nature of these lesions.

Bronchogenic cysts are congenital lesions thought to result from abnormal budding of theembryonic foregut. At computed tomography (CT), bronchogenic cysts typically manifestas spherical masses of either water or soft-tissue attenuation. When bronchogenic cystsmanifest as water-attenuation masses on CT scans, differentiation from other mediastinalprocesses such as lymphadenopathy or neoplasia is not difficult. However, when bron-chogenic cysts manifest as soft-tissue–attenuation masses on CT scans, differentiationfrom solid lesions can be more problematic. We reviewed a large series of bronchogeniccysts to fully define the spectrum of CT, magnetic resonance (MR) imaging, and ultra-sonographic (US) features of bronchogenic cyst.

MATERIALS AND METHODS

All cases with a histopathologic diagnosis of bronchogenic cyst and cross-sectional imag-ing (CT, MR imaging, or US) that were referred to the Department of Radiologic Pathology,the Armed Forces Institute of Pathology between 1978 to 1997 were included. By defini-tion, all cysts fulfilled strict histopathologic criteria for diagnosing bronchogenic cyst: alining of respiratory epithelium associated with a wall containing glands, cartilage, andsmooth-muscle elements (Fig 1) (1). We studied 68 cysts in 68 patients (38 male and 30female patients; age range, newborn to 72 years; mean age, 22 years).

Clinical records were available in 66 cases and were reviewed, with notation made ofpresenting complaints. Surgical reports were reviewed for cyst location and presence of astalk and point or points of attachment to adjacent mediastinal structures such as the

Thoracic Imaging

441

esophagus or tracheobronchial tree. Pa-thology reports were reviewed for grossdescription of the maximal diameter ofthe lesion and the cyst contents.

Forty-two cysts were imaged with onlyCT, 20 cysts were imaged with both CTand MR imaging, four cysts were imagedwith only MR imaging, and two cystswere imaged with only prenatal US. Be-cause of the retrospective and multiinsti-tutional nature of the study, CT and MRtechniques were inconsistent. Of the 62CT scans obtained, 33 scans were onlycontrast material enhanced, 14 were onlynonenhanced, and 15 were obtained withand without contrast material enhance-ment. Of the 24 MR examinations, 23studies used a short repetition time/shortecho time sequence (T1-weighted), and18 used a long repetition time/long echotime sequence (T2-weighted). Gadolini-um-based contrast material was adminis-tered in three cases. All imaging studieswere reviewed by at least two thoracicradiologists (H.P.M., M.L.R., W.M.K., S.M.)simultaneously, and findings were re-corded per consensus.

CT scans were evaluated for cyst shapeand margin characteristics, visualizationof a cyst wall, presence of mass effect onadjacent structures, cyst location, atten-uation, homogeneity, calcification, andpatterns of enhancement following intra-venous administration of iodinated con-trast material. Lesions were classified asanterior mediastinal if they were locatedanterior to the heart or great vessels (pre-vasular space), posterior mediastinal if lo-cated in either paraspinal region, andmiddle mediastinal if located in the para-tracheal or subcarinal regions or alongthe course of the esophagus. On CTscans, the mediastinal cysts were classi-fied as either water or soft-tissue attenu-ation in one of two ways. If region-of-interest measurements within the cystwere available, the lesions were classifiedas water attenuation if their attenuationwas less than 20 HU and as soft-tissue

attenuation if their attenuation was greaterthan 20 HU. If region-of-interest measure-ments were not available, the cysts wereclassified as water attenuation if their at-tenuation was similar to fluid in the gall-bladder or spinal canal and as soft-tissueattenuation if their attenuation wasgreater than fluid in the gallbladder orspinal canal.

The soft-tissue–attenuation lesions onCT scans were subcategorized as cysticappearing, solid appearing, or indetermi-nate according to the following criteria:(a) If the attenuation of the lesion wasless than that of surrounding soft tissue,if the lesion was internally homoge-neous, if there was no internal enhance-ment, and if there was a well-defined thinwall, the lesion was categorized as cysticappearing. (b) If the attenuation was sim-ilar to that of the surrounding soft tis-sues, if the lesion was internally hetero-geneous, and if it did not have a well-defined thin wall, it was categorized assolid appearing. (c) Lesions were catego-rized as indeterminate if they did notmeet all stated criteria for classification aseither cystic appearing or solid appear-ing.

MR images were evaluated for signalintensity on T1- and T2-weighted images,homogeneity, and enhancement follow-ing intravenous administration of gado-linium-based contrast material. Signalintensity on T1-weighted images wasclassified as isointense to cerebrospinalfluid (CSF), isointense to skeletal muscle,or hyperintense to skeletal muscle. Signalintensity on T2-weighted images was clas-sified as hypointense to CSF, isointense toCSF, or hyperintense to CSF.

US images were specifically evaluated

for cyst location, visualization of a cystwall, internal echos or septations, andpresence or absence of acoustic enhance-ment.

RESULTS

Clinical Presentation

Fifty-seven (84%) of 68 affected pa-tients presented in the first 4 decades oflife (Table 1). Of the 66 patients forwhom a clinical history was available, 38patients (58%) were symptomatic withpain (n 5 15), dyspnea (n 5 8), respira-tory infection (n 5 7), wheezing (n 5 5),cough (n 5 5), dysphagia (n 5 1), pneu-mothorax (n 5 1), and superior vena cavasyndrome (n 5 1). Several patients hadmore than one symptom. Twenty-eightpatients (42%) were asymptomatic, andthe bronchogenic cyst was incidentallydiscovered because of an abnormality de-

TABLE 1Distribution of Bronchogenic Cystsby Patients’ Age

Age (y) No. of Cases (n 5 68)

0–9 2310–19 620–29 1630–39 1240–49 550–59 260–69 370–79 1 Figure 2. Bronchogenic cyst in a 19-year-old

asymptomatic man. Intraoperative photo-graph shows a bronchogenic cyst (C) with at-tachment to the distal esophagus (e) by a stalk(arrowheads).

Figure 3. Bronchogenic cyst in a 1-year-oldasymptomatic boy. Photograph of a specimenshows a pearly white translucent cyst. Noteblood vessels in the cyst wall.

Figure 1. Bronchogenic cyst in a 3-year-oldasymptomatic boy. High-power photomicro-graph shows characteristic features of a bron-chogenic cyst. Note cyst lining of pseudo-stratified columnar respiratory epithelium (ar-rowhead), cartilage plate (C), smooth muscle(S), and bronchial glands (arrow) in the cystwall. (Hematoxylin-eosin stain; original mag-nification, 3150.)

442 z Radiology z November 2000 McAdams et al

tected on a radiograph obtained for otherreasons.

Surgical Findings

Sixty-seven patients underwent com-plete surgical resection of their broncho-genic cyst. In one patient, biopsy andmarsupialization were performed of thecyst wall, but it was not completely re-sected. At surgery, a stalk connecting thecyst to an adjacent structure was identi-fied in 31 cases (Fig 2). The attachmentpoint of the stalk varied widely. In eightcases, the stalk was attached to the tra-chea or carina. A pleural attachment wasnoted in four cases and an esophagealattachment in four. Other points of at-

tachment included the diaphragm (n 53), hilum (n 5 1), aorticopulmonary win-dow (n 5 1), and sympathetic trunk (n 51). In the remaining nine cases in whicha stalk was surgically identified, the pointof attachment was not documented.

Gross Features

Descriptions of gross specimens wereavailable in all 68 cases, and measure-ments of the cyst size were available in67. Maximal cyst diameter ranged from1.3 to 11.0 cm, with a mean of 4.8 cm(Fig 3). The cyst content varied but was ofa liquid nature in all cases. Descriptionsof the cyst contents were available in 53cases. The fluid was most frequently de-scribed as milky and gelatinous (n 5 16).Other descriptions of the fluid includedgreen and mucoid (n 5 12), brown orhemorrhagic (n 5 8), yellow and gelati-nous (n 5 6), white and translucent (n 55), yellow and puslike (n 5 3), and thin

and serous (n 5 1). Two cysts were com-pletely air filled. The chemical composi-tion of the fluid was not analyzed in anycase.

Location

There were 58 mediastinal (85%) and10 extramediastinal (15%) bronchogeniccysts. Of the mediastinal cysts, 46 cysts(79%) were located in the middle medi-astinum, 10 (17%) were located in theposterior mediastinum, and two (3%)were located in the anterior mediasti-num. The extramediastinal cysts were lo-cated in the lung parenchyma in sevencases (Fig 4), the diaphragm in two, andthe pleura in one. The lung parenchymalcysts did not show a predilection for aparticular lobe and were located in theright upper (n 5 2), right lower (n 5 3),and left upper (n 5 2) lobes.

CT Findings

On CT scans (n 5 62), 60 cysts weresharply marginated masses with eithersmooth (n 5 35) or lobulated (n 5 25)borders. Indeterminate borders were seenin two cases, likely due to obscuration ofthe margins by associated atelectasis. Fif-ty-eight lesions were in the mediastinum,and four were in the lung parenchyma.The CT attenuation characteristics of thelesions are summarized in Table 2. Of the58 mediastinal cysts, 25 were classified aswater attenuation (Fig 5) and 25 wereclassified as soft-tissue attenuation (Fig6). Forty-eight of these 50 cysts were ofhomogeneous attenuation and two wereof heterogeneous attenuation. Of the re-maining eight mediastinal lesions im-aged with CT, six could not be classifiedbecause streak artifact obscured much of

TABLE 2CT Attenuation Characteristics ofBronchogenic Cysts

Attenuation Characteristicsby Location

No. of CT Scans(n 5 62)

Mediastinum (n 5 58)Water attenuation 25Soft-tissue attenuation 25

Cystic appearing 11Solid appearing 7Indeterminate 7

Milk-of-calcium 2Streak artifact–obscured

cyst 6Lung (n 5 4)

Completely air filled 2Air-fluid level 2

Figure 4. Prenatal diagnosis of bronchogenic cyst. (a)Transverse sonogram of the thorax obtained at 26 weeks gestation shows a unilocular cyst(C) in the right hemithorax. Note posterior acoustic enhancement (arrows). (b) Supine frontal and (c) cross-table lateral chest radiographs obtained3 days after birth show a thin-walled (arrows in b), air-filled cyst in the right lung, with a dependent air-fluid level (arrowheads in c).

Figure 5. Bronchogenic cyst in a 21-year-oldasymptomatic woman. Transverse contrast-en-hanced chest CT scan (mediastinal windowsettings) shows a well-circumscribed water-at-tenuation cyst in the middle mediastinum.Note the thin, enhancing wall medially (blackarrows) and peripheral punctate calcification(white arrow).

Volume 217 z Number 2 Bronchogenic Cyst: Imaging Features with Clinical and Histopathologic Correlation z 443

the cyst and two had dependent high-attenuating material consistent withmilk of calcium (Fig 7). On the basis ofpreviously stated criteria, the 25 soft-tis-sue–attenuation lesions at CT were sub-categorized as cystic appearing (n 5 11),solid appearing (n 5 7), or indeterminate(n 5 7). Of the four lung parenchymalcysts imaged with CT, two were com-pletely air filled and two had an air-fluidlevel (Fig 4).

Calcification was seen in six (10%) of62 cysts imaged with CT. In four cases,calcification was located peripherally inthe cyst wall and was punctate and dis-continuous in nature (Fig 5). In twocases, milk of calcium was present in thecyst fluid (Fig 7). Contrast-enhanced CT(n 5 48) did not show enhancement ofcyst contents in any case. Mural en-hancement, however, was observed in 16cysts and was helpful in the delineationof a thin wall (Fig 5).

Mass effect on surrounding structuressuch as bronchi, esophagus, or medias-tinal vessels was observed in 25 cases.Atelectasis and/or parenchymal consol-idation was seen in eight cases. Threepatients had an ipsilateral extralobar se-questration, one patient had a smallpericardial effusion, and one patienthad an ipsilateral congenital diaphrag-matic hernia.

MR Imaging Findings

On T1-weighted MR images (n 5 23),12 cysts were slightly hyperintense toskeletal muscle (Fig 6), six cysts wereisointense to skeletal muscle, and fivecysts were isointense to CSF (Fig 8). OnT2-weighted MR images (n 5 18), 11 cystswere isointense to CSF (Figs 6, 8), sixcysts were hyperintense to CSF, and onewas hypointense to CSF. On gadolinium-enhanced MR images (n 5 3), the cystcontents did not enhance, but enhance-ment of the cyst wall was seen, aidingdelineation of a thin wall (Fig 8).

Twenty cysts were imaged with bothMR imaging and CT. Of these, nine werecategorized at CT as either solid-appear-ing or indeterminate lesions (Fig 6). AtMR imaging, all nine cysts were of in-creased signal intensity compared to CSFon T1-weighted images and of markedlyincreased signal intensity (similar to thatof CSF) on T2-weighted images, the latterfinding being consistent with the cysticnature of these lesions.

Prenatal US

Prenatal US in one case showed aunilocular anechoic cyst in the right lung

with posterior acoustic enhancementand mass effect on the right atrium (Fig4). Prenatal US in a second case showed amultilocular cyst with dependent echo-genic material in the right lung. Bothcysts developed air-fluid levels on chestradiographs in the postnatal period.

DISCUSSION

Bronchogenic cysts are congenital le-sions thought to result from abnormalbudding of the ventral foregut that oc-curs between the 26th and 40th days ofgestation (2). This abnormal bud subse-quently differentiates into a fluid-filled,blind-ending pouch. Most cysts are lo-cated in the mediastinum, near the tra-cheal carina. In our series, 58 (85%) of 68cysts were mediastinal in location, and46 (79%) of 58 mediastinal cysts were inthe middle mediastinum. Less commonly,cysts may occur within the lung paren-chyma, pleura, or diaphragm (10 [15%]of 68 cysts). A stalk or pedicle attachingthe cyst to an adjacent mediastinal struc-

ture is common and was found in 31(46%) of our 68 cases. Bronchogenic cysts

Figure 6. Bronchogenic cyst in a 40-year-oldasymptomatic man. (a) Transverse contrast-enhanced chest CT scan (mediastinal windowsettings) shows a homogeneous soft-tissue at-tenuation lesion (M) without mural enhance-ment in the subcarinal region. Note the masseffect on the right inferior pulmonary veinand left atrium (arrows). This lesion cannot beconfidently differentiated from a solid neo-plasm on the CT scan. (b) Transverse T1-weighted MR image (1.3 T; repetition time msec/echo time msec, 500/25) shows a somewhatheterogeneous lesion (M) of increased signal in-tensity compared to CSF. (c) Transverse T2-weighted MR image (1.3 T, 2,000/80) shows thatthe lesion (M) is of similar signal intensity to thatof CSF, which suggests a cyst.

Figure 7. Bronchogenic cyst in a 55-year-oldman with dysphagia. Transverse contrast-en-hanced CT scan (mediastinal window settings)shows a middle mediastinal mass (m) of waterattenuation. Note the dependent high-attenu-ating material (arrows) consistent with milk ofcalcium and the mass effect on the distal con-trast material–filled esophagus (arrowhead).

444 z Radiology z November 2000 McAdams et al

are sometimes found in association withother congenital pulmonary malforma-tions such as sequestration and lobar em-physema (3). Accordingly, three cysts inour series were associated with ipsilateralextralobar sequestration, and one was as-sociated with a congenital diaphragmatichernia.

The frequency of extramediastinal cystsin our series (15%) was lower than thatoften reported in the literature (4–6).This may be a reflection of the strict cri-teria we used to diagnose bronchogeniccyst (1). We excluded all lesions that didnot have cartilage in their wall, becauseof potential confusion with other cysts,both congenital and acquired (1). Otherseries (5–7), however, included lesionsthat did not have cartilage in their wall.For example, in the series of St-Georges etal (7), only 37 (43%) of 86 lesions classi-fied as bronchogenic cysts had cartilagein their wall, and approximately 25% ofthe 86 cysts were extramediastinal in lo-cation. Because intrapulmonary cystsmay contain fewer bronchial wall compo-nents than do mediastinal cysts, we mayhave excluded some true extramediastinalbronchogenic cysts from our series.

In our series, as in others (7,9), most

affected patients presented in the firstfew decades of life. Presentation beyond50 years of age is distinctly unusual. Mostof our patients for whom clinical historywas available (38 [58%] of 66) were symp-tomatic at presentation. Chest pain wasthe most common presenting complaint.Symptoms, when present, were likely re-lated to mass effect on adjacent struc-tures. A sizable minority of patients (28[42%] of 66), however, were asymptom-atic, with the cyst detected as an inciden-tal radiographic finding.

The fluid within bronchogenic cysts isusually a mixture of water and protein-aceous mucus. The contents of the cystsin our series were not chemically ana-lyzed. However, descriptions of the fluidranged from a thin, watery liquid to hem-orrhagic fluid to a very viscous, mucoidmaterial. This variability in cyst contentis likely responsible for the variability ofattenuation seen on CT scans and signalintensity characteristics seen on MR im-ages (10,11). It has been suggested, basedon analysis of cyst fluid, that calcium is amajor factor contributing to high atten-uation on CT scans (12,13). However,most of the cysts that we classified ashaving soft-tissue attenuation on CTscans (based on region-of-interest mea-surements) were hyperintense to CSF onT2-weighted MR images and isointense

to skeletal muscle on T1-weighted MRimages. This finding suggests that mucusand proteinaceous debris, not calcium,within the cyst is the most likely expla-nation for the increased attenuation seenon CT scans (14). We did, however, ob-serve two cysts with clear evidence ofcalcium-containing fluid within the cystmanifesting as dependent milk of cal-cium on CT scans.

Twenty-five (43%) of 58 mediastinalbronchogenic cysts in our series wereclassified as having soft-tissue attenua-tion on CT scans based on measurementsof internal attenuation. However, 11 ofthese lesions had features that were oth-erwise characteristic of fluid-filled cysts(including homogeneous internal atten-uation, no central enhancement, and athin, often enhancing wall), and mostoccurred in a characteristic location suchas the subcarinal or paratracheal region.Given that another 25 lesions were clearlyfluid filled on CT scans (according tomeasurement of internal attenuation)and that two other lesions contained cal-cium-fluid levels, we believe that a confi-dent diagnosis of bronchogenic cyst couldbe rendered with CT in almost two-thirdsof our cases (38 [66%] of 58 cases). How-ever, we believe that the remaining 20lesions could not be confidently diag-nosed as cysts with CT, either because ofinternal heterogeneity, high attenuationnumbers, streak artifact, lack of mural en-hancement, or atypical location. In suchdifficult cases, we believe that MR imag-ing, by demonstrating markedly increasedsignal intensity within the lesions on T2-weighted images, can be useful for suggest-ing the true cystic nature of the lesion.

Thus, based on our experience, we be-lieve that the following imaging featuresmay be helpful for differentiating non–water-attenuation bronchogenic cysts fromsolid masses. First, the presence of a well-defined, thin, smooth wall suggests thatthe lesion is a cyst. Delineation of thewall is facilitated by intravenous admin-istration of iodinated or gadolinium-based contrast material. A thick or irreg-ular wall is not a typical feature ofbronchogenic cyst and suggests necroticneoplasm or lymphadenopathy. Second,bronchogenic cysts with high CT attenu-ation numbers are usually of homoge-neous attenuation and have cyst con-tents that do not enhance followingadministration of contrast material. In fact,because the cyst contents do not enhance,attenuation differences between medias-tinal soft tissues and the cyst and its con-tents are often accentuated following ad-ministration of contrast material, which

Figure 8. Bronchogenic cyst in a 3-year-oldboy with cough. (a) Coronal T1-weighted MRimage (600/14) shows a well-marginated het-erogeneous mass (arrowheads) in the left pos-terior mediastinum. Signal intensity is similarto that of CSF. (b) Coronal T1-weighted MRimage (750/11) following administration of in-travenous gadolinium-based contrast materialshows mural enhancement (arrows). (c) Trans-verse T2-weighted MR image (2,300/80) showscyst contents of markedly increased signal in-tensity, similar to that of CSF. Note the normalthymus (t) in the anterior mediastinum.

Volume 217 z Number 2 Bronchogenic Cyst: Imaging Features with Clinical and Histopathologic Correlation z 445

facilitates recognition. Conversely, if thelesion is heterogeneous or enhances cen-trally, neoplasia must be excluded. Third,MR imaging can be very useful for differ-entiating high-attenuating cysts on CTscans from soft-tissue masses. Such cystsare typically isointense or hyperintenseto CSF with all pulse sequences. A lesionthat is hypointense to CSF on T2-weightedimages should be viewed with caution.

We acknowledge several limitationsof our study. Because of its retrospectiveand multiinstitutional nature, the CTand MR imaging techniques were in-consistent. In particular, region-of-in-terest measurements were not availablein all cases; attenuation characteristicsin these cases were based on visual cri-teria. It is possible that if region-of-in-terest measurements were available inall cases, fewer lesions would have beenclassified as being of soft-tissue attenu-ation on CT scans. It is also unfortunatethat so few cases were imaged with con-trast-enhanced MR imaging. It is thusdifficult to reach firm conclusions re-garding the utility of this technique forimaging bronchogenic cysts. Lack ofenhancement of cyst contents on con-trast-enhanced MR images, however, mayprove to be a reliable indicator of thecystic nature of problematic broncho-genic cysts.

Appropriate treatment of patients withbronchogenic cysts depends on the pa-tient’s age and symptoms at presenta-tion. Symptomatic cysts should be re-sected (either at thoracotomy or bymeans of video-assisted thoracoscopy) re-gardless of patient age, unless surgicalrisks are unacceptably high (15). It is gen-erally recommended that asymptomaticcysts in young patients be removed (16),

because of the low surgical risk and thepotential risk of late complications (al-beit rare) such as infection, hemorrhage,or neoplasia within the cyst. Conserva-tive (watch-and-wait) treatment has beenadvocated in asymptomatic adults orother high-risk patients (17). Percutane-ous catheter drainage (18) or sterile alco-hol ablation (19) has been performed inselected cases. In such patients, radio-logic diagnosis is of great importance, astissue confirmation will not be obtained.

In summary, bronchogenic cysts typi-cally manifest as sharply marginated,middle mediastinal masses of soft-tissueor water attenuation on CT scans. Mostbronchogenic cysts can be confidentlydiagnosed by using nonenhanced CT;however, administration of contrast ma-terial at CT or the addition of MR imag-ing can be useful for differentiating prob-lematic soft-tissue–attenuation cysts frommediastinal neoplasia.

References1. Stocker JT. The respiratory tract. In:

Stocker JT, Dehner LP, eds. Pediatric pa-thology. Vol 1. Philadelphia, Pa: Lippin-cott, 1992; 517–518.

2. Chapman KR, Rebuck AS. Spontaneousdisappearance of a chronic mediastinalmass. Chest 1985; 87:235–236.

3. Heitzman ER. The lung: radiologic-patho-logic correlation. Vol 1. St Louis, Mo:Mosby, 1984; 17–38.

4. Rogers LF, Osmer JC. Bronchogenic cysts:a review of 46 cases. AJR Am J Roentgenol1964; 91:273–283.

5. Ramenofsky ML, Leape LL, McCauley RG.Bronchogenic cyst. J Pediatr Surg 1979;14:219–224.

6. Hutchin P. Congenital cystic disease ofthe lung. Rev Surg 1971; 28:79–87.

7. St-Georges R, Deslauriers J, Duranceau A,et al. Clinical spectrum of bronchogeniccysts of the mediastinum and lung in theadult. Ann Thorac Surg 1991; 52:6–13.

8. Suen HC, Mathisen DJ, Grillo HC, et al.Surgical management and radiologicalcharacteristics of bronchogenic cysts. AnnThorac Surg 1993; 55:476–481.

9. Ribet ME, Copin MC, Gosselin B. Bron-chogenic cysts of the mediastinum. J Tho-rac Cardiovasc Surg 1995; 109:1003–1010.

10. Mendelson DS, Rose JS, Efremidis SC,Kirschner PA, Cohen BA. Bronchogeniccysts with high CT numbers. AJR Am JRoentgenol 1983; 140:463–465.

11. Glazer HS, Siegel MJ, Sagel SS. Low-atten-uation mediastinal masses on CT. AJRAm J Roentgenol 1989; 152:1173–1177.

12. Nakata H, Sato Y, Nakayama T, Yoshi-matsu H, Kobayashi T. Bronchogenic cystwith high CT number: analysis of con-tents (case report). J Comput Assist To-mogr 1986; 10:360.

13. Yernault JC, Kuhn G, Dumortier P, Roc-mans P, Ketelbant P, De Vuyst P. “Solid”mediastinal bronchogenic cyst: mineral-ogic analysis. AJR Am J Roentgenol 1986;146:73–74.

14. Lyon RD, McAdams HP. Mediastinalbronchogenic cyst: demonstration of afluid-fluid level at MR imaging. Radiology1993; 186:427–428.

15. Patel SR, Meeker DP, Biscotti CV, KirbyTJ, Rice TW. Presentation and manage-ment of bronchogenic cysts in the adult.Chest 1994; 106:79–85.

16. Cioffi U, Bonavina L, De Simone M, et al.Presentation and surgical management ofbronchogenic and esophageal duplica-tion cysts in adults. Chest 1998; 113:1492–1496.

17. Bolton JW, Shahian DM. Asymptomaticbronchogenic cysts: what is the best man-agement? Ann Thorac Surg 1992; 53:1134–1137.

18. Adam A, MacSweeney JE, Whyte MK,Smith PL, Ind PW. CT-guided extrapleu-ral drainage of bronchogenic cyst. J Com-put Assist Tomogr 1989; 13:1065–1068.

19. Johnston SR, Adam A, Allison DJ, SmithP, Ind PW. Recurrent respiratory obstruc-tion from a mediastinal bronchogeniccyst. Thorax 1992; 47:660–662.

446 z Radiology z November 2000 McAdams et al