Pituitary adenoma, craniopharyngioma, and Rathke cleft cyst involving both intrasellar and suprasellar regions

Pituitary adenoma, craniopharyngioma, and Rathke cleft cyst involving both intrasellar and suprasellar regions: differentiation using MRI S.H. Choia, B.J. Kwona,*, D.G. Naa, J.-H. Kima, M.H. Hana,b,c, K.-H. Changa,b,c
aDepartment of Radiology, Seoul National University College of Medicine, bClinical Research Institute, Seoul National University Hospital, and cInstitute of Radiation Medicine, Seoul National University Medical Research Center, Seoul, Republic of Korea
Received 28 June 2006; received in revised form 23 November 2006; accepted 7 December 2006
AIMS: To determine the differential magnetic resonance imaging (MRI) features of pituitary adenoma, craniopharyngioma, and Rathke cleft cyst involving both intrasellar and suprasellar regions. MATERIALS AND METHODS: The MRI images of 64 patients with pituitary adenoma (n¼38), craniopharyngioma( n¼13), or Rathke cleft cyst (n¼13) were retrospectively reviewed by three neuroradiologists. The following characteristics were evaluated: shape, volume, extent, component characteristics, signal intensities of solid portions on T2-weighted images, signal intensities of cystic portions on T1-weighted images, and enhancement patterns of solid portions and cyst walls of tumours. Fisher’s exact test applied with Bonferroni correction was used for multiple comparison. A ?owchart for differential diagnosis was constructed based on statistical analysis of the results. RESULTS: A snowman shape, solid characteristics, and homogeneous enhancement of the solid portion were more common in pituitary adenomas (p <0.017). A superiorly lobulated shape, third ventricle compression by superior tumour extension, mixed solid and cystic characteristics, and reticular enhancement of the solid portion were more common in craniopharyngiomas (p <0.017). Finally, an ovoid shape, a small tumour volume, cystic characteristics, and no or thin cyst wall enhancement were more common in Rathke cleft cysts (p <0.017). The ?owchart yielded diagnostic accuracies as follows: 92.1% in pituitary adenoma; 92.3% in craniopharyngioma; 92.3% in Rathke cleft cyst; and 92.2% overall. CONCLUSION: A combination of MRI ?ndings is helpful in the differential diagnosis of the three tumours involving both intrasellar and suprasellar regions. ª 2007 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction
Pituitary adenoma is the most common of the lesions involving both intrasellar and suprasellar regions.1e3 Although an uncomplicated pituitary adenoma typically shows a signal paralleling the grey matter, and homogeneously intense enhancement after contrast medium administration on magnetic resonance imaging (MRI) images, various
signal and enhancement patterns related to necrosis, cystic degeneration, and haemorrhage are not uncommon.2,4 Craniopharyngioma is an epithelial neoplasm arising from squamous epithelial rests of the Rathke pouch, and is the most heterogeneous of the lesions involving the sellar region due to their cystic and solid components.2 Although most craniopharyngiomas are con?ned to the suprasellar region, intrasellar involvement by a large craniopharyngiomas is found in about 21%.5 Rathke cleft cyst is a benign, epithelium-lined cyst that is thought to be a remnant of the Rathke pouch, and approximately 70% of Rathke cleft cysts involve both intrasellarand suprasellar regions.6e10
* Guarantor and correspondent: B.J. Kwon, Department of Radiology, Seoul National University Hospital, 28, Yongondong, Chongno-gu, Seoul 110-744, Republic of Korea. Tel.: þ82 2 2072 1969; fax: þ82 2 743 6385. E-mail address: bjkwon@radiol.snu.ac.kr (B.J. Kwon).
0009-9260/$ – see front matter ª 2007 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2006.12.001
Clinical Radiology (2007) 62, 453e462
Therefore, in cases with a tumour involving intrasellar and suprasellar regions, the differential diagnosis usually includes pituitary adenoma, craniopharyngioma, and Rathke cleft cyst.11,12 When a lesion is encountered that involves both intrasellar and suprasellar regions, the differential diagnosis is important, because surgical planning depends on the diagnosis: most pituitary adenomas are approached by the trans-sphenoidal route, craniopharyngiomas usually require craniotomy and radical surgery, and symptomatic Rathke cleft cysts are managed by surgical drainage with partial excision of the cyst wall.13e15 However, it is not always easy to make a differential diagnosis between these three tumours, especially when the tumour has a solid and cystic, or cystic characteristics. However, to the authors’ knowledge, there are no reports regarding the differential MRI ?ndings of these three tumours involving both intrasellar and suprasellar regions in a large series, although the respective MRI ?ndings of these three tumours are well known. Therefore, we sought to identify the differential MRI features of these three tumours and to produce a schematic approach based on their differential features.
Methods
Study population
A single investigator (S.H.C) searched the adult (over18yearsold)pathologydatabasefromJanuary 1998 through February 2004 using the terms: pituitary adenoma, craniopharyngioma, and Rathke cleftcyst,andidenti?ed232patients.Onehundred and seventy-eight patients were excluded from the analysis by the author (S.H.C), and the exclusion criteria were as follows: (a) MRI was not performed or MRI images were not available; (b) the tumour involved only the intrasellar or suprasellar region; (c) pathological examination revealed mixed disease; or (d) surgeryor biopsywas performed before MRI. The remaining 64 patients with pituitary adenoma (n¼38), craniopharyngioma (n¼13) and Rathke cleft cyst (n¼13) were included in this study. Patients’ ages ranged between 24 and 80 years (mean 48 years). Thirty-two patients were men(agerange,25e80years;meanage,49.6years) and 32 women (age range, 24e72 years; mean age, 46.7 years). For this type of study, institutional review board approval was not required by either of the participating institutions; informed consent was not required because patient anonymity was maintained.
In 38 patients with pituitary adenoma (10 men and 28 women; age range 25e80 years; mean age 50.1 years), masses were detected on work-ups for visual disturbance (n¼27), amenorrhoea (n¼4), headache (n¼4), acromegaly (n¼2), or cerebrospinal ?uid rhinorrhoea (n¼1). Eighteen of the 38 pituitary adenomas were functioning: four were growth hormone (GH) producing, seven prolactin producing, four adrenocorticotropin hormone (ACTH) producing, and three folliclestimulating hormone (FSH) producing. In the 13 patients (seven men and six women; age range 24e60 years; mean age 43.8 years) with craniopharyngioma (squamous-papillary type in four and adamantinous type in nine), MRI was performed because of visual disturbance (n¼12) or galactorrhoea (n¼1). Finally, the 13 patients (seven men and six women; age range 33e65 years; mean age 46.7 years) with Rathke cleft cyst presented with headache (n¼7) or visual disturbance (n¼6).
MRI protocols
In all 64 patients MRI was performed with 1.5 T MRI machine (Signa, GE Medical systems, Milwaukee, WI, USA, or Magnetom Vision, Siemens, Erlangen, Germany). Imaging sequences included at least axial T2-weighted, sagittal and coronal T1-weighted, and contrast-enhanced T1-weighted sagittal and coronal sequences, which were obtained at 30e60 s after an intravenous bolus injection of 0.1 mmol/kg gadopentetate dimeglumine (Magnevist, Schering AG, Berlin, Germany). Section thickness was 5 mm with a 0.5-mm intersection gap. T1-weighted spin-echo images were obtained at 440e650 ms/8e24 ms (repetition time/echo time) with two to four signals acquired. T2weighted fast spin-echo images were obtained at 4000e5000 ms/96e122 ms with two to four signals acquired.
Evaluation of MRI ?ndings
All MRI images were reviewed on a picture archiving and communication system (PACS; Marotech, Seoul, Korea) workstation. The cases were randomly listed on PACS by one author (S.H.C.). Two neuroradiologists (B.J.K. and D.G.N. with 10 and 4 years of experience, respectively) evaluated MRI ?ndings. The reviewers were blinded to the diagnosis and clinical history, but were informed that a ?nal diagnosis was one of the three tumours (pituitary adenoma, craniopharyngioma, or Rathke cleft cyst).
454 S.H. Choi et al.
Tumourshapeswereclassi?edasovoid,snowmanlike, or superiorly or inferiorly lobulated. A snowman shape was de?ned as a ?gure of eight-like shape, and superiorly or inferiorly lobulated shape as having two or more lobes in the suprasellar or intrasellar compartments, respectively. Tumour size was de?ned as the largest anterioposterior, superoinferior, and transverse dimensions on MRI images, and tumour volumes were calculated using the following formula: volume¼0.5 anteroposteriorsuperoinferiortransverse dimensions.16,17 Finally, tumour volumes were classi?ed as large (2 ml) or small (<2 ml). Superior extensions were graded as below the optic chiasm, compressing the optic chiasm, and compressing the third ventricle. Compression of the third ventricle was considered to occur if the third ventricle ?oor was indented. Inferior extension was de?ned based on the presence or absence of a sellar ?oor depression, which was in turn de?ned as a depression 10 mm below an imaginary horizontal line posteriorly drawn from the planum sphenoidale.18 A lateral extent was classi?ed as one within or beyond the lateral margin of the cavernous intracranial carotid artery (ICA). The enhancing portion of the tumours was considered to be the solid portion, and cystic portions were de?ned as homogeneous, non-enhancing, sharply delineated areas on MRI images.17e21 Tumour characteristics were classi?ed as solid, mixed solid and cystic, including predominantly solid and predominantly cystic, or cystic. Signal intensities of the solid portion of tumours on T2-weighted MRI images were classi?ed as hyperintense (higher than the signal intensity of grey matter) or not. Signal intensities of the cystic portion of tumours on T1-weighted MRI images were also classi?ed as hyperintense (higher than the signal intensity of whiter matter), or not. Enhancement patterns of solid portions were classi?ed as homogeneous or reticular (mesh-like enhancements with intervening non-enhancing tiny defects). Enhancement patterns of cyst walls were classi?ed as none, thin (enhancing wall thickness was less than 2 mm), or thick (2 mm or more than 2 mm). The largest solid or cystic portions were evaluated in cases with predominantly solid or predominantly cystic characteristics with reference to signal intensities and enhancement patterns.
Schematic approach for differential diagnosis
Signi?cantly different frequencies between the MRI ?ndings of tumours were revealed by the statistical analysis performed during this study.
Based on of these ?ndings, a ?owchart with a schematic approach was constructed to guide differential diagnosis. In addition, a consensus between three neuroradiologists (K.H.C, D.G.N and B.J.K.) was considered at each step of the schematic approach. Flowchart-based diagnosis was performed in every case and diagnostic accuracies were calculated versus the histopathological diagnostic results.
Statistical analysis
To evaluate MRI ?ndings, the frequencies of individual imaging ?nding were compared for the three tumour groups using Fisher’s exact test. Because of the three number of tests performed, to minimize a(ie, type I) error, Bonferroni correction was generally applied to all analyses in which multiple tests were performed.22,23 The signi?cance level (p¼0.05) was therefore reduced to an a-adjusted p level of 0.017. The diagnostic accuracy of ?owchart-based diagnoses was determined by using histopathological diagnoses as reference standards. All statistical analyses were performed using the SPSS software package (version 13.0; SPSS, Chicago, IL, USA).
Results
Evaluation of MRI ?ndings
The variety of the MRI features of pituitary adenoma, craniopharyngioma, and Rathke cleft cyst are summarized in the Table 1. The most common shapes of pituitary adenomas, craniopharyngiomas, and Rathke cleft cyst were a snowman appearance (55.3%, 21 of 38 patients), superior lobulation (69.2%, nine of 13 patients), and ovoid (69.2%, nine of 13 patients), respectively. In addition, there were statistically signi?cant differences between groups (pituitary adenoma versus craniopharyngioma; p¼0.001, craniopharyngioma versus Rathke cleft cyst; p<0.001, and pituitary adenoma versus Rathke cleft cyst; p¼0.012). In terms of tumour volume, a small volume was more frequently observed in Rathke cleft cysts (84.6%, 11 of 13 patients) than in pituitary macroadenoma (p¼0.003) or craniopharyngioma (p¼0.015), and tumour volumes were not significantly different between pituitary macroadenoma and craniopharyngioma (p¼1.000). With respect to superior extent, compression of theopticchiasm(Figs.1and2)wasmostcommonin pituitary adenomas (71.1%, 27 of 38 patients) without a statistically signi?cant difference. By
Pituitary adenoma, craniopharyngioma, and Rathke cleft cyst 455
contrast, compression of the third ventricle (Figs. 3 and 4) was more commonly observed in craniopharyngiomas (61.5%, eight of 13 patients) than in pituitary adenoma (p¼0.004) or Rathke cleft cyst (p¼0.007). In Rathke cleft cysts (Fig. 5), a superior extent the below optic chiasm (53.8%, seven of 13 patients)wasmorecommonlyseenthaninothertumours without a statistically signi?cant difference. Sellar ?oor depression (Fig. 6) was more frequently observed in pituitary adenomas (44.7%, 17 of 38 patients) without a statistically signi?cant difference than in craniopharyngioma (p¼0.019) or Rathke cleft cyst (p¼0.019). In all cases of Rathke cleft cyst, lesion lateral extent was within the lateral wall of the ICA
(Fig. 5), and this was statistically signi?cant versus both pituitary adenoma (p¼0.005) and craniopharyngioma (p¼0.002). In terms of tumour characteristics, pituitary adenomas (Fig. 1) were more frequently of the solid type (52.3%, 20 of 38 patients), craniopharyngiomas (Figs. 3 and 7) most commonly had mixed characteristics, including both predominantly solid and predominantly cystic characteristics (76.2%, 10 of 13 patients), and in Rathke cleft cysts (Figs. 5 and 8), the most common was the cystic type (100%, all 13 patients). In addition, there were statistically signi?cant differences between groups (pituitary adenoma versus craniopharyngioma;p<0.001,craniopharyngiomaversusRathke
Table 1 Magnetic resonance imaging ?ndings of pituitary adenoma, craniopharyngioma, and rathke cleft cyst involving both intrasllar and suprasellar regions MRI ?ndings Pituitary adenoma (n¼38) Craniopharyngioma (n¼13) Rathke cleft cyst (n¼13) Shape Ovoid 2 (5.3) 2 (15.4) 9 (69.2)a Snowman-like appearance 21 (55.3)a 1 (7.7) 1 (7.7) Superior lobulation 8 (21.1) 9 (69.2)a 3 (23.1) Inferior lobulation 7 (18.4) 1 (7.7) 0 (0) Tumour volume Large (2 ml) 25 (65.8) 9 (69.2) 2 (15.4) Small (<2 ml) 13 (34.2) 4 (30.1) 11 (84.6)a
Superior extent Below optic chiasm 6 (15.8) 1 (7.7) 7 (53.8) Compressing optic chiasm 27 (71.1) 4 (30.1) 5 (38.5) Compressing third ventricle 5 (13.1) 8 (61.5)a 1 (7.7)
Inferior extent Absence of sellar ?oor depression 21 (55.3) 12 (92.3) 12 (92.3) Presence of sellar ?oor depression 17 (44.7) 1 (7.7) 1 (7.7) Right lateral extent Within lateral wall of cavernous ICA 30 (78.9) 10 (76.9) 13 (100)a Beyond lateral wall of cavernous ICA 8 (21.1) 3 (23.1) 0 (0) Left lateral extent Within lateral wall of cavernous ICA 29 (76.3) 7 (53.8) 13 (100)a Beyond lateral wall of cavernous ICA 9 (23.7) 6 (46.2) 0 (0) Component characteristicsb Solid 20 (52.3)a 0 (0) 0 (0) Predominantly solid 11 (28.9) 4 (30.1)a 0 (0) Predominantly cystic 4 (10.5) 6 (46.2)a 0 (0) Cystic 3 (7.9) 3 (23.1) 13 (100)a
Signal intensity of solid portion on T2 weighted image
Hyperintense 29 (87.9) 9 (90.0) 0 (0) Iso- or hypointense 6 (18.2) 1 (10.0) 0 (0)
Signal intensity of cystic portion on T1-weighted image
Hyperintense 2 (11.1) 5 (38.5) 7 (53.8) Iso- or hypointense 16 (88.9) 8 (61.5) 6 (46.2)
Enhancement pattern of solid portion
Homogeneous 33 (94.3)a 0 (0) 0 (0) Reticular 2 (5.7) 10 (100)a 0 (0)
Enhancement pattern of cyst wallc
No 0 (0) 1 (7.7) 6 (46.2)a Thin (<2 mm) 5 (29.4) 4 (30.8) 6 (46.2)a Thick (2 mm) 12 (70.6) 8 (61.5) 1 (7.7) Data in parentheses are percentages of patients. ICA, internal carotid artery. a There were statistically signi?cant differences from both of the other two tumour groups (p<0.017, Fisher’s exact test with application of Bonferroni correction). b Mixed solid and cystic tumours including predominantly solid and predominantly cystic tumours were considered as a single group in the Fisher’s exact test. c Lack of and thin cyst wall enhancement patterns were considered as a single group in the Fisher exact test.
456 S.H. Choi et al.
cleft cyst; p<0.001, and pituitary adenoma versus Rathke cleft cyst; p <0.001). There was no statistically signi?cant difference in the signal intensity of the solid portion on T2weighted images between pituitary adenoma and craniopharyngioma (p¼1.000). In addition, in terms of the signal intensity of the cystic portion on T1-weighted images, although the high signal intensity of the cystic portion (Fig. 5) was more frequently observed in Rathke cleft cyst (53.8%, seven of 13 patients) than in other tumours, there were no statistically signi?cant differences. In terms of the enhancement patterns of solid portions, 33 of 35 pituitary adenomas (94.3%) with a solid portion showed a homogenous pattern (Figs. 1B and 6), whereas all 10 craniopharyngioma
cases with a solid portion had the reticular pattern (Figs. 3 and 7B). The statistical difference between the frequencies of these enhancement patterns was strong for pituitary adenomas and craniopharyngiomas (p <0.001). The enhancement patterns of the cystic portions were also different for the three tumours. Lack of or thin enhancement pattern of the cyst wall (Figs. 5c and 8) were most frequently observed in Rathke cleft cyst (92.3%, 12 of 13 patietns), and there was a statistically signi?cant difference in comparison with cyst wall enhancement pattern of both pituitary adenoma (p¼0.001) and craniopharyngioma
Figure 1 Solid pituitary adenoma in a 50-year-old woman. (a) Non-enhanced sagittal T1-weighted MRI image (450/12) shows an isointense, snowman-shaped, homogeneous pituitary adenoma with sellar ?oor depression (arrow). (b) Coronal T1-weighted MRI image (500/12) obtained after gadopentetate dimeglumine administration shows a homogeneously enhancing pituitary adenoma with optic chiasm compression (white arrow), which invades the left cavernous sinus and surrounds the left internal carotid artery (black arrow).
Figure 2 Cystic pituitary adenoma in a 43-year-old man. Coronal T1-weighted MRI image (536/17) obtained after gadopentetate dimeglumine administration shows a cystic pituitary adenoma with thick wall enhancement (long arrow), which compresses the optic chiasm (short arrow).
Figure 3 Craniopharyngioma with a predominantly solid characteristic in a 46-year-old woman. Sagittal T1weighted MRI image (570/17) obtained after gadopentetatedimeglumineadministrationshowsthenon-enhancing cysticportion(longarrow), and a solidportion with reticular enhancement (short arrow). The superior portion compresses the third ventricle ?oor.
Pituitary adenoma, craniopharyngioma, and Rathke cleft cyst 457
(p¼0.011). These enhancement patterns are summarized in Table 2. The retrospective application of signi?cant ?ndings in terms of the frequencies of the above in each tumour type yielded sensitivities and speci?cities for tumour diagnosis. The sensitivities and speci?cities of statistically signi?cant MRI ?ndings are summarized in Table 3.
Flowchart-based diagnoses
The ?owchart using MRI ?ndings for the differential diagnosis of the three tumours is illustrated on Fig. 9, in which major MRI ?ndings used in the ?owchart were tumour characteristics, and enhancement patterns of solid potions and cyst walls. When the ?owchart-based diagnosis was tested for diagnostic accuracy, it resulted in the correct diagnosis of 35 of 38 pituitary adenomas (92.1%), 12 of 13 craniopharyngiomas (92.3%), and 12 of 13 Rathke cleft cysts (92.3%). The overall accuracy of this model of diagnosis was 92.2%.
Discussion
Clinically, the differential diagnosis of pituitary adenoma, craniopharyngioma, and Rathke cleft cyst involving both intrasellar and suprasellar regions is not easy because patients with these tumours usually present with nonspeci?c features such as headache, visual disturbance, or hypopituitarism.2,11,24,25 In contrast, the characteristic MRI ?ndings of these tumours are well known, and their differential diagnosis is not problematic
Figure 4 Cystic craniopharyngioma in a 38-year-old man. Sagittal T1-weighted MRI image (500/14) obtained after gadopentetate dimeglumine administration shows a cystic tumour with thick wall enhancement (long arrows), which compresses the third ventricle ?oor. In addition, a superior lobulation of the tumour was also noted.
Figure 5 Rathke cleft cyst in a 45-year-old man. (a) Non-enhanced sagittal T1-weighted MRI image (500/14) shows an ovoid, homogeneous, hyperintense tumour with sellar ?oor depression (arrow). (b) Sagittal T1weighted MRI image (500/14) obtained after gadopentetate dimeglumine administration shows a cystic tumour without cyst wall enhancement. (c) Coronal T1-weighted MRI image (500/14) obtained after gadopentetate dimeglumine administration shows a snowman-shaped Rathke cleft cyst with superior extent below the optic chiasm (short arrow) and with lateral extent within the lateral wall of the internal carotid artery.
458 S.H. Choi et al.
in the majority of cases.2,5,12,13,17e21,24e30 Pituitary adenomas with suprasellar extension typically have a ‘‘?gure of eight’’ or ‘‘snowman’’ appearance, and are the most strongly enhanced of the three.4 In craniopharyngiomas, two histological types, squamous-papillary and adamantinous, have been described in the literature, although in some cases craniopharyngiomas cannot be divided into distinct histological types.17,25 On MRI the typical features of squamous-papillary craniopharyngioma include a predominantly solid or mixed solid and cystic spherical tumour in the suprasellar region.25 The solid tumour region is inhomogeneous but intensely enhancing with small necrotic areas.25 Conversely, adamantinous craniopharyngioma is a cystic or predominantly cystic, lobulate tumour, which is often observed in the intrasellar or suprasellar regions.25 On T1weighted images, typically single or multiple hyperintense cysts with thin peripheral enhancing rims are present in adamantinous craniopharyngioma.25 Rathke cleft cysts usually have the following imaging features: a sellar epicentre, smooth contour, lack of calci?cation, lack of internal enhancement, and a homogeneous signal intensity within the lesion.26 Despite these well-known imaging ?ndings, it is challenging to arrive at differential diagnosis in a subset of these tumours. The present results show that several imaging ?ndings were signi?cantly more frequent in some tumour types than in others. Tumour shapes, extents, and volumes were signi?cantly different for the three tumours, although all of the tumours
involved both intrasellar and suprasellar regions. Most pituitary adenomas arise from the pituitary gland extending through the diaphragm sella up to the optic chiasm; most craniopharyngiomas arise from the suprasellar region extending up to the third ventricle with lobulation and down to the intrasellar region; and most Rathke cleft cysts arising from the pituitary fossa have small volumes due to relatively small extents. The presence of the lateral extent beyond the lateral wall of the ICA may helpfully rule out a Rathke cleft cyst. In addition to these differential ?ndings, the results of the present study also show that tumour characteristics and enhancement patterns can be accurately used in the diagnostic ?owchart produced to differentiate these three tumours.
Figure 6 Pituitary adenoma with a predominantly cystic characteristic in a 44-year-old woman. Sagittal T1-weighted MRI image (500/12) obtained after gadopentetate dimeglumine administration shows a predominantly cystic tumour composed of homogeneously enhancing solid portion (short arrow) and cystic portion with thick wall enhancement (long arrow). In addition, sellar ?oor depression was also noted.
Figure 7 Craniopharyngiomawithpredominantlyacystic characteristic in a 36-year-old man. (a) Axial T2weighted MRI image (5000/104) shows a heterogeneous hyperintense sellar tumour. (b) Sagittal T1-weighted MRI image (500/12) obtained after gadopentetate dimeglumine administration shows a craniopharyngioma composed of a solid portion with a reticular enhancement pattern (short arrow) and a cystic portion with thin wall enhancement (long arrow).
Pituitary adenoma, craniopharyngioma, and Rathke cleft cyst 459
Typical component characteristics of each tumour was the most accurate MRI ?nding (67%, 43 of 64 patients) in the differential diagnosis of the three tumours, the enhancement pattern of solid portions was the most accurate MRI ?nding (96%, 43 of the 45 patients) in the differential diagnosis in the tumours containing solid portions, and the cyst wall enhancement pattern was the most accurate MRI ?nding (84%, 16 of the 19 patients) in differentiating Rathke cleft cysts from the cystic tumours (Tables 1 and 2). In pituitary adenomas, solid characteristics proved speci?c to make a diagnosis. Many cases of pituitary adenomas and craniopharyngiomas having mixed solid and cyst characteristics were included in the present study, and this fact might have been due to the large volumes of the tumours. Large tumours might be prone to
have cystic or haemorrhagic changes resulting in mixed or cystic characteristics. Homogeneous and reticular enhancement patterns were found to be key elements in the differential diagnosis of solid containing pituitary adenomas and craniopharyngiomas, respectively. Although squamous-papillary craniopharyngiomas presenting as homogeneously enhancing solid masses were reported by Sartoretti-Schefer et al.,25 a reticular enhancement pattern was observed in all solid-containing craniopharyngiomas in the present study. Histopathological reviews of published studies suggest that the observed reticular enhancement pattern of solidcontaining craniopharyngiomas is caused by small regions of necrosis, keratin debris, or calci?cation.17,25 A cystic nature was found to be highly sensitive (100%) and speci?c (88%) for a diagnosis of Rathke cleft cyst, and lack of or thin cyst wall enhancement pattern is the main MRI ?nding in terms of differentiating Rathke cleft cyst from other neoplastic cysts involving both intrasellar and suprasellar regions. Recent studies suggest that the thin rim of peripheral enhancement of Rathke cleft cyst could be ascribed to changes due to squamous metaplasia, in?ammation, or deposition of haemosiderin or cholesterol crystals in the cyst wall.15,26e29 However, cystic portions of both pituitary adenoma and craniopharyngioma are frequently found to have a thick wall enhancement pattern in the present. The cystic portions of the three tumours are believed to be composed of various entities: necrosis, haemorrhage, and cystic degeneration in pituitary adenoma2,4; and a high protein concentration, cholesterol, or methaemoglobin in craniopharyngioma and Rathke cleft cyst.21,24e26 In a previous study, Hua et al.30 concluded that the signal intensity of the cystic ?uid
Figure 8 Rathke cleft cyst in a 33-year-old woman. Sagittal T1-weighted MRI image (500/14) obtained after gadopentetate dimeglumine administration shows a cystic tumour with smooth, thin wall enhancement (long arrow).
Table 2 Enhancement patterns of the solid portion and cyst wall on contrast enhanced magnetic resonance imaging images Component characteristics Enhancement patterns Pituitary adenoma (n¼38) Craniopharyngioma (n¼13) Rathke cleft cyst (n¼13) Solid tumour Solid enhancement Homogeneous 20 (100) 0 (0) 0 (0) Reticular 0 (0) 0 (0) 0 (0)
Predominantly solid tumour
Solid enhancement Homogeneous 10 (90.9) 0 (0) 0 (0) Reticular 1 (9.1) 4 (100) 0 (0) Cyst wall enhancement Thin (<2 mm) 2 (18.2) 2 (50.0) 0 (0) Thick (2 mm) 9 (81.8) 2 (50.0) 0 (0)
Predominantly cystic tumour
Solid enhancement Homogeneous 3 (75.0) 0 (0) 0 (0) Reticular 1 (25.0) 6 (100) 0 (0) Cyst wall enhancement Thin (<2 mm) 3 (75.0) 2 (33.3) 0 (0) Thick (2 mm) 1 (25.0) 4 (66.7) 0 (0) Cystic tumour Cyst wall enhancement None 0 (0) 1 (33.3) 6 (46.2) Thin (<2 mm) 1 (33.3) 0 (0) 6 (46.2) Thick (2 mm) 2 (66.7) 2 (66.7) 1 (7.7) Data in parentheses are percentages of patients.
460 S.H. Choi et al.
did not help in distinguishing Rathke cleft cyst from cystic neoplasms. The results of the present study are concordant in part with this conclusion, i.e., thathyper-intensitiesonT1-weightedimagesofcystic portion were most frequently observed in Rathke cleft cysts without statistical signi?cance. Apart from the intrinsic limitations of retrospective analysis, there are other limitations of the present study. First, MRI ?ndings in the cases studied were not correlated with detailed pathological ?ndings such as solid portions with reticular enhancement in craniopharyngioma, the chemical characteristics of cystic components, and enhancing components of cyst walls. However, possible or proven explanations in previous studies may indicate a pathological ?nding. Second, the most frequently occurring lesions involving both intrasellar and suprasellar regions were included in the
present study, but lesions such as meningiomas, aneurysms, granulomatous diseases, metastatic diseases, and gliomas were not included.2 This is because the aim was to determine the differential MRI ?ndings in the most common three lesions, it would be outside of the scope of this article to describe other rarely encountered lesions. In conclusion, MRI features, including tumour shape, superior extent, characteristics, and enhancement patterns of solid portions and cyst walls are useful in the differentiation of pituitary adenoma, craniopharyngioma, and Rathke cleft cyst involving both intrasellar and suprasellar regions. Moreover, the schematic approach was found to differentiate accurately these three tumours, and a prospective study based on these results is warranted.
References
1. Smith MV, Laws Jr ER. Magnetic resonance imaging measurements of pituitary stalk compression and deviation in patients with nonprolactin-secreting intrasellar and parasellar tumors: lack of correlation with serum prolactin levels. Neurosurgery 1994;34:834e9. 2. Johnsen DE, Woodruff WW, Allen IS, et al. MR imaging of the sellar and juxtasellar regions. RadioGraphics 1991;11: 727e58. 3. Osborn AG. Pituitary macroadenoma. In: Osborn AG, editor. Diagnostic imaging: brain. Salt lake city, Utah: Amirsys; 2004. p. II-2-24-27. 4. Osborn AG. Pituitary apoplexy. In: Osborn AG, editor. Diagnostic imaging: brain. Salt lake city, Utah: Amirsys; 2004. p. II-2-28-31. 5. HedlundGL.Craniopharyngioma.In:OsbornAG,editor.Diagnostic imaging: brain. Salt lake city, Utah: Amirsys; 2004. p. II-2-32-35. 6. Fager CA, Carter H. Intrasellar epithelial cysts. J Neurosurg 1966;24:77e81. 7. Shanklin WM. The histogenesis and histology of an integumentary type of epithelium in the human hypophysis. Anat Rec 1951;109:217e31.
Table 3 Sensitivity and speci?city calculated by using signi?cantly frequent magnetic resonance imaging ?ndings in the diagnosis of pituitary adenoma, craniopharyngioma, and Rathke cleft cyst Tumours Findings Sensitivity (%) Speci?city (%) Pituitary adenoma Snowman shape 55 (21/38) 92 (24/26) Solid characteristics 53 (20/38) 100 (26/26) Homogeneous enhancement of solid portion 94 (33/35) 100 (10/10)
Craniopharyngioma Superior lobulation shape 69 (9/13) 78 (40/51) Superior extent of compressing the third ventricle 62 (8/13) 90 (46/51) Mixed solid and cystic characteristics 77 (10/13) 71 (36/51) Reticular enhancement of solid portion 100 (10/10) 94 (33/35)
Rathke cleft cyst Ovoid shape 69 (9/13) 92 (47/51) Small tumour volume (<2 ml) 85 (11/13) 67 (34/51) Lateral extent within lateral wall of cavernous ICA 100 (13/13) 37 (19/51) Cystic characteristics 100 (13/13) 88 (45/51) No or thin enhancement pattern of cyst wall 92 (12/13) 63 (19/30) Data in parentheses are numbers of patients.
Figure 9 Flowchart showing the schematic approach used for differential diagnosis of tumours involving both intrasellar and suprasellar regions.
Pituitary adenoma, craniopharyngioma, and Rathke cleft cyst 461
8. Steinberg GK, Koenig GH, Golden JB. Symptomatic Rathke’s cleftcysts.Reportoftwocases.JNeurosurg1982;56:290e5. 9. Yoshida J, Kobayashi T, Kageyama N, et al. Symptomatic Rathke’s cleft cyst. Morphological study with light and electron microscopy and tissue culture. J Neurosurg 1977;47: 451e8. 10. Voelker JL, Campbell RL, Muller J. Clinical, radiographic, and pathological features of symptomatic Rathke’s cleft cysts. J Neurosurg 1991;74:535e44. 11. Poussaint TY, Barnes PD, Anthony DC, et al. Hemorrhagic pituitary adenomas of adolescence. AJNR Am J Neuroradiol 1996;17:1907e12. 12. Majos C, Coll S, Aguilera C, et al. Imaging of giant pituitary adenomas. Neuroradiology 1998;40:651e5. 13. Cappabianca P, Cirillo S, Al?eri A, et al. Pituitary macroadenoma and diaphragma sellae meningioma: differential diagnosis on MRI. Neuroradiology 1999;41:22e6. 14. Van Effenterre R, Boch AL. Craniopharyngioma in adults and children: a study of 122 surgical cases. J Neurosurg 2002; 97:3e11. 15. KimJE,KimJH,KimOL,etal.SurgicaltreatmentofsymptomaticRathkecleftcysts:clinicalfeaturesandresultswithspecial attention to recurrence. J Neurosurg 2004;100:33e40. 16. Lundin P, Pedersen F. Volume of pituitary macroadenomas: assessmentbyMRI.JComputAssistTomogr1992;16:519e28. 17. Eldevik OP, Blaivas M, Gabrielsen TO, et al. Craniopharyngioma: radiologic and histologic ?ndings and recurrence. AJNR Am J Neuroradiol 1996;17:1427e39. 18. Hagiwara A, Inoue Y, Wakasa K, et al. Comparison of growth hormone-producing and non-growth hormone-producing pituitary adenomas: imaging characteristics and pathologic correlation. Radiology 2003;228:533e8. 19. Ahmadi J, Destian S, Apuzzo ML, et al. Cystic ?uid in craniopharyngiomas: MR imaging and quantitative analysis. Radiology 1992;182:783e5.
20. Hald JK, Eldevik OP, Skalpe IO. Craniopharyngioma identi?cation by CT and MR imaging at 1.5 T. Acta Radiol 1995;36: 142e7. 21. Hayashi Y, Tachibana O, Muramatsu N, et al. Rathke cleft cyst: MR and biomedical analysis of cyst content. J Comput Assist Tomogr 1999;23:34e8. 22. Bonferroni CE. Teoria statistica delle classi e calcolo delle probabilita `. In: Pubblicazioni del R Istituto Superiore di Scienze Economiche e Commerciali di Firenze, vol 8. Florence, Italy: University of Florence; 1936. 23. Miller Jr RG. Simultaneous statistical inference. New York, NY: Springer-Verlag; 1981. 24. Kucharczyk W, Peck WW, Kelly WM, et al. Rathke cleft cysts: CT, MR imaging, and pathologic features. Radiology 1987;165:491e5. 25. Sartoretti-Schefer S, Wichmann W, Aguzzi A, et al. MR differentiation of adamantinous and squamous-papillary craniopharyngiomas. AJNR Am J Neuroradiol 1997;18: 77e87. 26. Naylor MF, Scheithauer BW, Forbes GS, et al. Rathke cleft cyst: CT, MR, and pathology of 23 cases. J Comput Assist Tomogr 1995;19:853e9. 27. Sumida M, Uozumi T, Mukada K, et al. Rathke cleft cysts: correlation of enhanced MR and surgical ?ndings. AJNR Am J Neuroradiol 1994;15:525e32. 28. Okamoto S, Handa H, Yamashita J, et al. Computed tomography in intra- and suprasellar epithelial cysts (symptomatic Rathke cleft cysts). AJNR Am J Neuroradiol 1985;6: 515e9. 29. Oka H, Kawano N, Suwa T, et al. Radiological study of symptomatic Rathke’s cleft cysts. Neurosurgery 1994;35: 632e6. 30. Hua F, Asato R, Miki Y, et al. Differentiation of suprasellar nonneoplastic cysts from cystic neoplasms by Gd-DTPA MRI. J Comput Assist Tomogr 1992;16:744e9.
462 S.H. Choi et al.
TO GET YOUR ASSIGNMENTS DONE AT A CHEAPER PRICE, PLACE THIS ORDER OR A SIMILAR ORDER WITH US NOW.