Glioblastoma multiforme with atypical diffusion ... - BIR Publications

The British Journal of Radiology, 78 (2005), 989–992 DOI: 10.1259/bjr/12830378

E

2005 The British Institute of Radiology

Glioblastoma multiforme with atypical diffusion-weighted MR findings 1,2

B HAKYEMEZ, MD, 1C ERDOGAN, MD, 1N YILDIRIM, MD and 1M PARLAK, MD

1

Department of Radiology, Uludag University Medical School, Bursa and 2Bursa State Hospital, Bursa, Turkey

Abstract. The aim of this study is to review the diffusion-weighted MRI findings of glioblastomas, to investigate those with atypical characteristics and to emphasise the reasons responsible for the atypical features on diffusion-weighted MR images. 48 cases of histologically proven glioblastomas were included in this study. In addition to conventional sequences of routine tumour protocol, diffusion-weighted MRI with spin-echo echoplanar sequence was performed. The cystic-necrotic components of the lesions, according to the conventional sequences, were determined on the diffusion-weighted MR images and were classified as typical or atypical. The presence of high signal intensity was accepted as an atypical feature while low signal intensity was accepted as typical. The apparent diffusion coefficient (ADC) values of the cystic components were calculated. The statistical significance of the typical and atypical glioblastomas was evaluated with the students t-test. In six of the cases apparent high signal intensity in diffusion weighted MR images was interpreted. In three cases the high signal intensity occupied all of the cystic component and in the other three most of the cystic component. The ADC values of the lesions varied between 0.8661023 mm2 s21 and 1.3961023 mm2 s21 (mean value 1.06¡0.1761023 mm2 s21). In 42 of the lesions the cystic-necrotic component demonstrated low signal intensity and the ADC values varied between 1.5661023 mm2 s21 and 3.3261023 mm2 s21 (mean value 2.36¡0.46 6 1023 mm2 s21). The difference between ADC values of atypical and typical lesions was statistically significant (p,0.001). The vast majority of glioblastomas do not exhibit restricted diffusion in diffusion-weighted MRI, but some of them display homogeneous or heterogeneous high signal intensity and decrease of ADC values. Diffusion-weighted MRI alone is not helpful in the differentiation of malignant tumours from abscesses with low ADC values and similar conventional MRI findings.

Gliomas are the most common primary tumours of the central nervous system in adults [1]. The histology varies between low-grade (benign) and high-grade (malignant). The histopathological grading and differentiation from other lesions play a very significant role in planning therapy, evaluating prognosis and assessment of the response to therapy [2, 3]. Conventional MR images are sometimes not adequate to differentiate cystic glioblastomas from brain abscesses [3]. Diffusion-weighted MRI has been used in various pathologies such as ischaemia, infection and tumours [4]. In this method, strong magnetic gradients are applied and images from the movement of water in the biological tissues are obtained. The amount of diffusion is affected by the microstructure and microdynamic processes in the tissue, and the diffusion coefficients of the tissues can be measured quantitatively via the ‘‘apparent diffusion coefficient’’ (ADC) method [5]. There are many studies emphasising the value of diffusionweighted MRI in differentiating malignant cystic tumours from brain abscesses. Generally, brain abscesses with their dense and viscous structure exhibit high signal intensity and their ADC value is low. On the other hand, cystic brain tumours are serous in nature and they display low signal intensity on diffusion-weighted images with high ADC values [6–9]. Brain tumours which contain infected or haemorrhagic material can show similar diffusionweighted MRI signal properties as abscesses [10–12]. Though there is a limited number of reports about Received 16 March 2005 and in final form 20 April 2005, accepted 12 May 2005.

The British Journal of Radiology, November 2005

atypical diffusion-weighted MRI findings of cystic or necrotic glioblastomas in the literature. In this study our goal is to evaluate diffusion-weighted MRI findings of glioblastomas in a large series of patients, and to investigate the imaging characteristics of the lesions and the causes of atypical imaging features.

Materials and methods 48 patients with histologically proven glioblastomas between 2002 and 2005 were included in this study. They were aged between 26 years and 75 years (mean 51.4 years). The examinations were performed in two different 1.5 T super-conducting machines with gradient strengths of 23 mT m21 and 25 mT m21. T1 weighted images following contrast material administration in three planes were obtained in addition to T1 and T2 weighted images in axial plane. Diffusion-weighted MRI was performed with spin-echo echo-planar sequence. The gradients susceptible to diffusion were applied in three different directions at the same time (frequency, phase and slice selection) and mean trace values were obtained. b values of 0 and 1000 s mm22 were used (b value represents the duration and the amplitude of the gradients used in diffusion-weighted sequences). ADC values were calculated automatically with the software systems of the equipment. The diffusion-weighted MR images matching the cysticnecrotic component of the lesion in conventional sequences were evaluated. High signal intensity depicted in those areas was accepted as atypical, while low signal 989

B Hakyemez, C Erdogan, N Yildirim and M Parlak

Figure 1. Glioblastoma, left basal ganglia. T1 weighted image (a) low signal intensity of the cystic component and post-contrast T1 weighted image (b) peripheral enhancement after contrast material administration. Diffusion-weighted MR image (c) low signal intensity, typical lesion. Apparent diffusion coefficient (ADC) map (d) depicts high signal intensity of the lesion, ADC value is 2.8261023 mm2 s21.

intensity was accepted as typical. The signal intensity characteristics at diffusion-weighted MRI were named versus normal brain parenchyma. Three different regions of interest (ROIs) of 10–20 mm2 were placed inside cystic components of the lesions in order to calculate the mean ADC values. The statistical analysis was made using the student’s t-test and p,0.05 was accepted as statistically significant.

high signal intensity occupied almost all over the cystic component (Figure 2), while it was heterogeneous in the rest of the lesions. The ADC values were between 0.8661023 mm2 s21 and 1.3961023 mm2 s21 (mean value 1.0861023 mm2 s21). There was a statistically significant difference between glioblastomas with atypical and typical diffusion-weighted MRI features (Figure 3).

Discussion Results 42 of 48 patients with glioblastoma demonstrated typical findings (88%), while 6 of them presented with atypical diffusion-weighted MRI features (12%). In 42 cases, low signal intensity was depicted at the cystic component of the lesion in diffusion-weighted MR images (Figure 1). ADC values varied between 1.5661023 mm2 s21 and 3.326 1023 mm2 s21 (mean value 2.36¡0.4661023 mm2 s21). Six of the masses with atypical features demonstrated apparent high signal intensity. In three of those lesions the

In this study, the diffusion-weighted MR images of glioblastomas were classified as typical and atypical. Typical glioblastomas (88%) do not demonstrate low signal intensity and restricted diffusion, while high signal intensity and restriction of diffusion is depicted in glioblastomas with atypical imaging findings (18%). Brain tumours exhibit various imaging findings due to their internal tissue characteristics, and factors influencing the microscopic diffusion of water are responsible for the difference. Some cystic tumours like epidermoid tumours

Figure 2. Glioblastoma, right frontal lobe. T1 weighted image (a) apparent low signal intensity and post-contrast T1 weighted image (b) lesion with strong peripheral enhancement after contrast material administration. Diffusion-weighted MR image (c) apparent high signal intensity, atypical lesion. Low signal intensity is visualized on apparent diffusion coefficient (ADC) map (d) with ADC value of 1.13610–3 mm2 s21.

990


Glioblastoma multiforme

show remarkably high signal intensity. The mucinous matrix of such tumours restricts the diffusion of water molecules significantly and causes an increase in signal intensity by which means epidermoid tumours can easily be differentiated from arachnoid cysts by diffusionweighted MRI [12]. Solid tumours such as lymphoma, medulloblastoma, germinoma and meningioma may also exhibit high signal intensity [13]. This is due to the diminution of extracellular space because of packed cell groups in the lesion, scant cytoplasms and wide nuclei [7]. Tumours contain serous liquid of low viscosity with inflammatory cells when necrosis and liquefaction occurs. This content is generally hypointense in diffusion-weighted MR images and the ADC value is usually high. The restriction of the microscopic diffusion of water molecules is not remarkable. Abscesses, on the other hand, contain inflammatory cells, large protein molecules and bacteria which are responsible for the exudative content. The ADC values are remarkably low and they are hyperintense in diffusion-weighted MR images, and all those features depicted in diffusion-weighted MR images are therefore helpful in differentiating necrotic brain tumours and abscesses which might mimic each other [6, 12]. In many studies the ADC values of tumours are reported to be higher than those of abscesses [6–9]. Noguchi et al reported ADC values of 2.2–3.26 1023 mm2 s21 in glial tumours and 0.761023 mm2 s21 in abscesses [7]. Stadnik et al reported ADC values between 0.7861023 mm2 s21 and 1.7961023 mm2 s21 in glial tumours [12]. In our study, in 42 of 48 cases with the tumoural cavity of low signal intensity the ADC values were 2.36¡0.4661023 mm2 s21 in accordance with the findings reported in the literature. In the rest of the patients, the tumour cavity was hyperintense and the ADC values were 1.06¡0.1761023 mm2 s21, lower than the others. These findings suggest restriction of diffusion inside the cavity, which might be the result of various processes. Intratumoural haemorrhage, cytotoxic oedema at the early phase of cell death, thick sterile liquefaction or pyogenic infection especially can contribute to this process [14–19]. Hartmann et al described restricted diffusion and low ADC values in a metastatic

Figure 3. Apparent diffusion coefficient (ADC) value of the 48 glioblastoma multiforme. The difference between typical and atypical forms of glioblastomas was statistically significant (p,0.001). The British Journal of Radiology, November 2005

adenocarcinoma. It was shown that low ADC values are not specific to abscesses, and metastatic tumours of the brain could also show such characteristics [11]. Holtas et al reported similar findings in a patient with metastatic adenocarcinoma [14]. Parks et al reported low ADC values in two of seven metastatic lesions [19]. It is also shown that some of the lesions, like squamous cell carcinoma metastases and radiation necrosis, exhibit decrease of ADC values and restriction of diffusion inside the cavity [17]. In these studies; sterile liquefactive necrosis, intratumoural superinfection and intracavitary microhaemorrhage are accepted to be responsible for the high signal intensity in diffusion-weighted MR images of malignant necrotic masses. In the literature the atypical diffusionweighted MR findings in metastatic lesions are defined and no study in a wide series of patients with glioblastomas exists. As a result, most of the glioblastomas do not exhibit restriction of diffusion in diffusion-weighted MRI, although some of them may show homogeneous or heterogeneous high signal intensity with decrease in the ADC values in diffusion-weighted MRI. Diffusionweighted MRI alone is not helpful in the differentiation of malignant tumours from abscesses with low ADC values and similar conventional MRI findings.

References 1. Burnetti A, Alfano B, Soricelli A, et al. Functional characterization of brain tumors: an overview of the potential clinical value. Nucl Med Biol 1996;23:699–715. 2. Bagley LJ, Grossman RI, Judy KD, et al. Gliomas: correlation of magnetic susceptibility artifact with histologic grade. Radiology 1997;202:511–6. 3. Mamelak AN, Mampalam TJ, Obana WG, Rosenblum ML. Improved management of multiple brain abscesses: a combined surgical and medical approach. Neurosurgery 1995;36:76–85. 4. Castillo M, Mukherji KS. Diffusion-weighted imaging of intracranial lesions. Semin US, CT and MRI 2000;21:405–15. 5. Hajnal JV, Doran M, Hall AS, et al. MR imaging of anisotropically restricted diffusion of water in the nervous system: technical, anatomic, and pathologic considerations. J Comput Assist Tomogr 1991;15:1–18. 6. Ebisu T, Tanaka C, Umeda M, et al. Discrimination of brain abscess from necrotic or cystic tumors by diffusion-weighted echo planar imaging. Magn Reson Imaging 1996;14:1113–6. 7. Noguchi K, Watanabe N, Nagayoshi T, et al. Role of diffusion-weighted echo-planar MRI in distinguishing between abscess and tumor: a preliminary report. Neuroradiol 1999;41:171–4. 8. Desprenchins B, Stadnik T, Koerts G, et al. Use of diffusionweighting MR imaging in differential diagnosis between intracerebral necrotic tumors and cerebral abscess. AJNR Am J Neuroradiol 1999;20:1252–7. 9. Ping HL, Jih TH, Wei L, et al. Brain abscess and necrotic brain tumor: discrimination with proton MR spectroscopy and diffusion-weighted imaging. AJNR Am J Neuroradiol 2002;23:1369–77. 10. Chan JH, Tsui EY, Chau LF, et al. Discrimination of an infected brain tumor from a cerebral abscess by combined MR perfusion and diffusion imaging. Comput Med Imaging Graph 2002;26:19–23. 11. Hartmann M, Jansen O, Heiland S, Sommer C, Munkel K, Sartor K. Restricted diffusion within ring enhancement is not pathognomonic for brain abscess. AJNR Am J Neuroradiol 2001;22:1738–42.

991

B Hakyemez, C Erdogan, N Yildirim and M Parlak 12. Hakyemez B, Yildiz H, Ergin N, Uysal S, Parlak M. Flair and diffusion weighted MR imaging in differentiating epidermoid cysts from arachnoid cysts. Tani Girisim Radyol 2003;9:418–26. 13. Stadnik TW, Chaskis C, Michotte A, et al. Diffusionweighted MR imaging of intracerebral masses: comparison with conventional MR imaging and histologic findings. AJNR Am J Neuroradiol 2001;22:969–76. 14. Holtas S, Geijer B, Stromblad LG, Maly-Sundgren P, Burtscher IM. A ring-enhancing metastasis with central high signal on diffusion-weighted imaging and low apparent diffusion coefficients. Neuroradiol 2000;42:824–7. 15. Tung GA, Evangelista P, Rogg JM, Duncan JA. Diffusionweighted MR imaging of rim enhancing brain masses: is markedly decreased water diffusion specific for brain abscess? AJR Am J Roentgenol 2001;177:709–12.

992

16. Batra A, Tripathi RP. Atypical diffusion-weighted magnetic resonance findings in glioblastoma multiforme. Australas Radiol 2004;48:388–91. 17. Dorenbeck U, Butz B, Schlaier J, Bretschneider T, Schuierer G, Feuerbach S. Diffusion-weighted echo-planar MRI of the brain with calculated ADCs: a useful tool in the differential diagnosis of tumor necrosis from abscess? Neuroimaging 2003;13:330–8. 18. Mori H, Abe O, Aoki S, Masumoto T, Yoshikawa T, Kunimatsu A, et al. Hemorrhagic brain metastases with high signal intensity on diffusion-weighted MR images. A case report. Acta Radiol 2002;43:563–6. 19. Park SH, Chang KH, Song IC, Kim YJ, Kim SH, Han MH. Diffusion-weighted MRI in cystic or necrotic intracranial lesions. Neuroradiology 2000;42:716–21.
