This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright
Author's personal copy
Clinical Radiology 66 (2011) 768e777
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.elsevierhealth.com/journals/crad
Pictorial Review
Imaging of primary central nervous system lymphoma Y.Z. Tang*, T.C. Booth, P. Bhogal, A. Malhotra, T. Wilhelm Royal Free Hospital, London, UK
article in formation Article history: Received 18 January 2011 Received in revised form 1 March 2011 Accepted 8 March 2011
Primary central nervous system lymphoma (PCNSL) comprises 5% of all primary brain tumours. PCNSL demonstrates a variety of well-documented imaging findings, which can vary depending on immune status and histological type. Imaging features of PCNSL may overlap with other tumours and infection making definitive diagnosis challenging. In addition, several rare variants of PCNSL have been described, each with their own imaging characteristics. Advanced imaging techniques including 2-[18F]-fluoro-2-deoxy-D-glucose (18FDG) and 11 C positron-emission tomography (PET), 201Tl single-photon emission computed tomography (SPECT), 1H-magnetic resonance spectroscopy (MRS), and MR perfusion, have been used to aid differentiation of PCNSL from other tumours. Ultimately, no imaging method can definitively diagnose PCNSL, and histology is required. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Primary central nervous system lymphoma (PCNSL) causes approximately 5% of all primary brain tumours. PCNSL is defined as lymphoma in the central nervous system (CNS) without primary tumour elsewhere. It is less common than secondary CNS involvement by systemic lymphoma. PCNSL is a diffuse large B-cell lymphoma in 90% of cases and is usually high grade. Less common PCNSL histological types are Burkitt’s lymphoma and T-cell lymphoma.1
Background There has been a dramatic rise in PCNSL incidence in recent decades. Between 1973 to 1992, the incidence increased from 2.5 to 30 cases per 10 million population. The rise was independent of age and gender and has outpaced the minimal rise of systemic non-Hodgkin’s lymphoma and the * Guarantor and correspondent: Y.Z. Tang, Royal Free Hospital, Pond Street, London NW32QG, UK. Tel.: þ44 7967656116 (mobile). E-mail address:
[email protected] (Y.Z. Tang).
relatively static rates of gliomas.2 The acquired immunodeficiency syndrome (AIDS) epidemic contributed to this phenomenon but the rise was also observed in the immunocompetent population. Incidence of PCNSL has since stabilized in the developed world with a slight decrease since the mid 1990s observed in the United States.3 This reflects the decrease in the incidence of AIDS since the introduction of highly active anti-retroviral therapy (HAART).4 Unlike systemic lymphoma, PCNSL typically presents with neurological symptoms. In a study of 248 patients with lymphoma, patients presented with symptoms of raised intracranial pressure, focal deficits, seizures, ocular and neuropsychiatric symptoms.5 “B symptoms” such as fever, weight loss, and night sweats are uncommon and reported in only 8% of 466 patients in one study.6 This is modulated by immune status and is commoner in AIDS-related PCNSL patients.7 The CNS is devoid of endogenous lymphoid tissue and the aetiology of PCNSL remains unclear. PCNSL is associated with EpsteineBarr virus8 and cytomegalovirus infection.9 The male population is affected more and contributes to 71% of cases.3 The only established risk factor for PCNSL is immunodeficiency and PCNSL is the most common brain tumour in this population. AIDS accounts for the largest
0009-9260/$ e see front matter Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2011.03.006
Author's personal copy
Y.Z. Tang et al. / Clinical Radiology 66 (2011) 768e777
769
Imaging PCNSL has many documented imaging features, which may mimic other diseases or have atypical imaging features. There are also rare variants of PCNSL, which adds to the diagnostic difficulty. In one study, only 39% of cases were accurately diagnosed as PCNSL on imaging alone.13 Histology is, therefore, required for a definitive diagnosis. Steroids disrupt cellular morphology and can cause falsenegative results,12,14,15 and are, therefore, contraindicated before biopsy. A unique finding in PCNSL is the dramatic initial response of CNS lymphoma to steroids, which occurs after only 48 hours. Imaging demonstrates the marked reduction in both tumour size and peritumoural oedema very clearly; however, most patients relapse. The length of remission is variable and has been reported as up to 2.5 years.14,16
Imaging in immunocompetent patients
Figure 1 Enhanced T1-weighted MRI demonstrating multifocal, enhancing PCNSL. It has spread across the genu of the corpus callosum, involving both frontal lobes. A further PCNSL mass is seen in the left subinsular region.
group of immunocompromised patients with PCNSL and is an AIDS-defining diagnosis.10 Patients with AIDS have a 3600-fold increased risk than the baseline population,11 and it is recommended to exclude HIV in all patients with PCNSL.12 Immunocompromised patients typically have a higher histological grade, are younger (30e50 years) and have a shorter survival time than immunocompetent patients.7
Figure 2 Unenhanced CT head depicts a hyperdense, left frontal PCNSL. There is surrounding vasogenic oedema but no midline shift.
PCNSL is typically a solitary and supratentorial lesion. In a study of 100 immunocompetent patients with PCNSL, the most common location was the cerebral hemispheres seen in 38% of cases. This was followed by basal ganglia and thalamic lesions seen in 16%.17 A further characteristic site of invasion is the corpus callosum.18 Similar to glioblastome multiforme (GBM), PCNSL can spread across the corpus callosum, involve both frontal lobes and genu of the corpus callosum giving a typical butterfly pattern (Fig 1). These lesions tend to be bigger than other PCNSL lesions.17 Less commonly the cerebellum and brainstem may be infiltrated.13,17
Figure 3 Enhanced CT head corresponding to Fig 9 demonstrates homogeneous enhancement of the tumour.
Author's personal copy
770
Y.Z. Tang et al. / Clinical Radiology 66 (2011) 768e777
Figure 4 Gadolinium-enhanced T1-weighted MRI. There is multifocal, bilateral enhancing subependymal PCNSL in the occipital horns of the lateral ventricles.
The high nuclear to cytoplasmic ratio in PCNSL results in hyperdensity on computed tomography (CT), but it may also be isodense19 (Fig 2). PCNSL lacks calcification or intratumoural haemorrhage prior to chemotherapy.20 Peritumoural oedema is seen to a lesser extent than in other CNS malignant tumours and may also be absent.13,17 A negative CT examination does not exclude CNS lymphoma and there is a reported 13e38% false-negative rate.7,21 PCNSL typically enhances homogeneously; however, a complete lack of enhancement has been reported17 (Fig 3). Contrast-enhanced magnetic resonance imaging (MRI) is the optimal imaging technique.22 PCNSL is typically isohypointense on T1-weighted imaging. It is iso-hypointense to grey matter on T2-weighted imaging, differing from most other brain tumours, which are typically high signal on T2-weighted imaging; however, PCNSL can also demonstrate T2-weighted hyperintensity. Contrast enhancement
Figure 5 Enhanced T1-weighted MRI. There are numerous, small, ring-enhancing PCNSL lesions in the basal ganglia and white matter.
tends to be homogeneous. In immunocompetent patients, necrotic, thus ring-enhancing, lesions are rare. PCNSL tends to demonstrate ependymal spread.19 If at least one of the lesions does not contact the ependyma or meninges, PCNSL is doubtful.17,19 Diffuse enhancement of the ependyma is highly specific for PCNSL in the context of AIDS and this is best seen on enhanced T1-weighted imaging (Fig 4). Leptomeningeal spread is difficult to diagnose on MRI, which has a sensitivity of 20%. The only imaging manifestation might be hydrocephalus. Guidelines on PCNSL from the British Committee for Standards in Haematology therefore recommend that, unless contraindicated, lumbar puncture for cerebrospinal fluid (CSF) cytological examination should be routinely performed in PCNSL patients to detect leptomeningeal spread.12 Imaging features, if present, include faint, meningeal enhancement.24
Imaging in immunocompromised patients Imaging features in immunocompromised patients are more variable and share some features with the immunocompetent patients. Some features are more typical in the immunocompromised population. Here, multiple lesions are more common and seen in up to 60% of patients,25 a higher percentage than the 35e38% found in immunocompetent patients6,17 (Fig 5). These lesions also tend to be smaller than in the immunocompromised with a mean diameter of 1.6 cm25 compared with 2.4 cm in the immunocompetent.17 PCNSL in the immunocompromised demonstrates rapid growth. The tumour outstrips its blood
Figure 6 Enhanced T1-weighted MRI in a patient with AIDS demonstrating a left frontal PCNSL with irregular ring enhancement. There is surrounding hypointensity as well as mass effect with midline shift, features consistent with vasogenic oedema.
Author's personal copy
Y.Z. Tang et al. / Clinical Radiology 66 (2011) 768e777
771
Figure 7 Enhanced T1-weighted MRI in an HIV-negative patient. There is an enhancing dural lymphoma, which has also infiltrated the skull vault (arrow). There is extensive dural thickening (thin arrow).
supply and becomes necrotic centrally. This leads to ring enhancement after the administration of contrast medium, and is characteristic of PCNSL in the immunocompromised patient. The ring-like enhancement may be irregular and nodular (Fig 6). Necrosis may also cause other heterogeneous enhancement patterns. Additionally, intra-tumoural haemorrhage may be seen, which is rare in the immunocompetent patient.26
T-Cell lymphoma T-cell lymphoma comprises 5% of PCNSLs27 PCNSL and systemic T-cell lymphoma is more common in Asia with a reported T-cell PCNSL incidence of 9e14%.6,28 They have a comparable presentation, outcome, and prognosis to B-cell PCNSL. It differs from B-cell PCNSL in that there are more “B symptoms” at presentation.29 Typical imaging findings are unclear because of the paucity of systemic evidence. Nonetheless, many different imaging features have been reported. A review of 25 cases suggested that features were similar to those found in immunocompetent patients with B-cell PCNSL.29 Indeed, tumours are multiple (29%) or single, have a supratentorial predilection,30,31 and are homogeneous before and after contrast medium administration, which demonstrates avid enhancement. Small case series suggest that leptomeningeal spread, neurolymphomatosis (peripheral nervous system involvement), and a subcortical distribution are common manifestations of T-cell PCNSL.31 Given the variety of imaging findings diagnosis is reliant on histopathology.
Primary dural lymphoma Primary dural lymphoma is a rare form of PCNSL seen in only 2% in a study of 335 patients with PCNSL.32 It is
Figure 8 T2-weighted MRI. There are multifocal lesions, predominately in the deep and subcortical white matter, which are hyperintense on T2-weighted imaging.
typically a low-grade B-cell marginal-zone lymphoma with a female predominance. It arises from the dura and can often be misdiagnosed as a meningioma as they share several imaging features. Indeed, dural lymphoma may demonstrate hyperdensity, a dural tail, vasogenic oedema, hyperostosis, and bone erosion (Fig 7). Furthermore, they enhance diffusely following contrast medium administration. Half these extra-axial lesions are single and half are multiple. Additionally direct leptomeningeal spread occurs in 63% of cases.33
Figure 9 Sagittal T1-weighted unenhanced MRI corresponding to Fig 8 confirms the gyriform high signal in the right occipital cortex (arrow), in keeping with pseudolaminar necrosis. The hypointensity in the right frontal lobe is again seen.
Author's personal copy
772
Y.Z. Tang et al. / Clinical Radiology 66 (2011) 768e777
Figure 10 DWI MRI corresponding to Fig 8 displays restricted diffusion in the areas of signal abnormality. The ADC map showed low signal in these areas. This was found on histology to be a case of intravascular malignant lymphomatosis. The patient’s immune status was not established.
Intravascular malignant lymphomatosis Intravascular malignant lymphomatosis (IML), also known as malignant angioendotheliomatosis, is a rare form of PCNSL. It is an aggressive subtype of diffuse large B-cell lymphoma affecting the CNS or skin. In a study of 104 cases of PCNSL only two cases were IML,1 and less than 50 cases have been reported. There is massive intravascular growth of lymphoid cells distending the affected small and intermediate size vessels. Diagnosis is difficult due to the nonspecific clinical and imaging features and often occurs only at autopsy. Histology in the majority of cases demonstrates widely distributed, multiple areas of recent or resolving infarcts. Sometimes, there is haemorrhagic transformation of infarcts.34 Reported clinical presentations include nonfocal neurological deficits, seizures, and change in mental state.34,35 Focal neurological deficits have also been reported.36 Various imaging findings have been reported but are non-specific. The most common feature is multifocal abnormalities in keeping with its growth pattern.35 On CT, low density in the white matter has been described.34 In a review of 11 cases, the most common MRI finding is abnormal hyperintensity on T2-weighted sequences in the deep white matter (Fig 8). This appears to correlate with oedema and gliosis pathologically. Infarct-like lesions were seen in 36% of cases, and less commonly
Figure 11 Enhanced T1-weighted MRI corresponding to Fig 8 at a more inferior level. There are multiple hypointensities in the right cerebral hemisphere. There is patchy enhancement within the right frontal lobe (black arrow). There is gyriform high signal in the right occipital cortex in keeping with pseudolaminar necrosis (white arrow). (Case contributed by Dr Indran Davagnanam of National Hospital for Neurology and Neurosurgery London UK.)
Figure 12 A case of lymphomatosis cerebri in an immunocompetent patient. Axial FLAIR MRI demonstrates diffuse hyperintensity in the white matter of both hemispheres, more severe on the left.
Author's personal copy
Y.Z. Tang et al. / Clinical Radiology 66 (2011) 768e777
773
Lymphomatosis cerebri A further rare presentation of PCNSL is lymphomatosis cerebri (LC). Here, there is diffuse lymphomatous parenchymal infiltration with no discernable tumour mass and preservation of cerebral architecture.38 It is analogous to gliomatosis cerebri and thus coined lymphomatosis cerebri. Clinically, it presents as a rapidly progressive dementia with personality change. MRI reveals widespread T2-weighted signal hyperintensities in the white matter thought to be due to infiltration of lymphomatous cells39,40 (Fig 12). The diffuse white matter lesions do not enhance and have been misdiagnosed as multiple sclerosis, encephalomyelitis,39 Binswanger’s disease, and CreutzfeldeJacob disease40 (Fig 13).
Additional imaging techniques
Figure 13 Post-gadolinium T1-weighted axial image corresponding to Fig 12 shows a distinct lack of contrast enhancement. Biopsy specimen showed a large B-cell lymphoma. (Case contributed by Dr Ryuichi Kanai of Shizuoka City Shimizu Hospital Japan.)
enhancing parenchymal mass lesions35,37 (Figs 9 and 10). Contrast enhancement has been described as gyriform, ring-like, or homogeneous37 (Fig 11). Meningeal enhancement is documented and thought to be due to direct infiltration of meningeal blood vessels by the tumour.35
Imaging techniques other than conventional CT and MRI are increasingly being used to aid diagnosis of PCNSL, particularly in cases with atypical imaging features. However, histology is still required for the final diagnosis as no imaging finding is definitively diagnostic. Differentiation from gliomas is essential as surgery is ineffective in the treatment of PCNSL.
Nuclear medicine The role of positron-emitted tomography (PET) in PCNSL is not clearly defined. PCNSL has a very high cellular density and increased glycolytic metabolism and shows avid uptake of 2-[18F]-fluoro-2-deoxy-D-glucose (18FDG) with PET (Fig 14). However, normal high uptake of FDG in the basal
Figure 14 (a) Unenhanced CT head in an immunocompetent patient. PCNSL presenting as an ill-defined hyperdensity is seen in the region of the right lateral ventricle (arrow). It appears to extend into the body of the right lateral ventricle. Mass effect is exerted causing midline shift to the left. The tip of a ventricular drain is also seen in this image. (b) PET-CT fusion image from a FDG-PET study. The PCNSL lesion shows avid FDG uptake. No further pathological FDG uptake was demonstrated in the remainder of the whole-body FDG-PET study.
Author's personal copy
774
Y.Z. Tang et al. / Clinical Radiology 66 (2011) 768e777
Figure 15 (a) Enhanced T1-weighted MRI. A homogeneously enhancing, periventricular lymphoma is seen abutting the left lateral ventricle. (b) Corresponding thallium-201 SPECT. There is increased uptake of thallium-201 within the lymphomatous mass.
ganglia, thalamus, and grey matter can mask underlying PCNSL. In cases with atypical imaging findings, such as no or disseminated lesions on MRI, FDG-PET has not been shown to aid localization or diagnosis.41 Lymphoma cells are dependent on the external supply of methionine and show increased 11C methionine uptake with PET. Background methionine is low in the brain parenchyma
and PCNSL is better visualized than in FDG-PET. This is useful for monitoring the therapeutic effect of irradiation, demonstrating missed residual tumour not detected on CT and MRI.42 Similarly, FDG-PET has been shown to detect therapy responses at an earlier stage than MRI after chemotherapy.43 FDG-PET has also proven useful in detecting systemic disease. In one study, 7% of patients were found to have systemic NHL on FDG-PET but corresponding CT and bone marrow biopsies were negative.44 FDG-PET is not included in
Figure 16 DWI MRI corresponding to Fig 9. There is restricted diffusion within the PCNSL mass as demonstrated by its hyperintensity.
Figure 17 ADC map corresponding to Fig 9 demonstrates hypointensity within the mass. This confirms restricted diffusion within the PCNSL.
Author's personal copy
Y.Z. Tang et al. / Clinical Radiology 66 (2011) 768e777
775
Figure 18 (a) Enhanced T1-weighted MRI demonstrates a homogeneously enhancing PCNSL in the cerebellum. (b) On visual evaluation of the cerebral blood volume (CBV) map, there is increased rCBV within the mass particularly at its periphery. The rCBV helps to distinguish it from GBM, which would have greater rCBV. Conversely, toxoplasmosis would have reduced rCBV. (Case contributed by Dr Teresa Cabadag of Hospital De Navarra Spain.)
staging PCNSL in the current guidelines from the British Committee for Standards in Haematology.12 In the immunocompromised patient PET is able to help differentiate PCNSL from toxoplasmosis, the two most common cerebral mass lesions in this population. Distinguishing the two diseases with cross-sectional imaging is problematic as enhancement patterns and multiplicity overlap.23 Here, nuclear medicine plays a key role as decreased FDG is seen in toxoplasmosis in contrast to high uptake in PCNSL. Previously, patients were given empirical anti-toxoplasmosis treatment for 2 weeks and if lesions resolved it was presumed to be toxoplasmosis. This is no longer performed as the delay significantly shortens survival in PCNSL.45 201Tl 30 m brain single-photon emission CT (SPECT) can also differentiate the two diseases. Toxoplasmosis demonstrates no 201TI uptake, whereas PCNSL displays intense uptake (Fig 15). The sensitivity and specificity of 201TI SPECT in detecting cerebral lymphoma in AIDS patients was demonstrated as 100% and 93%, respectively.46
ADC. In a cohort of 18 PCNSL patients, all cases with 25% percentile ADC values greater than 690 106 mm2/s demonstrated a complete response to methotrexate-based chemotherapy. Conversely, most patients with lower 25% percentile ADC values were more likely to have progressive disease.50 Studies have also been performed to investigate perfusion using dynamic susceptibility contrast (DSC) magnetic resonance in the diagnosis of PCNSL. This should be measured before steroid administration at it induces a reduction in bloodetumour barrier permeability and regional cerebral blood volume.51 The maximum relative cerebral blood volume ratio (rCBVmax) in PCNSL(1.3e2.3) pre-treatment is lower than in GBM (4.9e6.3).52,53 It also demonstrates less perfusion than anaplastic astrocytomas
Advanced imaging techniques Water diffusion is often restricted in lymphoma as it is hypercellular and made up of large lymphoid cells. PCNSL is therefore hyperintense on diffusion-weighted imaging (DWI) and hypointense on apparent diffusion coefficient (ADC) maps (low ADC values) (Figs 16 and 17). Ninety percent of pre-treatment patients in one study demonstrated restricted diffusion, but PCNSL may also demonstrate unrestricted diffusion. After treatment, restricted diffusion is more variable.47 The PCNSL ADC values are lower than in astrocytomas, GBM, and gliomatosis cerebri. Reported ADC values in PCNSL are 0.7e0.9 103 mm2/s.48,49 There is no established imaging biomarker predictive of prognosis yet, but one study has shown the possible use of
Figure 19 Single voxel, long TE 1H-MRS corresponding to Fig 9. There is a raised choline peak and low NAA peak typical of PCNSL.
Author's personal copy
776
Y.Z. Tang et al. / Clinical Radiology 66 (2011) 768e777
Acknowledgements The authors thank Dr Ryuichi Kanai of Shizuoka City Shimizu Hospital Japan, Dr Indran Davagnanam of National Hospital for Neurology and Neurosurgery London UK, and Dr Teresa Cabadag of Hospital De Navarra Spain for the provision of some of the images.
References
Figure 20 The single voxel, short TE 1H-MRS corresponding to Fig 9 demonstrates an exaggerated lipid peak in a solid mass. Lipid peaks are typically demonstrated in necrotic lesions, but in solid PCNSL lesions the peak is due to the increased macrophage content. (Case contributed by Dr Teresa Cabadag of Hospital De Navarra Spain.)
and metastases.52 PCNSL lacks tumour neovascularization accounting for its lower rCBVmax (Fig 18). 1 H-magnetic resonance spectroscopy (MRS) provides information on metabolic change in vivo. The most specific finding for PCNSL on MRS is an increase in lipid resonance. This is typically a signature of cell death; however, a lipiddominated spectrum is found in PCNSL that is not macroscopically necrotic.54,55 This appears to be due to numerous macrophages and the increased turnover of membrane components in transformed lymphoid cells. Harting et al. found significantly higher lipid peaks in solid PCNSL than solid low and high-grade astrocytomas. PCNSL cases also demonstrates raised choline (Cho) resonances relative to creatine (Cr) and N-acetyl aspartate (NAA), which are non-specific and are also demonstrated in astrocytomas and GBM (Figs 19 and 20).
Conclusion Diffuse large B-cell PCNSL may present with a wide range of established imaging findings. This makes diagnosis with imaging alone challenging. Additionally, the radiologist needs to be aware of the rare PCNSL variants that present with their own unique imaging features. Although MRI is the imaging of choice, it cannot enable the definitive diagnosis of PCNSL without histology. Much interest has been placed on advanced imaging techniques, including MR perfusion, MRS, and nuclear medicine, to aid diagnosis. Their findings have been shown to complement conventional MRI findings, supporting a diagnosis of PCNSL. These techniques have also demonstrated the potential to contribute to assessment of prognosis, and monitoring response to therapy. With further research, imaging may be able to accurately diagnose PCNSL and have an established role in prognosis and monitoring of patients.
1. Miller DC, Hochberg FH, Harris NL, et al. Pathology with clinical correlation of primary central nervous system non-Hodgkin’s lymphoma. Cancer 1994;74:1383e97. 2. Corn BW, Marcus SM, Topham A, et al. Will primary central nervous system lymphoma be the most frequent brain tumour diagnosed in the year 2000? Cancer 1997;15(79):2409e13. 3. Kadan-lottick NS, Skluzacek MC, Gurney JG. Decreasing incidence rates of primary central nervous system lymphoma. Cancer 2002;95:193e202. 4. Guiguet M, Boue F, Cadranel J, et al. Effect of immunodeficiency, HIV viral load, and antiretroviral therapy on the risk of individual malignancies (FHDHANRS CO4): a prospective cohort study. Lancet Oncol 2009;10:1152e9. 5. Bataille B, Delwail V, Menet E, et al. Primary intracerebral malignant lymphoma: report of 248 cases. J Neurosurg 2000;92:261e6. 6. Hayabuchi N, Shibamoto Y, Onizuka Y. Primary central nervous system lymphoma in Japan: a nationwide survey. Int J Radiat Oncol Biol Phys 1999;44:265e72. 7. Remick SC, Diamond C, Migliozzi JA, et al. Primary central nervous lymphoma in patients with and without the acquired immune deficiency syndrome. Medicine 1990;69:345e60. 8. Bashir R, Luka J, Cheloha K, et al. Expression of EpsteineBarr virus proteins in primary CNS lymphoma in AIDS patients. Neurology 1993;43:2358e62. 9. Morgello S, Petito CK, Mouradian JA. Central nervous lymphoma in the acquired immunodeficiency syndrome. Clin Neuropathol 1990;9:205e15. 10. Gill PS, Levine AM, Meyer PR, et al. Primary central nervous system lymphoma in homosexual men. Clinical, immunologic, and pathologic features. Am J Med 1985;78:742e8. 11. Cote TR, Manns A, Hardy CR, et al. Epidemiology of brain lymphoma among people with or without acquired immunodeficiency syndrome. AIDS/Cancer Study group. J Natl Cancer Inst 1996;88:675e9. 12. Guidelines on the diagnosis and management of adult patients with primary CNS lymphoma (PCNSL) and primary intra-ocular lymphoma (PIOL). British Committee for Standards in Haematology; 2007. 13. Zhang D, Hu LB, Henning T. MRI findings of primary CNS lymphoma in 26 immunocompetent patients. Korean J Radiol 2010;11:269e77. 14. De Angelis LM, Yahalom J, Heinemann MH, et al. Primary CNS lymphoma: combined treatment with chemotherapy and radiotherapy. Neurology 1990;40:80e6. 15. Damek D. Primary central nervous system lymphoma. Current treatment options in Neurology 2003;5:213e22. 16. Bent MR, Vanneste JAL, Ansink BJJ. Prolonged remission of primary central nervous system lymphoma after discontinuation of steroid therapy. J Neurooncol 1992;43:237e9. 17. Kuker W, Nagele T, Korfel A, et al. Primary central nervous system lymphomas (PCNSL): MRI features at presentation in 100 patients. J Neurooncol 2005;72:169e77. 18. Erdag N, Bhorade RM, Alberico R, et al. Primary lymphoma of the central nervous system: typical and atypical CT and MRI imaging appearances. Am J Roentgenol 2001;176:1319e26. 19. Jack Jr CR, O’neill BP, Banks PM, et al. Central nervous system lymphoma: histologic types and CT appearance. Radiology 1988;167:211e5. 20. Zimmerman RA. Central nervous system lymphoma. Radiol Clin North Am 1990;28:697e721. 21. Haldorsen IS, Krakenes J, Krossnes BK, et al. CT and MRI imaging features of primary central nervous system lymphoma in Norway. 1989e2003. AJNR Am J Neuroradiol 2009;30:744e51. 22. Batchelor T, Loeffler JS. Primary CNS lymphoma. J Clin Oncol 2006;24:1281e8. 23. Dina TS. Primary central nervous system lymphoma versus toxoplasmosis in AIDS. Radiology 1991;179:823e8.
Author's personal copy
Y.Z. Tang et al. / Clinical Radiology 66 (2011) 768e777 24. Slone HW, Blake JJ, Shah R, et al. CT and MRI findings of intracranial lymphoma. AJR Am J Roentgenol 2005;184:1679e85. 25. Johnson BA, Fram EK, Johnson PC, et al. The variable MR appearance of primary lymphoma of the central nervous system: comparison with histopathologic features. AJNR Am J Neuroradiol 1997;18:563e72. 26. Thurnher MM, Rieger A, Kleibl-Popov C, et al. Primary central nervous system lymphoma in AIDS: a wider spectrum of CT and MRI findings. Neuroradiology 2001;43:29e35. 27. Levin N, Soffer D, Grissaru S, et al. Primary T-cell CNS lymphoma presenting with leptomeningeal spread and neurolymphomatosis. J Neurooncol 2008;90:77e83. 28. Hayakawa T, Takakura K, Abe H, et al. Primary central nervous co-operative study by CNS Lymphoma Study Group in Japan. J Neurooncol 1994;19:197e215. 29. Shenkier TN, Blay JY, O’neill BP, et al. Primary CNS lymphoma of T-cell origin: a descriptive analysis from the international primary CNS lymphoma collaborative group. J Clin Oncol 2005;23:2233e9. 30. Liu D, Schelper R, Carter DA, et al. Primary central nervous system cytotoxic/suppressor T-cell lymphoma. Am J Surg Path 2003;27:682e8. 31. Kim EY, Kim SS. Magnetic resonance findings of primary central nervous system T-cell lymphoma in immunocompetent patients. Acta Radiol 2005;46:187e92. 32. Iwamoto FM, DeAngelis LM, Abrey LE. Primary dural lymphomas: a clinicopathologic study of treatment and outcome in eight patients. Neurology 2006;66:1763e5. 33. Iwamoto FM, Abrey LE. Primary dural lymphoma: a review. Neurosurg Focus 2006 Nov 15;21(5):E5. 34. Martin-Duverneuil N, Mokhtari K, Behin A, et al. Intravascular malignant lymphomatosis. Neuroradiology 2002;44:749e54. 35. Williams RL, Meltzer CC, Smirniotopoulos JG, et al. Cerebral MR imaging in intravascular lymphomatosis. AJNR Am J Roentgenol 1998;19:427e31. 36. Iijima M, Fujita A, Uchigata M, et al. Change of brain MRI findings in a patient with intravascular malignant lymphomatosis. Euro J Neurol 2007;15:e4e5. 37. Imai H, Kajimoto K, Taniwaki M, et al. Intravascular large B-cell lymphoma presenting with mass lesions in teh central nervous sytem: a report of five cases. Pathol Int 2004;54:231e6. 38. Bakshi R, Mazziotta JC, Mischel PS, et al. Lymphomatosis cerebri presenting as a rapidly progressive dementia: clinical, neuroimaging and pathologic findings. Dement Geriatr Cogn Disord 1990;10:152e7. 39. Kanai R, Shibuya M, Hata T, et al. A case of “lymphomatosis cerebri” diagnosed in an early phase and treated by whole brain radiation: case report and literature review. J Neurooncol 2008;86:83e8. 40. Vital A, Sibon I. A 64-year-old woman with progressive dementia and leukoencephalopathy. Brain Pathol 2007;17:117e8. 121.
777
41. Kawai N, Okubo S, Miyake K, et al. Use of PET in the diagnosis of primary CNS lymphoma in patients with atypical MR findings. Ann Nucl Med 2010;24:335e43. 42. Ogawa T, Kanno I, Hatazawa J, et al. Methionine PET for follow-up of radiation therapy of primary lymphoma of the brain. RadioGraphics 1994;14:101e10. 43. Palmedo H, Urbach H, Bender H, et al. FDG-PET in immunocompetent patients with primary central nervous system lymphoma: correlation with MRI and clinical follow-up. Eur J Nucl Med Mol Imaging 2006;33: 164e8. 44. Mohile N, Deangelis LM, Abrey LE. The utility of body FDG PET in staging primary central nervous system lymphoma. Neuro Oncol 2008;10:223e8. 45. Rosenblum ML, Levy RM, Bredesen DE, et al. Primary central nervous system lymphomas in patients with AIDS. Ann Neurol 1988;23(Suppl.): S13e6. 46. Kessler LS, Ruiz AM, Post MJD, et al. Thallium-201 brain SPECT of lymphoma in AIDS patients: pitfalls and technique optimization. AJNR Am J Roentgenol 1998;19:1105e9. 47. Zacharia TT, Law M, Naidich TP, et al. Central nervous system lymphoma characterization by diffusion-weighted imaging and MR spectroscopy. J Neuroimaging 2008;18:411e7. 48. Horger M, Fenchel M, Nagele T, et al. Water diffusivitiy: comparison of primary CNS lymphoma and astrocytic tumour infiltrating the corpus callosum. AJR Am J Roentgenol 2009;193:1384e7. 49. Guo AC, Cummings TJ, Dash RC, et al. Lymphomas and high-grade astrocytomas: comparison of water diffusibility and histologic characteristics. Radiology 2002;224:177e83. 50. Barajas Jr RF, Rubenstein JL, Chang JS, et al. Diffusion-weighted MR imaging derived apparent diffusion coefficient is predictive of clinical outcome in primary central nervous system lymphoma. AJNR Am J Neuroradiol 2010;31:60. 51. Ostergaard L, Hochberg FH, Rabinov JD, et al. Early changes measured by magnetic resonance imaging in cerebral blood flow, blood volume, and bloodebrain barrier permeability following dexamethasone treatment in patients with brain tumors. J Neurosurg 1999;90:300e5. 52. Calli C, Kitis O, Yunten N, et al. Pefusion and diffusion in MR imaging in enhancing malignant cerebral tumours. Eur J Radiol 2006;58:394e403. 53. Hartmann M, Heilanda S, Harting I, et al. Distinguishing of primary cerebral lymphoma from high-grade glioma with perfusion-weighted magnetic resonance imaging. Neurosci Lett 2003;338:119e22. 54. Tallibert S, Guillevin R, Menuel C, et al. Brain lymphoma: usefulness of the magnetic resonance spectroscopy. J Neurooncol 2008;86:224e9. 55. Harting I, Hartmann M, Jost G, et al. Differentiating primary central nervous system lymphoma from glioma in humans using localised proton magnetic resonance spectroscopy. Neurosci Lett 2003;342:163e6.
