Tumefactive lesions are an uncommon manifestation of demyelinating disease and can pose a diagnostic challenge in patients without a pre-existing diagnosis of multiple sclerosis. Choosing when to biopsy a tumefactive lesion to exclude alternative pathology can be difficult. Other questions include how best to treat an acute attack as well as the optimal timing of therapy to prevent relapse. This article aims to review the available literature for tumefactive demyelination and to propose an approach to diagnosis and management. We argue that disease modifying therapy should be considered for acute tumefactive demyelinating lesions only once criteria of dissemination in time and space are fulfilled and the diagnosis of multiple sclerosis is confirmed.
- Multiple Sclerosis
Statistics from Altmetric.com
The advent of MRI has revolutionised the diagnosis and management of multiple sclerosis (MS). Typical demyelinating lesions appear as small, often ovoid, T2 and fluid attenuated inversion recovery hyperintensities in the brain and/or spinal cord. Acutely, these lesions may enhance with gadolinium contrast whereas more chronic lesions are non-enhancing and may have the appearance of hypointense ‘black holes’ on T1 sequences.1 In the brain, the distribution of the lesions is helpful in distinguishing from other causes including leukoariosis, lymphoma, granuloma, vasculitis and infection. Locations that suggest demyelination include the corpus callosum, periventricular and deep white matter, juxtacortical regions and the infratentorial compartment.
Tumefactive demyelinating (TD) lesions are larger than those seen in typical MS (ie, >2 cm) and may occur in patients either with or without established MS. Lesion size, accompanying oedema and mass effect mean that patients may present clinically with symptoms and signs that are atypical for MS and have imaging findings which may lead to the lesions being mistaken for primary tumours or other space occupying lesions such as abscess or other infection, metastasis or infarct. The detection of TD lesions can therefore lead to diagnostic uncertainty—particularly if they occur in a patient without a known diagnosis of MS. For this reason they are often biopsied to help establish a diagnosis, although there are radiological features such as open ring enhancement that can be helpful in distinguishing them from neoplastic or other lesions.
The aim of this review is to revise the clinical presentation, diagnosis, prognosis and management of TD lesions and to offer a possible approach to diagnosis and treatment.
The prevalence of tumefactive demyelination has not been formally evaluated but it is estimated to be approximately 1–2 per 1000 cases of MS.2 There is no clear gender predilection for tumefactive lesions and patients in their 20s and 30s are most likely to be affected although paediatric and older patients have been described.3 The clinical presentation is often different from that seen in typical MS and tends to resemble that seen with more sinister space occupying lesions. New onset cortical signs are unusual in a conventional MS relapse but are much more common in TD presumably reflecting the larger size of the lesion and the effect of surrounding oedema in the brain. Depending on the anatomical site, clinical features that may suggest a tumefactive lesion include decreased level of consciousness, seizures, visual field deficits, cognitive dysfunction including dysphasia, hemiparesis and hemisensory disturbance.3 ,4 Headache and vomiting appear to be more common in children.5
TD lesions may be defined in radiological terms as pseudotumoural demyelinating lesions greater than 2 cm. Although there are no pathognomonic imaging signs to indicate a TD lesion there are characteristics which may be helpful in favouring TD over a neoplasm or abscess. Most commonly TD lesions are well circumscribed supratentorial lesions with a predilection for the frontal and parietal lobes3 ,6 (figure 1). Butterfly lesions involving the corpus callosum as well as basal ganglia, infratentorial and spinal cord lesions can occur. Mass effect and perilesional oedema in TD lesions are usually less substantial than that seen with malignancy but increases with larger lesions and those of a more acute nature, that is, <3 weeks.3 ,5 ,6 The majority of TD lesions range in size from 2 to 6 cm but lesions of up to 12 cm have been reported.3
The great majority of TD lesions enhance with gadolinium contrast and almost any pattern of enhancement can be seen (eg, homogeneous, heterogeneous, nodular, punctate, ring). Most pathologically proven tumefactive lesions have a closed ring appearance.6 If present, however, open ring enhancement with the incomplete portion of the ring on the grey matter side of the lesion is widely held to be an important diagnostic clue to a tumefactive lesion7 (figure 1). The ring enhancement is thought to represent the advancing area of active inflammation away from a central and more chronic non-enhancing core.8 Particularly useful is the relatively recent observation that a non-contrast CT brain in addition to MRI imaging can improve diagnostic accuracy of TD lesions compared with MRI alone. Specifically, CT hypoattenuation of MRI enhanced regions appears to be more predictive of TD (93%) than tumour9 (4%).
Other features typical of a demyelinating aetiology include a T2 hypointense rim, peripheral restriction on diffusion weighted imaging and venular enhancement. Unusually for demyelination there may be cortical as well as subcortical involvement.3 ,10 Magnetic resonance diffusion imaging shows that TD lesions have mildly increased diffusion coefficients which may help distinguish them from abscesses in which diffusion is reduced but is not helpful in making the distinction from neoplasm.11
The role of magnetic resonance spectroscopy (MRS) in distinguishing tumefactive lesions is yet to be clearly defined. TD lesions may cause an increased choline to N-acetyl-aspartate ratio but this is also a common finding in neoplastic lesions.11 Increased glutamate/glutamine peaks appear to favour TD but this has not been extensively studied.12 Serial MRS may be more useful than one-off imaging because as TD lesions age their MRS findings change whereas those associated with neoplasm tend to remain stable.13
Fluoro-deoxyglucose positron emission tomography (FDG-PET) scanning may be a useful adjunct to MRI and CT in the investigation of TD lesions. While acute demyelinating lesions, including TD lesions, may demonstrate increased hypermetabolism on FDG-PET, in general the degree of hypermetabolism appears to be less than that seen in neoplastic lesions.14 In five patients with pathologically verified TD, FDG-PET hypermetabolism was only marginally increased above that of normal cortex.15 In this study using methionine PET, a modality for evaluating amino acid metabolism in a tissue, hypermetabolism was significantly less than that seen for anaplastic astrocytomas or glioblastoma multiforme (GBM) prompting the authors to suggest that the combination of methionine PET and MRS warrants further research in the assessment of inflammatory demyelinating lesions.15 An important clinical point is that increased FDG uptake seen in primary central nervous system lymphoma (PCNSL) is generally not changed significantly following corticosteroids.16
When to biopsy?
Not all patients presenting with tumefactive lesions require stereotactic brain biopsy. MRI is often sufficient to make a diagnosis of tumefactive demyelination and the results of biopsy can be inconclusive or misleading. Additional white matter lesions are present in 85% of patients with TD and these may be reassuring if their distribution is typical for MS and if the tumefactive lesions have radiological characteristics in keeping with a demyelinating aetiology.6
Of course, it is important that other pathologies such as vasculitis, granuloma, infection and malignancy are excluded as far as possible. A serum screen incorporating antinuclear antibody, antineutrophil cytoplasmic antibody, extractable nuclear antigens, rheumatoid factor, complement, lupus anticoagulant, anticardiolipin antibodies, serum paraproteins, ACE, erythrocyte sedimentation rate, C reactive protein, aquaporin 4 antibodies and infection screen including HIV, hepatitis B, C, toxoplasma, cytomegalovirus and treponemal and borrelia serology is a reasonable starting point. Depending on the clinical circumstances, serum tumour markers, JC polyoma virus and an interferon-γ release assay or Mantoux could be considered. We would also recommend an initial screening chest x-ray with contrast-enhanced CT of the chest, abdomen and pelvis which can be performed at the same time as PET (CT-PET).
The cerebrospinal fluid (CSF) may also be a useful adjunct in identifying TD lesions over alternative pathology if a lumbar puncture is safe to perform. In 33 patients with a pre-established diagnosis of MS who developed TD, 90% had positive unmatched oligoclonal bands (OCBs) in the CSF compared with 52% who presented with a tumefactive lesion as part of their first clinical event.6 As with a conventional clinically isolated syndrome (CIS), or as in the investigation of probable MS, the presence of unmatched OCBs in the CSF could provide an additional level of diagnostic reassurance for the clinician in favour of demyelination rather than neoplasm, although it may not immediately alter management in that radiological follow-up of a TD lesion is still necessary. A finding of positive unmatched OCBs in the CSF may also provide prognostic information in terms of later conversion to MS. In some patients, for example, those who are HIV positive or who are otherwise immunocompromised, CSF is important for excluding cytological evidence of neoplasm or evidence of infection.
In practice, the patients who usually proceed to biopsy are those without a pre-existing diagnosis of MS, with inconclusive or suspicious imaging including PET, negative OCBs and/or those in whom a diagnosis of MS would be unusual, such as in older or very young patients. In those patients who appear to have a picture consistent with a typical TD lesion and in whom no other cause has been identified, we would advocate treating as presumed TD with corticosteroids (see Treatment section) and monitoring clinically and radiologically for response.
Unfortunately, there is no current evidence to guide the timing of further imaging. TD lesions may take more than 12 weeks to show significant resolution postcorticosteroids and 2% of demyelinating lesions will show persistent gadolinium enhancement after 6 months.8 Neoplastic lesions such as GBM often respond initially to corticosteroids as associated perilesional oedema resolves but the improvement is not sustained. Over the ensuing weeks, the oedema reaccumulates and patients develop rebound neurological deficits. PCNSL is more difficult to exclude based on corticosteroid response as there may be a more sustained clinical and radiological improvement with PCNSL typically re-declaring itself within 6–12 months.17–19 Longer term remission of GBM or PCNSL following corticosteroids is rare but delays of up to 5 years have been described in PCNSL.19–21 Complicating the picture is that so-called ‘sentinel’ demyelinating lesions have been described several months prior to the onset of PCNSL in some patients.17 As prior corticosteroids appear to reduce the success of a diagnostic biopsy in PCNSL we would suggest withholding corticosteroids for as long as is feasible until alternative diagnoses such as malignancy are excluded (eg, with FDG-PET, see Radiology section) or biopsy performed.
If a patient with a typical TD lesion responds well to corticosteroids we propose that it would be reasonable to consider reimaging 6–8 weeks later. If the patient is improving (or has recovered) and there is a good radiological response then perhaps further surveillance imaging could be obtained in 3 months’ time unless new or recurrent symptoms developed. If the lesion at 6–8 weeks now looked atypical for TD then it would be reasonable to proceed to biopsy bearing in mind that demyelination and GBM can coexist in a single patient and within the same lesion.22 If the patient is improving (or has recovered) but there is radiological progression of the lesion and the lesion continues to have appearances typical of TD then we would suggest a further MRI scan in another 6–8 weeks’ time to guide further decision making or sooner if the patient's condition worsened. These recommendations are based on an initial treatment regimen with intravenous methylprednisolone (IVMP) 1 g daily for 3 days. They may not apply as well to patients who are given tapering courses of oral corticosteroids post-IVMP.
Immunopathogenesis and pathology
The pathology of tumefactive demyelination has been well described.3 Importantly, there are no histological features that distinguish isolated TD from the demyelination seen in more typical MS lesions. Active lesions consist of areas of demyelination with hypercellularity and reactive astrocytes which may contain multiple nuclei (Creutzfeld cells) closely intermingled with myelin-containing foamy macrophages. There is relative sparing of axons with perivascular and parenchymal lymphocytic infiltrates. Features which may mimic tumour include areas of necrosis with nuclear atypia, astrocytic pleiomorphism and mitotic figures.3
The immunopathogenesis of TD remains unknown. A small number of early cases of TD were reported in the context of vaccination which prompted the suggestion that TD may represent an intermediate phenotype between MS and acute disseminated encephalomyelitis (ADEM) but subsequent studies have not noted an association.3 ,6 ,23 ,24 The close relationship between TD lesions and MS suggest a common aetiology. The frequent presence of positive OCBs in the CSF and anecdotal reports of response of TD lesions to plasma exchange (PEX) and Rituximab (see Treatment section) seem to infer some role for antibody and B cell mediated immunological mechanisms. Developmentally immature myelin basic protein has been implicated in the development of the aggressive Marburg variant of MS that can be associated with tumefactive lesions.25
Even less clear is why some patients develop both TD lesions and conventional demyelinating lesions. Immunohistochemical studies in the immunopathogenesis of ADEM have suggested a potential role for the inflammatory cytokines tumour necrosis factor-α and interleukin-1β which are toxic to myelin and oligodendrocytes. These cytokines appear to be expressed differentially in the neuraxis suggesting that there might be different molecular microenvironments within the CNS.26 It is conceivable that subtle differences in these microenvironments may not only influence why typical demyelinating lesions develop in different regions of the brain at different times but why some lesions such as in TD and ADEM are larger in size and some smaller. The larger size of TD lesions may also reflect dysfunction of inhibitory immunoregulatory cells in these microenvironments in susceptible individuals.
TD lesions can occur in diseases other than MS and there are case reports of these lesions being due to viral infections including HIV, other autoimmune diseases such as systemic lupus erythematosis, Sjogren's syndrome, Behcet's disease and neuromyelitis optica (NMO) and drugs such as tacrolimus.27–33 There are also cases describing an association with malignancy particularly renal cell carcinoma.23 ,34
Retinal vasculopathy with cerebral leukodystrophy
An important but very rare differential diagnosis for TD lesions is retinal vasculopathy with cerebral leukodystrophy. This is an autosomal dominant condition associated with mutations in the TREX1 gene.35 Patients with this condition present, commonly within the fourth or fifth decade, with pseudotumours, white matter lesions and visual disturbance that sometimes, but not always, may be associated with systemic signs including renal and/or hepatic impairment, Raynaud's phenomenon or rash. The pseudotumoural lesions may be difficult to distinguish from either tumour or TD and often demonstrate a good clinical and radiological response to corticosteroids. As in MS, there may be relapses but the course of the disease is usually more relentlessly progressive leading to death within 10 years. Unfortunately, there are no treatments that have been shown to modify the course of the neurological disease in retinal vasculopathy with cerebral leukodystrophy.36 Family history is key but may not always be appreciated. White matter lesions pathologically resemble postradiation vascular damage. Pathology of pseudotumours demonstrates areas of coalescing ischaemic necrosis accompanied by vasculopathy with fibrinoid occlusion of small vessels. There may be perivascular lymphocytic infiltrates but not frank vasculitis. Ultrastructural studies show characteristic multilamellar thickening of subendothelial basement membranes. Interestingly, different mutations in the TREX1 gene are associated with the autoimmune condition systemic lupus erythematosis.37
What is the likelihood of further attacks?
While the initial diagnosis of tumefactive demyelination may be difficult, once a diagnosis is established what does this mean for a patient's later clinical course? While long term data are limited, the available evidence suggests that approximately two-thirds of patients will follow a relapsing remitting course typical of conventional MS after a first TD lesion (see also Prognosis section). Almost all of the remaining patients will have no further attacks of demyelination consistent with a monophasic CIS or ADEM. Primary progressive neurological decline from onset is very rare.3 ,6
Interestingly, a small group of patients who suffer a second clinical episode will relapse with further tumefactive lesions.34 ,38 In 54 patients with TD disease, 16.7% of patients developed new tumefactive lesions over a median follow-up period of 38 months.6 On occasion, tumefactive relapses occur in the context of other more typical MS demyelinating lesions but there may be a rare subset of individuals who only experience relapse with TD lesions.34 It may be that recurrent TD is simply a phenotypic variant of conventional MS but the possibility exists that patients with recurrent TD lesions may have a distinct subtype of demyelinating disease.2 At this stage, it is not possible to predict which patients are more likely to develop a further tumefactive event and which will develop conventional MS lesions.
A suggested approach to the treatment of a typical TD lesion in adults is shown in figure 2.
Treatment of acute lesions
In occasional patients with minimal or no disability from a small tumefactive lesion an expectant approach can be adopted. In most instances, however, the size of a tumefactive lesion and its clinical sequelae mean that some form of treatment is necessary. TD is sufficiently rare that randomised controlled therapeutic trials have not been possible and treatment is based on the results of physician experience supported by individual reports and case series. There is broad agreement that, as with any disabling MS relapse, treatment with corticosteroids should be first line therapy for an acute symptomatic tumefactive lesion. The largest case series in which treatment of TD was assessed showed that more than 80% of patients make a significant response to treatment with corticosteroid.6
PEX appears to be a reasonable second line approach in those patients in whom corticosteroids are ineffective. One such case of biopsy-proven TD had no clinical response to IVMP on two separate occasions but improved considerably with PEX over 3 months despite radiological progression on the scan. When this patient relapsed 2 years later with another tumefactive lesion they were once again poorly responsive to IVMP but improved following PEX.40 In two further cases of biopsy proven TD, PEX was effective where prior IVMP had failed but on these occasions there was clinical and radiological response.41 ,42 In another study, two more cases responded to PEX after corticosteroids were ineffective with one of these cases requiring as many as 21 cycles of PEX.24 These reports would be in keeping with data from a randomised trial that showed PEX is beneficial in patients with mixed central nervous system inflammatory demyelinating disease who have failed to respond to corticosteroids.43
A variety of other treatments have been proposed in refractory cases. Many of these agents have the potential benefits of exerting a disease modifying effect as well as acutely improving disability due to a single large lesion in a refractory relapse. A recent case report describes beneficial effects with the anti-CD20 B cell monoclonal antibody Rituximab in a patient who had initial response to dexamethasone but relapsed on two occasions during attempts to wean corticosteroids. This patient improved clinically following Rituximab infusion with evidence of radiological improvement when rescanned at 1 month. At 12 months follow-up off steroids he had remained relapse-free.44 Rituximab was also tried with success, immediately following PEX, in a patient with MS and tumefactive lesions whose individual relapses responded to IVMP but who had rapidly evolving disease that did not respond to intravenous immunoglobulin, Mitoxantrone, interferon β or Natalizumab.45 Cyclophosphamide and intravenous immunoglobulin appear to have been used with more success in the paediatric population than in adults.39
Disease modifying therapy
At the moment there is insufficient evidence to recommend commencing MS disease modifying therapy (DMT) after an initial TD event in the absence of clinical or collateral radiological evidence of dissemination in space and time. Many clinicians would favour using DMTs only after MS diagnostic criteria have been fulfilled according to the 2010 revised McDonald criteria.46 ,47 An argument for considering a first, isolated TD event as the equivalent of a demyelinating CIS could certainly be made and there is evidence that DMT treatment following a conventional demyelinating CIS delays a second clinical attack and therefore conversion to MS. There may also be minimal improvement in long term disability as a result of this more aggressive approach.48 Hence, a DMT could be commenced if the TD CIS was particularly severe or disabling or if there were other markers to suggest a high risk of conversion to MS, for example, positive unmatched OCBs in the CSF or multiple other typical asymptomatic demyelinating lesions. What is not certain, however, is to what extent a single isolated TD event at presentation is representative of other conventional CIS patients in terms of risk of progression to MS. At present, the available evidence indicates that the time to a second attack following an initial TD lesion is already delayed compared with conventional CIS and that a subset of patients may have a more benign course (see Prognosis section). The commencement of DMT for CIS is also not without the potential for adverse effects48 and this would need to be weighed carefully in the shared decision making between an individual patient and their treating clinician.
When DMT therapies are commenced, the available data suggest that traditional first line MS therapies such as interferon-β and glatiramer acetate are the most commonly used. In one study, interferon-β was used in 48.1% and glatiramer acetate was used in 9.3% of TD patients.6 There are no data available to determine whether the efficacy of these first line treatments is different in patients with TD lesions compared with typical relapsing remitting MS (RRMS) or whether interferon-β should be preferred to glatiramer acetate.
A potential role for Fingolimod in the treatment of MS with TD lesions is yet to be determined but recent published cases would indicate that a cautious approach to using this agent is warranted. One patient with relatively mild MS discontinued interferon-β due to side effects. Within 4 months of commencing Fingolimod she relapsed due to two tumefactive lesions in the right hemisphere which regressed on serial follow-up imaging postcorticosteroids.49 Another patient with known MS who switched from interferon-β to Fingolimod suffered a severe relapse required corticosteroids, PEX and eventually craniectomy for subfalcine herniation related to vasogenic oedema. When this patient had a further TD relapse corticosteroids, PEX and cyclophosphamide were required.50 A third patient suffered a relapse secondary to a tumefactive lesion 6 weeks after switching from interferon-β to Fingolimod and later stabilised on Natalizumab.51 A fourth patient developed tumefactive demyelination 16 weeks after switching from Natalizumab to Fingolimod.52 A fifth patient with MS participating in the FREEDOMS trial developed a TD-like lesion described as haemorrhaging focal encephalitis with additional evidence of markedly increased MS disease activity 7 months after commencing Fingolimod.53 On a related note, a patient with NMO spectrum disorder and positive aquaporin 4 antibodies treated with Fingolimod (when the diagnosis was thought to be MS) developed an extensive tumefactive lesion.54
Clearly, these observations raise the possibility of more than just a chance association between TD lesions and Fingolimod that warrants further attention. If there is an association, then the underlying mechanism is speculative. Fingolimod is a sphingosine-1-phosphate receptor modulator that acts to reduce egress of lymphocytes from peripheral lymph nodes. If inhibitory immune cells were preferentially affected in a subset of individuals then a paradoxical effect of Fingolimod could be imagined in which large demyelinating lesions and increased disease activity could result.49 ,51
Natalizumab has been in clinical use for a longer duration than Fingolimod. While tumefactive lesions have been reported in patients with MS and NMO undergoing treatment with Natalizumab55–57 these cases are less frequent in the literature. In addition, there is one report of Natalizumab being beneficial in a patient with rapidly evolving relapsing tumefactive disease who remained relapse free 12 months after initiation of treatment.40
As previously discussed, many TD lesions do not require biopsy and so those lesions that are atypical enough that a biopsy appears warranted may not necessarily reflect the wider TD cohort. Therefore, data regarding prognosis in biopsy proven TD need to be interpreted carefully when extrapolating to include all TD. It may also be that patients with TD lesions associated with MS do not share the same prognosis as solitary TD lesions which appear to occur without supporting evidence of MS or as patients who have relapsing TD disease without conventional MS lesions. Bearing this in mind, the largest and most rigorous study of patients with tumefactive demyelination examined 168 patients with biopsy confirmed disease.3 In this study, in a cohort with both MS and non-MS associated TD lesions, the time to a second event in patients presenting with a TD lesion as their first clinical event (ie, fulfilling dissemination in time) was 4.8 years compared with 1.9–3 years for a typical demyelinating event cited in other studies.23 ,58–60 However, the presence of TD lesions did not appear to greatly impact on overall prognosis. There was a slight trend for more favourable prognosis in those patients with tumefactive lesions when accompanied by more typical MS demyelinating lesions on MRI. This group of patients was marginally slower to reach expanded disability scale score 3.0 at median 38 months of follow-up compared with those who had a single unaccompanied TD lesion at presentation.
Another study examined 14 patients with TD lesions as their first clinical event, of which eight had biopsy proven TD.24 This smaller cohort was more homogeneous in their baseline characteristics in that none of the patients had any additional typical demyelinating lesions at the time of their presentation with TD and only one had positive OCBs in the CSF. At median follow-up of 3.5 years only two patients had developed clinically definite MS. A more benign course following an isolated TD event has been reported in another small case series.61
The reason for a delay to a second clinical episode is unclear and could, at least in part, reflect more aggressive immunomodulatory treatments given to patients with TD lesions. It could also reflect a difference in the fundamental immunopathogenesis of TD lesions as compared with lesions of conventional RRMS but too little is known and more research is needed.
Tumefactive demyelination is an uncommon cause of neurological symptoms. The presence of radiological features such as an open, ring-enhancing lesion on MRI imaging can be a useful clue to diagnosis and can spare patients the need for brain biopsy. It is not clear if tumefactive demyelination is immunopathologically distinct from MS. Treatment of acute lesions is with corticosteroids but the timing of longer term DMT is less clear. When a diagnosis of MS is fulfilled, treatment should therefore be with conventional first line MS DMT agents and escalated to second line DMTs if those agents fail to control further events.
Contributors TAH conceived the idea for the article, performed the literature search, wrote the article and is the guarantor. JC was involved in the planning of the article, revised it critically for important intellectual content and gave final approval of the version to be published.
Competing interests TAH: Wishes to acknowledge assistance with travel to conferences from Bayer-Schering and Novartis in the past 3 years. JC: Has no relevant competing interests.
Provenance and review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.