Книги по МРТ КТ на английском языке / MR Imaging in White Matter Diseases of the Brain and Spinal Cord - K Sartor Massimo Filippi Nicola De Stefano Vincent Dou

phocytes and larger mononuclear cells (Vanderzant et al. 1988). In addition, involvement of the ventricular vessels has been reported in the past (Reik et al. 1983). Segmental inflammation and necrosis involving the small leptomeningeal and parenchymal blood vessels is more commonly seen; however larger vessels may at times be involved (Moore and Cupps 1983).

Fibrin thrombosis has also been reported (Vanderzant et al. 1988). Kolodny et al. (1968) reported a higher predilection for leptomeningeal involvement compared with parenchymal vessels. The experience of Barrow Neurological Institute supports this observation. The surrounding brain parenchyma may also demonstrate areas of both gross and microscopic ischemic and hemorrhagic infarction in evolving stages in addition to reactive astrocytosis (Kolodny et al. 1968). The presence of reactive changes in the surrounding brain parenchyma is not a consistent feature. In addition, loss of myelin with or without axonal degeneration may also be present (S. Coons, pers. commun.). The inflammatory infiltrate is comprised of macrophages, lymphocytes, multinucleated giant cells, and plasma cells,with or without the presence of granulomata (Moore and Cupps 1983). The focal, segmental, and heterogeneous nature of the disease process may lead to patchy granulomata formation as well as areas free of granuloma.A necrotizing vasculitis may be seen along side a non-granulomatous PACNS (Lie 1992). Necrotic changes within granulomata and presence of eosinophils are usually not observed (Koo and Massey 1988). Focal and segmental involvement is not limited solely to brain parenchyma and meninges, but can also be present in the spinal cord.

Aside from the standard pathological techniques utilizing a variety of stains, other techniques are also available to determine the presence of PACNS. The role of electron microscopy is yet to be determined in the diagnosis of PACNS; however, it is useful for assessing the presence of viral inclusion particles. Immunofluorescence is useful in detecting the presence of infectious agents but is of little value in the diagnosis of PACNS.

20.4

Immunological Mechanisms

The primary events that initiate this disease process remain unknown at present. In addition, the immunological mechanisms that result in the histopathological changes seen in PACNS are not completely understood. In some cases an immune complex-mediated process

is responsible for damage to the intimal surface of blood vessels and cell-mediated immune mechanisms, as is suggested by the presence of granulomata. This is also supported by the presence of a lymphomononuclear cell infiltrate seen in addition to endothelial cell proliferation (Moore 1995).

20.5

Clinical Presentation

Nearly every neurological symptom has been reported at least once in cases of PACNS (Calabrese et al. 1997). Headache and confusion are the most common presenting symptoms. Hemiparesis, language difficulties (expressive and receptive aphasia), hemisensory loss, cranial nerve abnormalities, and seizures may also be observed during the disease process. Myelopathic symptoms may also be seen as a result of spinal cord involvement (Giovanini et al. 1994). In addition, patients may also present with symptoms correlating with diffuse white matter disease (Finelli et al. 1997). Visual disturbances and hallucinations have also been reported (Cravioto and Feigin 1959; Nurick et al. 1972). Focal and multi-focal abnormalities occur in more than 80% of cases (Moore and Cupps 1983) with focal seizures reported in approximately 5% of cases (Moore and Richardson 1998). Some cases presenting with intracranial hemorrhage have also occurred (Kumar et al. 1997; Yasuda et al. 1993). In the case series by Duna et al. (1995), CNS hemorrhage was documented in 11% (18 of 168) of patients studied.

Diffuseneurologicaldysfunctionisseeninaggressive cases with resultant coma and death.In PACNS,systemic symptoms, such as fever, rash, arthralgias, myalgias, or arthritis, are usually not observed. When systemic symptoms are present, the focus must be directed towards a systemic disease process with secondary CNS involvement. The progression of symptoms may vary greatly in cases of PACNS owing to the heterogeneity of the natural history of the disease. There are no specific symptoms pathognomonic for PACNS.

20.6

Benign Angiopathy of the Central Nervous System

Reports of a benign angiopathy of the CNS (BACNS) have been proposed as a subset of PACNS in previously published studies (Bettoni et al. 1994;

Calabrese et al. 1993; Hajj-Ali et al. 2002; Snyder and McClelland 1978). These patients differ with respect to their clinical presentation, requirements for long-term immunosuppression, and prognosis.

Calabrese et al.(1993) first proposed the existence of patients with PACNS who had a more “benign” course compared to individuals with pathologically documented granulomatous angiitis of the CNS. Their course was described as monophasic rather than progressive. It was observed that they were primarily females (female:male ratio 4.3:1) who presented most commonly with complaints of headache with or without the presence of a focal neurological deficit.In addition, cerebrospinal fluid (CSF) studies were normal to mildly abnormal along with findings on cerebral angiogram consistent with vasculitis. Moreover, these individuals required less immunosuppressive treatment than traditionally used. Although the underlying pathogenesis was not clearly understood, reversible vasoconstriction or vasospasm was proposed as the most likely cause of BACNS as compared with an inflammatory process in PACNS.

An evaluation of 16 cases of presumed BACNS without histological confirmation was reported by Hajj-Ali et al. (2002). Angiographic follow-up data on 10 of 16 cases were analyzed, in addition to clinical outcomes as assessed by phone interview utilizing the Barthel index and a specifically designed cognitive index (Hajj-Ali et al. 2002). Results showed improvement in all repeat cerebral angiograms obtained after treatment had consisted of varying courses of glucocorticoids, calcium channel blockers, and cytotoxic medications. Clinical recovery was seen in 94% of cases. No deaths were reported; however, there was a 6% relapse rate observed. In those patients evaluated by the Barthel index, there was no lasting disability in 71% of cases. A different outcome was reported by Woolfenden et al. (1998) who published retrospective data on 10 patients with angiographically defined PACNS. They concluded that these patients neither had a benign outcome nor monophasic course.

Further data are needed to determine if such a distinct group of patients exist. The presumed pathological mechanisms underlying this proposed subtype is still lacking. It is noteworthy that the cases included in the study by both Calabrese et al. (1993) and Hajj-Ali et al. (2002) lacked histopathological confirmation. In previously published studies, a fair amount of patients who underwent brain biopsy for presumptive PACNS were found to have other diagnoses (Alrawi et al. 1999; Chu et al. 1998).

At our institution, we feel that a form of “benign” angiopathy or angiitis is a misnomer as there usually

is nothing benign about the patient’s clinical presentation or course. More studies are needed to determine whether in fact there is a separate subgroup and, if so, its natural history and pathogenesis.

20.7

Diagnostic Studies

20.7.1 Serological Tests

No specific serological diagnostic test(s) exist for PACNS. Instead, a comprehensive serological workup is performed to exclude other disorders. Recommended tests include a complete blood count, comprehensive metabolic panel including liver function tests, and C-reactive protein. Erythrocyte sedimentation rate (ESR), a common serological marker assessed in the PACNS, is usually negative. The ESR was found to be positive in 8–35% in two studies (Calabrese and Mallek 1988; Lie 1992). Patients should also be evaluated for connective tissue disorders (e.g., anti-nuclear antibody, anti-neutrophil cytoplasmic antibody). Depending on the clinical context, human immunodeficiency virus (HIV), hepatitis, syphilis, Lyme disease, and sarcoidosis also need to be considered in the differential diagnosis.

20.7.2

Cerebrospinal Fluid Analysis

In general, CSF studies are important in the diagnosis of PACNS. Abnormal CSF profiles are seen in 80–90% of pathologically documented cases of PACNS (Calabrese et al. 1997). Characteristically, a mononuclear pleocytosis with or without increased protein levels are seen (Oliveira et al. 1994; Vollmer et al. 1993). A previous report of 40 cases of pathologically proven PACNS demonstrated moderately to markedly elevated protein levels, and normal or mild decreased glucose levels (Vollmer et al. 1993). Oligoclonal bands and increased IgG synthesis may also be present. Oliveira et al. (1994) suggested that following the plasmatic albumin transudation and intrathecal IgG synthesis profiles may be a reliable marker for monitoring the disease process and therapeutic response. The appropriate stains, cultures, and studies [herpes simplex virus–polymerase chain reaction (HSV–PCR), Cocciodes immitis, syphilis, etc.] should be performed to evaluate for the presence of infectious etiologies.

20.7.3

Brain Computed Tomography

Computed tomography (CT) of the brain is abnormal in up 67% of cases (Siva 2001). This imaging modality, however, is not sensitive in the diagnosis of PACNS. Findings may include areas of low signal intensity suggestive of an ischemic event. Intraparenchymal or subarachnoid hemorrhage are uncommonly seen but have been reported.CT is also valuable during instances when magnetic resonance imaging (MRI) cannot be performed or is unavailable.

Stone et al. (1994) compared the diagnostic sensitivities of CT, lumbar puncture and magnetic resonance imaging in 20 patients with angiographically proven PACNS (ages ranging from 7-72 years). Of the 17 patients who underwent brain CT, 11 were found to be abnormal with 35% of cases failing to demonstrate any significant abnormality. They concluded that CT is a useful diagnostic tool when used as an adjunct with CSF studies.

20.7.4

Magnetic Resonance Imaging

Magnetic resonance imaging of the brain and spinal cord has been proven to be valuable in the evalu-

ation of PACNS (Fig. 20.1). Although abnormalities seen are not pathognomonic for PACNS, this noninvasive and reproducible technique is sensitive to parenchymal changes regardless of the size of vessels involved by a vasculitic process (Harris et al. 1994). Patients who fail to demonstrate abnormalities on MRI are unlikely to manifest changes on angiography or brain biopsy suggestive for PACNS. Cases have been reported, however, in which patients with angiographically proven PACNS have had negative brain CT and MRI scans (Greenan et al. 1992; Stone et al. 1994).

In their review of six patients (two biopsy proven) with PACNS, Campi et al. (2001) reported discrete or diffuse supraand infratentorial lesions involving the deep and subcortical white matter (Campi et al. 2001a). This is consistent with other reported cases of PACNS (Ehsan et al. 1995; Greenan et al. 1992; Scully 1989). The lesions in PACNS are usually enhancing in up to 90% of cases. In addition to the brain, enhancement has been noted in the Vir- chow-Robin spaces and spinal cord (cervical and thoracic segments). Other authors have reported a predominance of white matter involvement on T2weighted images suggestive of demyelinating disease (Alhalabi and Moore 1994). Non-specific scattered focal and linear areas of high-signal abnormality involving the brainstem, cerebellum, and cere-

bral white matter has also been seen in addition to areas of recent infarction (Koo and Massey 1988; Shoemaker et al. 1994). Leptomeningeal enhancement with modest parenchymal involvement has also been reported (Negeshi and Sze 1993). Johnson et al. (1989) reported a case of granulomatous angiitis masquerading as a mass lesion with MRI findings revealing bilateral high signal intensity parietooccipital lesions extending across the splenium of the corpus callosum. Idiopathic granulomatous angiitis of the CNS was also reported as manifesting as diffuse white matter disease in a patient presenting with progressive spastic paraparesis (Finelli et al. 1997).

Harris et al. (1994) studied the utility of MRI as a screening tool in 92 patients who underwent angiography for“exclude vasculitis”as the indication. The MRI data from 70 of the 92 patients were evaluated. Of the 92 patients studied, 11 patients were found of have intracranial vasculitis, 8 of which displayed abnormalities on angiography. The MRI data was available on 9 of the 11 cases, with all 9 demonstrating significant abnormalities. The authors advocated that brain MRI be performed in patients with suspected intracranial vasculitis. They concluded that a negative MRI was more valuable than a negative angiogram when trying to exclude the possibility of an intracranial vasculitis.

Currently, there are no MRI findings that are pathognomonic for PACNS; however, this modality is excellent in monitoring the natural history of the disease in addition to utilizing it as a tool for assessing for the presence of a treatment response (Ehsan et al. 1995).

20.7.5

Cerebral Angiography

Angiography is reportedly the most sensitive diagnostic modality available for the diagnosis of PACNS, although it is variable with respect to abnormalities seen (Alhalabi and Moore 1994; Ferris and

Levine 1973; Leeds and Goldberg 1971; Moore

1995). Sensitivities do vary based on the study reviewed. A retrospective review of 30 consecutive patients who underwent brain biopsy and/or cerebral angiography for the evaluation of PACNS revealed that cerebral angiography had a sensitivity and positive predictive value (PPV) of <30% (Duna and Calabrese 1995). Vollmer et al. (1993) reported 56% of 41 angiograms were abnormal in biopsyproved cases of idiopathic granulomatous angiitis of the CNS; however, only 27% (11 of 41) of these cases

demonstrated findings thought to be diagnostic of vasculitis. On the other hand, a sensitivity of 89% was achieved with angiography by Alhalabi and Moore (1994) when retrospectively evaluating angiograms in 19 patients.

Angiography identifies large and medium-sized vessel abnormalities suggestive of vasculitis,but findings cannot unequivocally establish the diagnosis for PACNS, and a distinction between primary and secondary causes of vasculitis cannot be made based on the data. The angiographic abnormalities of PACNS cannot be differentiated from secondary causes of angiitis of the CNS; these include, among others, infectious causes (Engelter et al. 2002; Lehrer

1966; Mackenzie et al. 1981; Rabinov 1963; Ramos and Mandybur 1975; Rosenblum and Hatfield

1972; Walker et al. 1973), neoplasms (Giang 1994; Greco et al. 1976), collagen vascular diseases (Liem et al. 1996; Trevor et al. 1972), illicit drugs (Gertner and Hamlar 2002; Giang 1994; Glick et al. 1994; Margolis and Newton 1971), and other causes that may mimic a CNS vasculitis (Behcet’s disease; Siva et al. 2001), Moya-moya disease (Coakham et al. 1979), fibromuscular dysplasia (Houser and Baker 1968), vasospasm, atherosclerosis (Ferris and Levine 1973), radiation injury, hypertensive vasculopathy, sickle cell anemia (Merkel et al. 1978), atrial myxoma embolism (Marazuela et al. 1989), refractory dermatomyositis (Regan et al. 2001), and medication toxicity (Berlit 1994; Giang 1994; Regan et al. 2001; Rumbaugh et al. 1971; Siva 2001).

Unlike magnetic resonance angiography (MRA), conventional angiography is an invasive procedure with associated risks. Previously published complication rates vary depending on the institution at which the procedure is performed. The complication rate at our institution following the evaluation of 1000 prospectively studied consecutive cases performed on 688 patients revealed a 1% overall incidence for a neurological deficit and a 0.5% incidence of a persistent deficit (Heiserman et al. 1994). The complications observed were seen in patients with a history of a cerebrovascular accident, transient ischemic attack, or carotid bruit (Heiserman et al. 1994).

Focal and/or multifocal segmental involvement of both small and medium-sized leptomeningeal and parenchymal blood vessels are angiographic findings seen in PACNS (Fig. 20.2). More commonly, single areas of focal abnormality are observed in a number of vessels rather than extensive areas of involvement on a single vessel (Alhalabi and Moore 1994). Also seen are vascular occlusions, collateral vessel formation, and prolonged circulation time (Alhalabi and

Moore 1994; Lie 1992). Aneurysms have also been identified (Nishikawa et al. 1998). Abnormalities observed on angiography do not always correlate with significant findings observed on MRI of the brain, likely owing to the extent of involvement of smaller vessels.

Alhalabi and Moore (1994) compared retrospective angiograms in 19 patients with PACNS along with prospective angiograms during the course of the disease. They noted areas of segmental narrowing were reversible if treatment was instituted early in the course of the disease. The initial areas of segmental narrowing were hypothesized to be secondary to inflammation and vasospasm. They also observed that areas of narrowing increased over time and that later in the course of the disease they were less likely to resolve. They suggested that serial angiography may be used to guide immunosuppressive therapy.

Although the sensitivity is highly variable, angiography provides another modality to help diagnose PACNS; however, brain and meningeal biopsies remain the gold standard for diagnosis.

20.7.6

Brain Biopsy

Histopathological confirmation by leptomeningeal and brain parenchymal biopsy remains the gold standard for the diagnosis of PACNS. Brain biopsy is the most specific diagnostic modality in PACNS. The optimal technique is wedge cortical biopsy with lep-

tomeningeal tissue which permits better histological definition and a reduction in artifact (Moore 1989). Considerable debate exists as to whether stereotactic biopsy is as effective as open-wedge biopsy in the diagnosis of PACNS, although the experience of Alrawi et al. (1999) demonstrated no significant difference between these two approaches.

There are several retrospective studies which have calculated the specificity to be between 60 and 97% (Alrawi et al. 1999; Chu et al. 1998; Duna and Calabrese 1995). A false-negative rate of approximately 10% has been previously reported in premortem biopsies (Calabrese et al. 1993). False-positive results are rare and may occur in the setting of lymphoproliferative disease. Histological confirmation is necessary for a definitive diagnosis; however, the data must be correlated with the clinical history and information obtained from other diagnostic tests. Cases of PACNS with the presence of concomitant cerebral amyloid angiopathy have been reported in the past which may complicate the pathological diagnosis in some patients (Fountain and Eberhard 1996; Riemer et al. 1999; Schwab et al. 2003).

In addition to verifying the presence of a vasculitic process, the biopsy is also necessary to exclude other conditions that may affect the intracranial vasculature and mimic vasculitis. These include demyelinating conditions [acute disseminated encephalomyelitis (ADEM), multiple sclerosis]; infections (HIV [Nogueras et al. 2002], cytomegolovirus [CMV], herpes zoster, progressive multifocal leukoencephalopathy, Creutzfeldt-Jakob disease, Coxsackie A9,

brain abscess, Cocciodes immitis); reaction to various toxic substances (heroin, cocaine, amphetamine use); neoplasia (lymphoma, malignant histiocytosis, metastatic small cell carcinoma); sterile infarction; and sarcoidosis.

Biopsy material should be examined with routine and special stains in addition to tissue cultures to exclude other diagnoses. It is of benefit to examine tissue by electron microscopy to identify the presence of viral inclusions or ultrastructural abnormalities that may be present.

Brain biopsy often is deferred due to concern for potential complications surrounding the procedure as well as the possibility of a false-negative result. The morbidity of the procedure has been reported to be 3.3% by Chu et al. (1998) and a similar morbidity rate of 4.9% was observed by Alrawi et al. (1999) in their analysis of 61 consecutive biopsy patients.

The site of the biopsy should be determined by location of active lesions observed on MRI. Biopsy of a radiographically abnormal region increases the sensitivity of the procedure and diagnostic accuracy is enhanced by sampling both the leptomeninges and cortical and subcortical tissues (Calabrese et al. 1992; Chu et al. 1998; Parisi and Moore 1994). Data from Alrawi et al. (1999) contradict this data as all identified cases of PACNS contained brain parenchyma involvement and only 77% demonstrated involvement of the leptomeninges. It is important to note, however, that sampling of the leptomeninges allows for the diagnosis of processes other than PACNS. The optimal location for biopsy would be a site of active disease on MRI, providing it is in a non-sensitive area of the brain; otherwise, sampling the anterior lip of the non-dominant temporal lobe is preferred and least harmful to the patient. Due to the focal and segmental nature of PACNS, biopsy results may vary with only 66–75% of biopsies being diagnostic (Siva 2001).

Duna and Calabrese (1995) reported results of 15 brain biopsies (7 stereotactic and 8 via craniotomy) in the diagnosis of PACNS. The sensitivity was 53%, specificity 87%, PPV 80%, and negative predictive value (NPV) 70%. One false-positive result was observed in a patient who was subsequently diagnosed with CNS lymphoma. Of the four true-positive biopsies, one demonstrated vasculitic changes in the cortex alone, one in the leptomeninges alone, and the other two with both parenchymal and leptomeningeal involvement (Duna and Calabrese 1995). A low sensitivity of 36% was observed in the report published by Alrawi et al. (1999). A specific diagnosis, however, was observed in 75% of the cases. This

highlights the value of brain biopsy in not only identifying cases of PACNS but also those processes that it may mimic (Alrawi et al. 1999). The low sensitivities observed in both of these studies suggest that a negative biopsy result does not rule out the possibility of PACNS.

A retrospective analysis of 25 patients, 10 treated with immunosuppressive medications and 15 untreated, with PACNS were studied to assess the prognosis of those patients who had a negative brain biopsy (Alreshaid and Powers 2003). The addition of immunosuppressive therapy did not significantly enhance the outcome of these patients compared with the untreated group. Patients were stratified into two groups; “good”, if the patient was residing at home; and “poor” if the patient resided in an extended-care facility or died. However, one should be cautioned that treatment regimens were not disclosed in this report and no specific treatment criteria were noted for those who received therapy. Additionally, it was not clear if other medical or social factors influenced their results.

20.7

Other Diagnostic Modalities

The role of positron emission tomography (PET), single-photon-emission computed tomography (SPECT), magnetic resonance spectroscopy (MRS), and transcranial doppler (TCD) is still undetermined (Ritter et al. 2002). These diagnostic modalities are neither sensitive nor specific in the diagnosis of PACNS.

Table 20.2 contains modified diagnostic criteria recommended by Moore (1989) for the diagnosis of PACNS.

Table 20.2. Antemortem diagnostic criteria for primary angiitis of the central nervous system modified from Moore (1989, 1998).

1.Clinical features consistent with recurrent, multifocal, or diffuse disease

2.Exclusion of an underlying systemic inflammatory process or infection

3.Neuroradiographic studies - cerebral angiography demonstrating findings supporting the diagnosis of a vasculopathy

4.Brain biopsy to establish the presence of vascular inflammation and exclude infection, neoplasia or alternate causes of vasculopathy

5.A CSF study consistent with CNS inflammation (pleocytosis, increased protein) and excluding infection and neoplasia

20.8 Treatment

Currently, the optimal treatment regimen for PACNS is not clearly delineated. To date, there have been no multicenter, randomized, or placebo-controlled studies investigating the efficacy of varying treatment regimens. This is most likely due to the paucity of documented cases. Our understanding and clinical approach therefore is based primarily on anecdotal experience and data from previously published individual case reports and modest case series.In general, there is agreement that the underlying pathogenesis of this condition is neuro-immunological, and thus treatment is directed towards modulating the immune system. Because of the clinical, radiological, and histopathological heterogeneity, it can be argued that the optimal treatment for one patient may not be as efficacious for another.

PACNS in previous years was viewed as uniformly fatal. There are now newer methods of modulating the patient’s immune system, and with this availability the prognosis has become more favorable.At present, standard treatment recommendations include the use of high-dose corticosteroids and immunosuppressive agents. Azathioprine, methotrexate, and antiplatelet treatment for maintenance therapy have also been used in the past (Craven and French 1985; Griffin et al. 1973; Oliveira et al. 1994; Siva 2001; Zivkovic and Moore 2000). The role of IVIg, plasmapheresis, the interferons, and monoclonal antibody therapy in PACNS remains unclear.

Currently, it is generally agreed upon that aggressive immunotherapy should be instituted in cases where progressive neurological decline is present. This is with the understanding that an attempt to confirm the diagnosis of PACNS by histopathology, as well as conventional diagnostic modalities, has been made, and that other causes for the presenting symptoms have been investigated and ruled out. This approach is also advocated by other investigators (Calabrese et al. 1997). The optimal duration of treatment is not clearly established. Recommendations for a treatment period of 6–12 months after the clinical symptoms have stabilized have been made, although there is no conclusive evidence to support this (Calabrese 1995; Moore 1994). Treatment duration should be guided by patient response to therapy and current clinical status.

Most commonly, the treatment regimen used in PACNS includes high-dose methylprednisolone at a dose of 1000 mg/day for 3–7 days followed by oral prednisone at 60 mg/day with slow taper. The expe-

rience with cyclophosphamide in the treatment of PACNS has been extensively reported (Barron et al. 1993; Fountain and Lopes 1999; Mandybur and

Balko 1992; Moore 1989; Riemer et al. 1999; Siva 2001; Woolfenden et al. 1998). Cyclophosphamide, oral or intravenous, is administered concomitantly with steroids at a dose of 2.0–2.5 mg/kg day–1. If intravenous therapy is selected, pulsed intravenous cyclophosphamide (500–1000 mg/m2) can be administered every second week for the first 6 weeks followed by monthly intervals.

Reviewing five patients with biopsy-proven PACNS, Moore (1989) recommended for the initial 6 weeks of therapy a regimen of prednisone at 40– 60 mg/day combined with cyclophosphamide, dose of 100 mg/day (Moore 1989). Oral prednisone alone at a dose of 40–100 mg/day has also been advocated in a report by Crane et al. (1991) in their review of 11 patients with angiographically proven IACNS. Of the 11 patients studied,10 achieved clinical remission following this regimen of steroid therapy alone. Cupps et al. (1983) reported four cases of biopsy proven (1) and unproven (3) PACNS. The single biopsy-proven case involved isolated spinal cord involvement as no brain abnormalities were identified. Sustained clinical remission was achieved with a combination of cyclophosphamide and alternate-day prednisone therapy in all patients (Cupps et al. 1983). The successful use of the prednisone and cyclophosphamide combination has also been seen in pediatric patients (Lanthier et al.2001; Barron et al.1993).Lanthier et al. (2001) reported on two pathologically proven cases of IACNS. One case was successfully treated with a regimen of prednisone 2 mg/kg day–1 and cyclophosphamide 2 mg/kg day–1, the other with prednisone alone. The patient who received cyclophosphamide in addition to steroids remained symptom free for 6 years after discontinuation of therapy. A successful response was achieved in a 12-year-old patient with cerebral angiogram presumed IACNS with monthly pulse cyclophosphamide (750 mg/m2) for 6 months followed by treatment every 3 months for 1 year (Barron et al. 1993).

The most sensitive as well as specific modality for monitoring the response to treatment in PACNS continues to be investigated. At present, serial brain MRI scans is the most sensitive means of monitoring disease progression. This imaging modality has been utilized in the past by other investigators to monitor both disease progression and response to treatment (Ehsan et al. 1995). Serial cerebral angiograms have also been used to assess the efficacy of the treatment regimen utilized (Alhalabi and Moore 1994; Stein

et al. 1987). Some authors have proposed the use of subsequent serial CSF profiles as a guideline for monitoring disease activity and response to therapy (Oliveira et al. 1994). There have been, however, no reports of repeat histopathological analysis to assess therapeutic success. Since PACNS is such a heterogeneous disorder, it is unclear if repeat analysis of the leptomeninges or brain parenchyma would provide conclusive evidence of therapeutic success, since a lack of non-inflammatory findings may represent disease suppression rather than cure.

Further investigations are necessary to not only determine the optimal therapeutic regimen for PACNS, but also to prevent treatment complications. The sequelae of shortand long-term steroid use are well known. Cyclophosphamide has several wellknown side effects primarily related to bone marrow suppression (Bradley et al. 1989). Riemer et al. (1999) reported on a single case of PACNS that was treated for 5 years with cyclophosphamide (cumulative dose 100 g). The patient ultimately died of cyclo- phosphamide-induced myelodysplastic syndrome. Post-mortem evaluation revealed vascular scarring and amyloid angiopathy; however, evaluation of the brain parenchyma revealed the absence of inflammation (Riemer et al. 1999).

Ideally, a national database encompassing patients with biopsy-proven and unproven PACNS, and their response to treatment, may help to define the optimal treatment regimen for this disorder.

20.9 Prognosis

The prognosis for patients with PACNS is variable due to the heterogeneity of the disease process with respect to progression and response to treatment. In addition, the natural history of the disease and long-term outcomes, for the most part, are unknown. Clearly important in improving the overall outcome is the rapid definitive diagnosis of PACNS with institution of appropriate treatment.

20.10

Case Discussion

The following case seen at this institution exemplifies the heterogeneity of PACNS, both clinically and pathologically, and the inherent difficulties in diag-

nosis and treatment. Some of the histopathological findings described here are unique and have not yet been reported in cases of PACNS.

A previously healthy right-handed 31-year-old man with no significant medical history initially presented with complaints of headache and transient right arm and leg weakness which began 7 days prior. His headache was described as being constant and unremitting with sharp pain located bi-frontally.

His general medical examination was unremarkable. No focal findings were observed on neurological examination. A complete blood count (CBC), comprehensive metabolic panel (CMP), including liver function tests, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), urine toxicology screen (U-TOX), serum protein electrophoresis (SPEP), extended anti-nuclear antibody panel (ANA), complement level, serum amino acids, and lactate level, were ordered and found to be within normal limits. On CSF analysis his opening pressure was 17 cm H20, RBC 1/mm3, nucleated cells 11/mm3 (88% lymphocytes, 4% polymorphonuclear cells), glucose 74 mg/ dl, and protein 35 mg/dl. The CSF cultures were negative for growth. The CSF cytology was negative for malignant cells and herpes simplex virus–poly- merase chain reaction (HSV–PCR) test was negative.

An electroencephalogram (EEG) demonstrated focal slowing of the background rhythm over the left hemisphere. No electrographic seizure events or epileptiform activity were observed.An MRI of the brain demonstrated non-contrast enhancing, diffusionnegative T2 signal hyperintensities involving the left frontal and temporal lobe regions (Fig. 20.3a,b).

The patient was treated with high-dose methylprednisolone (1 g/day) for 3 days and was discharged to home where he completed a 15-day course of acyclovir. Improvement of his symptoms was reported at the time of discharge.

One month later, the patient returned to the hospital with complaints of language difficulty. He reported “mixing up”his words during conversation.He also had complaints of increased headache, confusion, muscle aches, fatigue, and left thumb twitching lasting seconds to minutes. Neurological examination was again found to be non-focal.A CBC,CMP,ESR,and CRP were again ordered and found to be within normal limits. Coagulation studies and rheumatoid factor were also found to be within normal limits. Repeat CSF analysis revealed nine nucleated cells per cubic millimeter (74% polymorphonuclear cells, 17% lymphocytes), RBC 0/mm3, glucose 74 mg/dl, and protein 58 mg/dl. The CSF cultures and HSV–PCR were negative. The CSF lactate was within normal limits.

The patient underwent a repeat EEG with focal background slowing observed over the left frontotemporal region. No discrete electrographic seizure events were captured and no epileptiform activity was observed. A repeat MRI of the brain with gadolinium demonstrated non-contrast-enhancing and diffusion-negative T2-hyperintensities involving the left frontal, temporal, and occipital lobes (Fig. 20.3c). The patient underwent the same treatment course as his previous hospitalization with modest improvement of his symptoms and was discharged.

Two months following his second discharge from the hospital, the patient presented after a generalized tonic–clonic seizure preceded by reports of automatisms and abdominal discomfort. He also had complaints of right ear numbness and fullness, left forearm numbness, chest discomfort, and right hemispheric headaches. Neurological examination again was non-focal. An HIV test was ordered which was negative. Serological tests were ordered once again (p-antineutrophil cytoplasmic antibody, c-antineu- trophil cytoplasmic antibody, extended ANA panel, serum lactate, Lyme disease serologies, CRP) and found to be normal. He underwent a muscle biopsy to exclude the possibility of a mitochondrial disorder. Histological evaluation of the muscle architecture revealed normal findings. A four-vessel cerebral angiogram was normal with no signs of vasculitis observed. Magnetic resonance imaging revealed abnormalities now involving the right hemisphere with marked improvement of T2-hyperintensities observed previously involving the left hemisphere (Fig. 20.3d,e).

A stereotactic biopsy of the right frontal region was performed. Histopathology revealed lymphocytic invasion of both the meninges and vasculature (Fig. 20.4a–d). The surrounding parenchyma was significant for areas of infarction. There were no signs of hemorrhage observed. Staining of the cortex with von Kossa was positive, demonstrating spiculated calcium deposits throughout infarcted brain tissue (Fig. 20.4e). This was further confirmed by electron microscopy studies which had been performed initially to determine the presence of fungal and viral inclusion particles. Electron microscopy of brain tissue was negative for any infectious process. The sampled tissue was, however, significant for areas of dystrophic calcification in addition to abnormal calcium deposits adjacent to dendrites (Figs. 20.5, 20.6). Dystrophic calcification was also observed involving the mitochondria (Fig. 20.5c,d). Although the role of calcium in cell injury has been greatly studied,

this finding has not been described or previously reported in cases of vasculitis involving the CNS (Trump et al. 1980).

The patient was treated with daily oral prednisone (60 mg/day with slow taper over 1 year) and cyclophosphamide (2 g IV every month for 6 months followed by infusions once every 2 months) with good result. Azathioprine was also added to his treatment regimen. He continued on phenytoin and lamotrigine for his seizures. Since completing the treatment regimen, the patient has remained symptom free for 18 months. Serial MRI scans of the brain have demonstrated signs of encephalomalacia but no progression of disease or areas of increased T2-signal abnormality.

20.11 Conclusion

Physicians are faced daily with the challenges of accurately diagnosing patients by employing the current diagnostic criteria, recommended structural neuroimaging techniques, and/or guidelines put forth by ad hoc committees, consensus groups, and data from large, multicenter, placebo-controlled studies. A comprehensive understanding of the approach to PACNS from the etiology to optimal treatment is currently not clearly defined. This is a rare disorder that represents a unique diagnostic challenge for the neurologist given the lack of uniformity with respect to the clinical presentation, radiological and histopathological findings, and the varying natural history of the disease. This condition is restricted to the CNS with no signs of systemic involvement and biopsy of the brain parenchyma and leptomeninges is paramount in the diagnosis.A delay in the diagnosis and proper treatment of the disorder may cause increased morbidity and death. Current treatment strategies should be initiated aggressively, but only after confirmation of the diagnosis has been made. The diagnosis should not be made based on any single study. Instead, historical, clinical, serological, radiological, and histopathological correlation is needed for an accurate diagnosis. It is only by this approach that appropriate treatment interventions can be instituted in a timely manner to increase the likelihood of a better outcome. Future studies may focus on the genetic predilection for this condition which may provide insight into its pathogenesis and varied clinical course as well as treatment response.