Correspondence *Johns Hopkins University School of Medicine, Department of Neurology, Pathology Building 2-210, 600 North Wolfe Street, Baltimore, MD 21287, USA

Email toker@jhu.edu

The case

A 39-year-old man with remote traumatic brain injury sustained as a teenager and well-controlled seizures presented to a private neurology clinic with recent falls and dysphagia. In the few months before his presentation, he had developed insidiously progressive personality changes, social withdrawal, bradykinesia, hypophonia, and gait impairment. A review of his medical history revealed that he had also experienced mild fever and diarrhea during this same time frame. Outpatient laboratory studies, a brain MRI scan and electroencephalogram (EEG) were unremarkable at the time of presentation, and he had been commenced on ropinirole by his primary neurologist for possible early-onset Parkinson's disease. His social and family histories were unremarkable.

Following his presentation to the private neurology clinic, the patient was admitted to hospital. He was found to be anxious and in mild distress, and his pulse measured 100–110 beats per minute. He was awake and alert, but was markedly hypophonic to the extent that he appeared functionally mute. He communicated by using the 'thumbs up' and 'thumbs down' hand signals for yes and no, or by typing slowly. He was oriented, able to follow complex commands, attended to tasks such as "serial sevens" and typing paragraphs, had good calculation abilities, and displayed logical and abstract thinking. He had an emotionless, masked facial expression, and an incomplete, mixed vertical and horizontal supranuclear gaze palsy, with frequent saccadic (opsoclonic) and non-saccadic (nystagmoid) intrusions (for detailed description see Box 1 and Supplementary Video 1 online with legend). The patient also demonstrated rigidity of the upper extremities without cogwheeling. His strength, sensation, and reflexes were all normal. His gait was narrow-based with small steps and had a marked tendency toward retropulsion.

Laboratory tests for comprehensive metabolic panel, complete blood cell count, HIV, erythrocyte sedimentation rate, anti-nuclear antibodies, Lyme disease, anti-neutrophilic cytoplasmic antibody, anti-Ro and anti-La, angiotensin-converting enzyme, anti-thyroglobulin antibody, thyroid-stimulating hormone, vitamin B₁₂, rapid plasma reagin, Whipple's polymerase chain reaction (PCR), and serum protein electrophoresis were all normal. The test results for his paraneoplastic profile (anti-Hu, anti-Ma1, anti-Ma2, anti-Yo, anti-Ri, anti-CAR, anti-LEMS, and anti-CV2 antibodies) remained pending throughout this initial hospitalization. A contrast-enhanced brain MRI scan revealed frontal encephalomalacia and fluid-attenuated inversion-recovery (FLAIR) hyperintensity without enhancement, which was thought to be consistent with his remote head injury (Figure 1A). There was also subtle FLAIR hyperintensity of the right substantia nigra and left temporal lobe (Figure 1B). No remote MRI scans were available, but comparison of a head CT scan from 2 years previously with his recent admission head CT scan indicated that there had been an expansion of the frontal white matter hypodensities, particularly on the right side. Because of the patient's clinical presentation, the midbrain lesion was also considered to be new. Cerebrospinal fluid (CSF) testing revealed two white blood cells per high-power field (WBCs/hpf), no red blood cells, a total protein level of 0.6 g/l [60 mg/dl] (normal range: 0.15–0.45 g/l [15–45 mg/dl]) and a normal glucose level. CSF viral PCRs, cryptococcal antigen, acid-fast bacillus culture, cytology, flow cytometry, and 14-3-3 protein were all normal. Chest, abdomen and pelvis CT scans with contrast were also normal, as was a total body ¹⁸F-fluorodeoxyglucose positron emission tomography (FDG-PET) CT scan.

Figure 1 MRI scan results for the patient.

(A) Axial fluid-attenuated inversion-recovery (FLAIR) MRI scan of the brain at initial clinical presentation showing left greater than right frontal encephalomalacia and white matter hyperintensity. (B) MRI scan showing FLAIR signal changes in the right substantia nigra and left temporal lobe. The cerebral changes were deemed radiographically consistent with the patient's remote history of head injury. Mesencephalic changes were not noted in the radiology report, but are unlikely to have resulted from prior head injury. The extent to which cerebral changes were remote or subacute could not be definitively determined, as the patient had no prior MRI scans for comparison. Comparison of recent CT scans with remote CT scans, however, suggested evolution of the white matter pathology in the 2 years before clinical presentation.

Full figure and legend (19K)Figures & Tables indexDownload Power Point slide (144K)

Because the patient displayed a midbrain-predominant encephalitis with vertical gaze palsy and intrusive eye movements with no evidence of systemic malignancy to suggest a paraneoplastic syndrome, a tentative diagnosis of CNS Whipple's disease was made. He was commenced on treatment with intravenous ceftriaxone 2 g every 12 h for 3 weeks, which was temporally coincident with clinical improvement, including a clear reduction in the amount of intrusive eye movements and a mild improvement in his gait. Small bowel biopsy and CSF PCR assays to confirm Whipple's disease were negative. Because the apparent response to treatment was only partial, and the diagnosis of CNS Whipple's disease was unsupported by laboratory evidence, we elected to pursue diagnostic brain biopsy in the hopes of definitively identifying an underlying etiology. Unfortunately, open right frontal brain and meningeal biopsy (directed towards a region of MRI signal abnormality believed to be new) revealed only thickening of the cortical vessels, without inflammation, infection, or neoplasm. On completion of his course of intravenous ceftriaxone, he was discharged on trimethoprim 160 mg and sulfamethoxazole 800 mg twice daily with a diagnosis of possible CNS Whipple's disease. Over the ensuing weeks the patient's symptoms remained relatively stable, although he experienced an increase in his anxiety, prompting treatment with a selective serotonin reuptake inhibitor.

Three months later, however, the patient was admitted to another hospital with severe spasticity and tachypnea requiring intubation. The laboratory test results for his paraneoplastic profile from his earlier admission had by this time been reported, revealing anti-Ma2 antibodies which were indicative of paraneoplastic encephalitis. On further analysis, after normalizing for the total IgG in the serum and CSF, the serum anti-Ma2 antibody titer was 1:8,000 and the CSF titer was 1:32,000, indicating intrathecal synthesis of anti-Ma2 antibodies. A testicular examination was undertaken which was said to be normal, but a testicular ultrasound scan demonstrated a 1.4 cm mass. The patient was transferred to another hospital for a unilateral orchiectomy, which revealed a 1 mm embryonal carcinoma surrounded by extensive fibrous tissue. After resection of the primary tumor, his severe paraneoplastic neurological syndrome was treated with high-dose intravenous methylprednisolone 1,000 mg daily and intravenous immunoglobulin 4 mg/kg daily, both for 5 days, which was followed by continuous immunosuppression with prednisone 60 mg daily and mycophenolate mofetil 500 mg twice daily, but without apparent clinical improvement. His spasticity worsened—particularly in the upper body—despite therapy with intrathecal and oral baclofen at doses as high as 20 mg every 6 h, tizanidine 8 mg every 6 h, and diazepam 30 mg every 6 h. A gastrostomy tube had to be fitted because of swallowing difficulties. A repeat CSF test showed anti-Ma2 antibodies, with a total protein level of 58 mg/dl, and 2 WBCs/hpf.

Six months later, an MRI scan revealed a 0.75 cm³ homogeneously enhancing extra-axial dural-based mass along the clivus, thought to possibly be a metastasis. A biopsy was considered to be too risky, and the patient was empirically treated with one course of intravenous bleomycin 30 units on 3 separate days over a period of 3 weeks, intravenous etoposide 100 mg/m² for 5 days, and intravenous cisplatin 20 mg/m² for 5 days for presumed metastatic testicular carcinoma. A repeat MRI scan revealed radiographic remission of the clival mass. The patient's spasticity continued to worsen, and was refractory to the addition of dantrolene 50 mg daily (stopped because of thrombocytopenia and elevated hepatic enzyme levels), botulinum toxin injections 50–100 units per muscle, dronabinol 5 mg twice daily, and 6% phenol injections of 5–10 ml per muscle. As his upper body and bulbar spasticity increased, he experienced recurrent aspiration pneumonia, necessitating the placement of a tracheostomy tube for airway protection. Shortly thereafter, he died of cardiopulmonary arrest.

At his autopsy, microscopic examination of the brain (see Figure 2) revealed old traumatic changes, reactive astrocytosis in the amygdala, basal forebrain, and septum, a microglial nodule in the basal forebrain, and mild perivascular cuffing. The midbrain showed inflammatory changes, with microglial nodules and perivascular cuffing. The cerebellum was histologically normal. The inflammatory cells stained positive for CD45 (common leukocyte antigen) and CD68 (a marker of microglia). Additional staining revealed rare microglia and perivascular cuffing in the pons and medulla. The deltoid muscle revealed atrophic fibers with endomysial fibrosis, scattered central nuclei, and regenerating fibers, consistent with a myopathic process. There was no evidence of residual cancer.

Figure 2 Histological results of microscopic examination of the brain at autopsy.

(A) Microglial nodule in basal forebrain (hematoxylin and eosin stain). (B) Perivascular cuffing in the pons (hematoxylin and eosin stain). (C) Microglial nodule in midbrain with T lymphocytes (CD3 immunostain). (D) Microglial nodule in the pons with microglia (CD68 immunostain). (E) Perivascular cuffing by lymphocytes in the midbrain (CD45 immunostain). Courtesy of Barbara Crain, MD; Johns Hopkins Department of Pathology, Baltimore, MD, USA.

Full figure and legend (51K)Figures & Tables indexDownload Power Point slide (208K)

Discussion of diagnosis

Neurological paraneoplastic syndromes incorporate a broad range of disorders thought to result from 'collateral damage' to the nervous system when the immune response to neoplastic antigen(s) results in cross-reactivity with antigen(s) that are incidentally expressed by healthy neural tissue. The primary neoplasm—typically malignant—is often small and difficult to identify, possibly because it is contained by the body with an aggressive immune response. Notably, the paraneoplastic syndrome often appears months or even years before the underlying malignancy is identified.^{1, 2} Extreme variability in clinical and laboratory manifestations of these protean syndromes has forced development of uniform diagnostic standards.³

The clinical syndrome associated with anti-Ma2 antibodies was reported in the late 1990s by Voltz and colleagues.⁴ The Ma2 gene encoding the target antigen is a member of the paraneoplastic 'Ma' (PNMA) family of onconeuronal proteins,^{5, 6} and is expressed by both testicular cancer cells and normal brain tissue⁴—particularly in limbic and brainstem regions.⁵ Anti-Ma2-associated encephalitis (anti-Ma2AE) has been identified in association with many different neoplasms, the most frequent of which is germ cell tumor of the testis.^{5, 7} Anti-Ma2AE is difficult to diagnose, and a testicular malignancy is often overlooked. In contrast to other paraneoplastic syndromes that usually evolve over fewer than 8 weeks, patients with anti-Ma2AE have a median time from symptom onset to stabilization or death of 6 months (range 5 days to 16 months).⁷ The protracted time course and confusing constellation of findings (see Table 1) frequently lead to diagnostic delays, as seen in this case.

Table 1 Neurological spectrum in patients with anti-Ma2-associated encephalitis (anti-Ma2AE) and CNS Whipple's disease, by presumed brain region affected.^{7, 13, 20}

Full tableFigures & Tables indexDownload Power Point slide (228K)

Routine diagnostic tests will often reveal normal results, or nonspecific abnormal results. In 50% of cases, a lumbar puncture will reveal a pleocytosis,⁷ and in 63% of cases there will be nonspecific markers of inflammation or altered protein synthesis (elevated protein or oligoclonal bands).⁷ In 74% of cases, brain MRI scans are abnormal, and often demonstrate T2/FLAIR hyperintensities in the limbic and brainstem regions that will occasionally be contrast-enhancing.⁷ The standard criterion for diagnosis is the identification of anti-Ma2 antibodies in either the serum or CSF in the appropriate clinical context. Of note, commercial paraneoplastic panel results may take weeks or months to return—in this case, the final results were sent from the laboratory 13 weeks after the specimen was drawn, well after the patient's discharge from the hospital.

In the present case, a postmortem histological examination revealed a variety of changes largely in the upper brainstem, basal forebrain, and amygdala. The patient's vertical gaze palsy and dominantly vertical and torsional intrusive eye movement disorder are attributable to involvement of the midbrain tegmentum and substantia nigra pars reticularis. The profound parkinsonism may have been the result of new disease superimposed upon old—encephalitic involvement of the substantia nigra pars compacta and basal forebrain in the setting of prior frontal lobe traumatic brain injury. The progressive anxiety associated with the syndrome could be explained by involvement of the amygdala.

Of particular interest in the present case were the negative abdominal and pelvic CT and PET-CT scans performed during his work-up. It is not surprising that routine CT imaging might miss a small malignancy, but PET scans can detect malignancies as small as 4–8 mm in diameter,⁸ and PET-CT has been shown to have a sensitivity of at least 83% in patients with paraneoplastic neurological syndromes and negative or inconclusive conventional imaging results.⁹ Although PET-CT is more sensitive than conventional CT for the identification of small malignant lesions,¹⁰ it is clear from this case that it cannot be used as the sole and final arbiter for the absence of malignancy, even in the setting of profound, clinically devastating neurological disease. In particular, with respect to detection of testicular malignancies, baseline high levels of metabolic activity in the testicles make lesion visualization difficult on FDG-PET.¹¹ In our patient, PET-CT scanning was unable to detect the 1 mm tumor ultimately deemed responsible for the anti-Ma2-associated syndrome. In patients with otherwise unexplained, subacute CNS deterioration, paraneoplastic disease should be given consideration even in the face of a negative PET scan. Testicular ultrasound, in large part because of its ability to visualize the thick, fibrous tissue surrounding the tumor, was the diagnostic test of choice here.

Differential diagnosis

The patient in this case presented with a subacute neurological decline characterized by symptoms that were suggestive of a prominent midbrain involvement (i.e. vertical gaze palsy and atypical parkinsonism). Although toxic disease (e.g. manganese toxicity), metabolic conditions such as Wilson's disease and mitochondrial disease, and degenerative diseases such as progressive supranuclear palsy were all theoretical considerations, the presence of MRI signal changes and an elevated CSF protein level suggested the presence of an encephalitic process. The predominant etiologies for consideration were an atypical bacterial cause (listeriosis, Lyme disease, syphilis or Whipple's disease), a viral condition (herpes simplex), prion disease (Creutzfeldt–Jakob disease), and immune-mediated causes (systemic lupus, paraneoplasia). Of these possibilities, the subacute progression lasting a period of months and the prominent intrusive eye movement disorder indicated that a paraneoplastic syndrome, Whipple's disease or Creutzfeldt–Jakob disease were the leading considerations.

CNS Whipple's disease is of particular interest, because it can mimic anti-Ma2AE. The condition is caused by the bacterium Tropheryma whippelii, and it classically presents with fever, polyarthritis, weight loss, and diarrhea, although other organs are frequently involved.¹² In almost 25% of cases, however, a CNS involvement is clinically isolated and prominent neurological manifestations are seen.¹³ Both Whipple's disease and anti-Ma2AE affect middle-aged men,^{7, 13} and as demonstrated in Table 1, they produce a similar spectrum of neurological symptoms and signs. They are also both associated with nonspecific brain MRI changes in the majority of cases.^{5, 7, 13} In one published series of patients with anti-Ma2AE,⁷ 16% (6/38) had undergone duodenal biopsy for suspected CNS Whipple's disease before being correctly diagnosed (J Dalmau, unpublished data).

Whipple's disease can be diagnosed by a histologic examination of endoscopically biopsied intestinal mucosa with a sensitivity of 75% (a percentage that can be slightly improved with the use of electron microscopy).¹³ In cases involving isolated neurological manifestations, however, the sensitivity of light microscopy on intestinal mucosa is reduced to 0%.¹³ In such cases, PCR detection of the bacterial 16S ribosomal RNA in intestinal specimens can still be of diagnostic relevance,¹⁴ and the involvement of lymph nodes as identified by abdominal CT scan can provide an additional source of diagnostic information.¹³ If there are no extracranial manifestations, a brain or meningeal biopsy should ultimately be pursued, but even this test can produce false negatives.¹² Although PCR analysis for Tropheryma whippelii RNA is thought to have a high sensitivity for detecting the disease¹⁵ and it can now be conducted on biopsy specimens, CSF, serum, and even feces,¹⁶ its specificity has been called into question,¹⁷ and large cohorts with isolated CNS disease have not been studied using these methods.

In the present case, the patient's history of fever and diarrhea and the negative initial investigations for underlying malignancy (including FDG-PET-CT) were misleading, as was the apparent response to antibiotic treatment, which, in retrospect, probably represented the natural progression of anti-Ma2AE. Although definitive bedside criteria for differentiating anti-Ma2AE from Whipple's disease would be desirable, this seems to be unachievable in most cases because of the substantial phenotypic overlap between the two syndromes (see Table 1). To date, oculomasticatory myorhythmia, found in only 8% of CNS Whipple's disease cases, remains the only neurological sign not to have been reported in both diseases.

Treatment and management

The cornerstone of treatment for paraneoplastic anti-Ma2AE remains the identification and eradication of the underlying malignancy. Effective cancer treatment can result in a reversal of neurological symptoms without the need for additional therapy, and it is associated with substantially improved outcomes compared with immunomodulatory therapy alone.^{5, 7} Anti-Ma2AE—unlike many other paraneoplastic syndromes—shows a reasonable potential for symptomatic improvement with treatment (35% across two case series, n = 39),^{5, 18} rather than merely a stabilization of symptoms. In cases where the paraneoplastic syndrome fails to respond to treatment of the primary malignancy, immunotherapies have been suggested,¹⁹ but none have proven highly effective. In the patient in this case study, the anti-Ma2AE progressed relentlessly, despite gross eradication of the cancer and combined immunomodulatory therapy with prednisone, mycophenolate mofetil, intravenous immunoglobulin, and plasmapheresis. Although these treatments may have slowed the disease progression, they certainly did not alter the final outcome.

Conclusion

Anti-Ma2AE is a rare condition, but it may manifest with symptoms that overlap with or mimic those of other neurological diseases—particularly CNS Whipple's disease. In order to be able to treat anti-Ma2AE at the earliest possible stage, it is essential to maintain a high level of suspicion and to pursue the diagnosis aggressively. During this process, it is important to bear in mind that commercial paraneoplastic panel results may take months to return, and that such delays increase the risk of communication gaps among providers, thereby increasing the risk of misdiagnosis. Furthermore, paraneoplastic diseases can be caused by tumors too small for detection by routine CT or even FDG-PET-CT imaging—a negative scan does not rule out the possibility of a small, occult malignancy. Testicular ultrasound is essential in male patients with possible anti-Ma2AE, even in the face of negative results from other studies.

Acknowledgments

The patient is thanked for allowing publication of the supplementary online material, for which his written consent was obtained.

References

1. Posner JB (1995) Neurologic Complications of Cancer. Philadelphia: FA Davis Company
2. Ahern G et al. (1994) Paraneoplastic temporal lobe epilepsy with testicular neoplasm and atypical amnesia. Neurology 44: 1270–1274 | PubMed | ChemPort |
3. Graus F et al. (2004) Recommended diagnostic criteria for paraneoplastic neurological syndromes. J Neurol Neurosurg Psychiatry 75: 1135–1140 | Article | PubMed | ISI | ChemPort |
4. Voltz R et al. (1999) A serologic marker of paraneoplastic limbic and brainstem encephalitis in patients with testicular cancer. New Engl J Med 340: 1788–1795 | Article | PubMed | ChemPort |
5. Rosenfeld M et al. (2001) Molecular and clinical diversity in paraneoplastic immunity to Ma proteins. Ann Neurol 50: 339–348 | Article | PubMed | ChemPort |
6. Schuller M et al. (2005) The human PNMA family: novel neuronal proteins implicated in paraneoplastic neurological disease. J Neuroimmunol 169: 172–176 | Article | PubMed | ChemPort |
7. Dalmau J et al. (2004) Clinical analysis of anti-Ma2 associated encephalitis. Brain 127: 1831–1844 | Article | PubMed | ISI |
8. Lowe V and Nauheim K (1998) Positron emission tomography in lung cancer. Ann Thorac Surg 65: 1821–1829 | Article | PubMed | ChemPort |
9. Younes-Mhenni S et al. (2004) FGD-PET improves tumour detection in patients with paraneoplastic neurological syndromes. Brain 127: 2331–2338 | Article | PubMed | ISI | ChemPort |
10. Wechalekar K et al. (2005) PET/CT in oncology—a major advance. Clin Radiol 60: 1143–1155 | Article | PubMed | ChemPort |
11. El-Haddad G et al. (2004) Normal variants in [¹⁸F]-fluorodeoxyglucose PET imaging. Radiol Clin North Am 42: 1063–1081 | PubMed |
12. James D and Lipman M (2002) Whipple's disease: a granulomatous masquerader. Clin Chest Med 23: 513–519 | Article | PubMed | ChemPort |
13. Gerard A et al. (2002) Neurologic presentation of Whipple disease: report of 12 cases and review of the literature. Medicine 81: 443–457 | Article | PubMed |
14. Lynch T et al. (1997) Polymerase chain reaction-based detection of Tropheryma whippelii in central nervous system Whipple's disease. Ann Neurol 42: 120–124 | Article | PubMed | ChemPort |
15. Ramzan N and Loftus E (1997) Diagnosis and monitoring of Whipple's disease by polymerase chain reaction. Ann Intern Med 126: 520–527 | PubMed | ChemPort |
16. Maibach RC et al. (2002) Detection of Tropheryma whipplei DNA in feces by PCR using a target capture method. J Clin Microbiol 40: 2466–2471 | Article | PubMed | ChemPort |
17. Ehrbar H et al. (1999) PCR-positive tests for Tropheryma whippelii in patients without Whipple's disease. Lancet 353: 2214 | Article | PubMed | ChemPort |
18. Gultekin S et al. (2000) Paraneoplastic limbic encephalitis: neurological symptoms, immunological findings, and tumor association in 50 patients. Brain 123: 1481–1494 | Article | PubMed | ISI |
19. Bataller L and Dalmau J (2003) Paraneoplastic neurologic syndromes. Neurol Clin 21: 221–247 | Article | PubMed | ISI |
20. Louis ED et al (1996) Diagnostic guidelines in central nervous system Whipple's disease. Ann Neurol 40: 561–568 | Article | PubMed | ChemPort |

Supplementary information

Supplementary Video 1 (mpg 29 MB)

Video showing complex ocular motility disorder seen in patient.

Video demonstrates an incomplete, mixed vertical and horizontal, supranuclear gaze palsy, with preferential restriction of upgaze, and relative sparing of pursuit, vergence, and vestibulo-ocular reflex movements. Static eyelid retraction and fluttering (myokymia) were present at rest, and exacerbated by attempted eye movements, particularly upward. At times, during such attempted movements, the upper lids appeared ptotic, rather than retracted. There were frequent, spontaneous, ocular intrusions—both saccadic and non-saccadic. The back-to-back saccadic intrusions were multivectorial (opsoclonus), but were generally low-amplitude and intermittent with the eyes at rest. During voluntary and reflexive eye movements (particularly with sustained convergence or upgaze), the amplitude and frequency of these intrusions would increase dramatically. Upper eyelid myokymia or ptosis was typically enhanced when the opsoclonic movements increased in frequency or amplitude. Occasionally the intrusive movements appeared convergent-divergent, particularly during upgaze elicited by the vertical vestibulo-ocular reflex. The non-saccadic, nystagmoid intrusions included slow movements (with or without a corrective quick phase) that were both non-rhythmic (torsional pendular) and semi-rhythmic (upward drift with downward correction). In general, a torsional pendular intrusion began with a slow movement towards the right shoulder (in both eyes, though more evident in the right eye), followed by a brief pause, followed by a slow return to primary position.

Contact the journal about this article

Subject areas under which this article appears: Neuro-oncology | Neuroimmunology and neuroinflammation