Skip to main content
BMJ Case Reports logoLink to BMJ Case Reports
. 2019 Nov 2;12(11):e230117. doi: 10.1136/bcr-2019-230117

Isolated leptomeningeal carcinomatosis and possible fungal meningitis as late sequelae of oesophageal adenocarcinoma

Richard Dumbill 1,✉, Sanja Thompson 2, Heiko Peschl 3, GDH Turner 4, Charles Woodrow 5
PMCID: PMC6827809  PMID: 31678919

Abstract

We describe a case of a 67-year-old man with known chronic obstructive pulmonary disease, type 2 diabetes mellitus, hypertension, osteoarthritis, previous history of excess alcohol intake, and oesophagectomy 3 years earlier for T3N0 adenocarcinoma, referred by his general practitioner with confusion, weight loss and several recent falls. CT of the chest, abdomen and pelvis revealed a right middle-lobe pulmonary embolism, while CT of the head revealed a communicating hydrocephalus. Lumbar puncture was performed, and empirical treatment for tuberculous and fungal meningitis was commenced. Unfortunately, he suffered a rapid neurological deterioration with markedly elevated cerebrospinal fluid (CSF) pressures, leading to an external ventricular drain. Cytological analysis of a CSF sample revealed a cellular infiltrate consistent with leptomeningeal carcinomatosis (adenocarcinoma), with the previous oesophageal malignancy the likely primary. He passed away 17 days after hospital admission. Prolonged culture of CSF later produced evidence of two distinct phaeomycotic moulds (Cladosporium sp and Exophiala sp), suggesting that fungal meningitis may also have contributed to the clinical picture.

Keywords: meningitis, oesophageal cancer, pathology, surgical oncology

Background

Oesophageal carcinoma is the 13th most common malignancy diagnosed in the UK, with around 9000 cases diagnosed each year.1 The 5-year survival rate is 15%.2 Surgical resection is appropriate for approximately one in five patients,3 and includes oesophagogastric resection, segmental oesophageal resection and total oesophagectomy, depending on the site of the tumour. Indications for primary surgical resection are T1 or T2 disease. Patients with T3 disease and selected patients with T4a disease may be offered resection after neoadjuvant chemoradiotherapy.

Neurological sequelae of oesophageal carcinoma are rare and may be due to cerebral or leptomeningeal metastases and paraneoplastic phenomena. Previous case reports illustrate the primary presentation of oesophageal carcinoma as classic meningitis,4 malignant meningitis causing sensorineural hearing loss,5 limbic encephalitis6 and confusion.7 Recurrence following treatment has also been described presenting with haemorrhagic brain metastasis.8 Here, we report the first case report of isolated leptomeningeal recurrence presenting as confusion, with rapid general deterioration and death, 3 years after apparently curative oesophagectomy for oesophageal adenocarcinoma.

Case presentation

A previously independent 67-year-old man presented to the hospital after a home visit by his general practitioner (GP), who was concerned about confusion, weight loss and several falls. The GP reported that his house was in a state of disrepair. His last contact with a healthcare professional had been 4 months previously with no cause for concern. His medical history was significant for well-controlled type 2 diabetes mellitus (as implied by haemoglobin A1c measurements from his GP up to 4 months prior to this presentation), hypertension, chronic obstructive pulmonary disease, osteoarthritis and oesophagectomy 3 years prior for adenocarcinoma of the gastro-oesophageal junction. Preoperative staging was T2N0M0, and he was treated with neoadjuvant cisplatin and capecitabine prior to a left thoracoabdominal oesophagectomy. Final pathological staging was T3N0M0.

On admission he was alert but confused, with an AMTS score of 3/10. His cardiovascular, respiratory and neurological examinations were normal. He was thin and unkempt. He had no close friends or family who were able to provide a collateral history. He admitted to alcohol intake between 10 and 20 units/week. His medications included lansoprazole 30 mg once daily, sitagliptin 100 mg once daily, amlodipine 10 mg once daily, insulatard 8–10 units once daily (in the morning), co-codamol, vitamin B12 supplements and inhaled salbutamol. His GP record indicated that his weight had been measured to be 77 kg 8 months prior to presentation, while in hospital, his estimated weight was around 65 kg. The initial impression was of acute delirium from urinary tract infection (UTI), likely superimposed on a chronic cognitive decline.

Investigations

Aside from confusion and cachexia, examination was unremarkable. The abnormal blood tests on admission were mild neutrophilia (5.58×1012/L), an international normalised ratio of 1.8 (PT 20 s), an iron level of 4.2 μmol/L, transferrin saturation of 8%, B12 level of 1392 ng/L, low serum folate level (<3 µg/L), alkaline phosphatase of 190 IU/L, C reactive protein level (CRP) of 11 mg/L and a blood glucose level of 17.8 mmol/L. Initial blood cultures were negative. He was treated with 10 mg oral vitamin K, folic acid replacement and standard hospital alcohol detox regime (high vitamin B replacement) and chlordiazepoxide as needed.

CT of the head revealed an enlargement of the supratentorial and infratentorial ventricular systems, with significant parenchymal volume loss with no transependymal cerebrospinal fluid (CSF) resorption and no acute injury (figure 1A).

Figure 1.

Figure 1

(A) CT head shortly after admission, reported as showing no acute intracranial pathology, with enlargement of the supratentorial and infratentorial ventricular systems, and no transependymal cerebrospinal fluid resorption. (B) MRI scanning showed ventriculomegaly (T2-weighted axial image). (C) MRI showing brainstem atrophy (T1-weighted sagittal section). (D) The CT scan (high resolution) immediately prior to external ventricular drain insertion, with global ischaemic changes and global mass effect on paramedial sagittal plane with the loss of grey-white matter differentiation and sulcal effacement. (E) The CT scan (high resolution) postdrain. Despite aggressive ventricular decompression by the drain, the changes shown in D had progressed on the subsequent postdrain CT.

CT of the chest, abdomen and pelvis revealed a right middle lobe pulmonary embolism, and he was therefore anticoagulated. He received empirical treatment for a UTI with a course of oral co-amoxiclav. Over the following few days, his cognition and mobility steadily declined, and intravenous acyclovir, amoxicillin and ceftriaxone were administered. An MRI scan (figure 1B,C) showed communicating hydrocephalus and brainstem atrophy, disproportionate to the patient’s age, raising the possibility of an underlying neurodegenerative disorder.

A lumbar puncture performed on day 8 of his admission revealed an opening pressure of 7 cmH2O, and CSF biochemistry showed glucose level of 3.7 mmol/L (paired serum sample 12.3 mmol/L), lactate level of 11 mmol/L and a protein concentration of 1197 mg/dL. Microscopy revealed a CSF white blood cell count of 34×106/L (lymphocytes 30×106/L). A BioFire meningitis/encephalitis panel run on the CSF was negative, and antibiotics and acyclovir were discontinued. The CSF sample sent for cytology was delayed in reaching the laboratory due to problems with labelling, which may have contributed to its degradation, showing only debris and poorly preserved cells, some of which were identified as small lymphocytes.

Clinical examination after lumbar puncture showed transient but significant improvement; he had progressed to being able to sit up and communicate in short sentences. He was reviewed by a neurologist, who suggested performing further laboratory analyses, including the Venereal Disease Research Laboratory test for syphilis, anti-nuclear antibodies, immunoglobulins, serum electrophoresis, beta-2-microglobulin and paraneoplastic antibodies. We were also advised to arrange electroencephalogram, CT-positron emission tomography, and, after an interval, repeat MRI of the brain and spine with gadolinium.

An autoimmune panel (run on serum and CSF) was negative for voltage-gated potassium channel antibodies, fixed-cell NDMA receptor antibodies, and Hu, Yo, and Ri antibodies. Serum antibodies to HIV, syphilis, toxoplasma and Borrelia were negative, and galactomannan was negative. The serum beta-d-glucan antigen level was above the normal range at 230 pg/mL.

This clinical improvement persisted for approximately 5 days and was followed by general deterioration, with progressively impaired mobility. A lumbar puncture was repeated on day 15, revealing hypercellular CSF containing large, atypical cells with cytoplasmic blebbing, vacuolation and basophilia, and once again high protein and low glucose. The beta-d-glucan concentration in the second CSF sample was elevated at 346 pg/mL. Empirical treatment for possible fungal and tuberculous (TB) meningitis was commenced, consisting of amphotericin, rifampicin, isoniazid, pyrazinamide, pyridoxine and ethambutol.

On day 16, his Glasgow Coma Scale score suddenly deteriorated to 4, and he was intubated and transferred to the intensive care unit. Repeat CT of the head showed a global mass effect on the paramedial sagittal plane with the loss of grey-white matter differentiation and sulcal effacement (figure 1D). On day 17, due to further deterioration and the development of sluggish pupils bilaterally, a third lumbar puncture was performed, which showed an opening pressure of 37 cmH2O. Following this, an external ventricular drain with intracranial pressure monitor was inserted, and a third CT of the head (figure 1E) following this showed progressive loss of grey-white differentiation consistent with extensive ischaemic/hypoxic brain injury.

At this point, cytology results from the second CSF sample became available. The CSF cytospin showed small mature lymphocytes and a variably sized population of atypical plasmacytoid cells with some nuclear enlargement, multiple nuclei and focal marked cytoplasmic vacuolation. Some perinuclear hoffs and a deep cyanophilic cytoplasm were present. Flow analysis initially raised the possibility of a plasma cell malignancy; however, on immunostaining, the enlarged atypical cells showed variable positive staining for pan-cytokeratin and CK7, in keeping with an epithelial adenocarcinoma. They were negative for lymphocytic markers CD3, CD20 and CD79a, and plasma cell markers, including MUM-1, cyclin D1 and CD56. No light chain staining or restriction was demonstrated, in keeping with the flow result. The features were felt to be those of a metastatic adenocarcinoma, which, despite their plasmacytoid morphology and CD138 positivity on flow cytometry, were felt likely to represent metastatic spread from the oesophageal primary. He continued to decline and passed away later that day.

Prolonged culture of the two CSF samples later yielded evidence of two distinct phaeomycotic moulds. Cladosporium sp (confirmed by molecular identification) was cultured at the reference laboratory from the first sample. Exophiala sp was identified by molecular methods from the second sample but failed to grow on subculture.

Differential diagnosis

The differential diagnosis changed as the case evolved and included UTI, normal pressure hydrocephalus, an underlying neurodegenerative disorder and intracranial infection. Early in the course of the admission, the initial diagnosis of UTI was superseded by normal pressure hydrocephalus or a neurodegenerative process; however, his rapid deterioration was out of keeping with this. The lymphocytic nature of the CSF in combination with the appearance of hydrocephalus on brain imaging raised the possibility of a malignant meningitis, a lymphomatous meningitis or a chronic infection such as TB or fungal meningitis. The diagnosis was complicated by the isolated nature of the pathology and a transient improvement after initial lumbar puncture and treatment for a possible UTI. The lack of lymphadenopathy on imaging initially suggested a pathology other than malignant or TB meningitis. CSF cytology eventually yielded the diagnosis, but only after repeated lumbar punctures, later in the course of his admission.

Treatment

His treatment initially involved antibiotics for UTI, but his failure to improve, with further deterioration of cognition and mobility, initiated further tests, including brain imaging and lumbar puncture. The initial lumbar puncture revealed a normal CSF pressure, which contributed to a provisional diagnosis at this point of normal pressure hydrocephalus; therefore, earlier external ventricular drainage was not pursued. Treatment thereafter was guided by the test results; specifically, anti-TB therapy was started due to clinical deterioration, the imaging findings and a significantly lymphocytic CSF. Liposomal amphotericin was started due to the high beta-d-glucan levels in serum and CSF. Neurosurgical intervention with an extraventricular drain was eventually performed prior to definitive CSF cytology results becoming available, in response to rising CSF pressures and neurological deterioration.

Outcome and follow-up

The patient unfortunately passed away on day 17 of his admission, after 2 days in intensive care.

Discussion

Our patient had evidence of both carcinomatous and fungal meningitis, both of which can cause chronic cognitive decline with more rapid deterioration associated with ischaemia and communicating hydrocephalus. Carcinomatous meningitis is a rare cause of delirium, and recurrence of a solid malignancy as isolated leptomeningeal metastases not visible on imaging is rarer still. Neurological complications resulting from leptomeningeal carcinomatosis secondary to a primary oesophageal cancer have previously been described,4 5 7 but this is the first report of a recurrence presenting in this fashion. A possible mechanism for the development of leptomeningeal metastases without nodal disease could be dural injury and seeding at the time of resection; however, the description of previous similar cases, particularly that described by Lobo et al, which presented with leptomeningeal seeding without any surgical intervention,7 raises the possibility that this could be a natural, rather than iatrogenic, phenomenon. Of note, in the case described by Lobo et al, lesser curvature, coeliac, and para-aortic lymphadenopathy was noted on CT scanning; in our case, no lymphadenopathy was found.

No fungal agent was definitively identified antemortem, but there were high serum and CSF beta-d-glucan levels. In a study of CSF samples relating to a US outbreak of fungal meningitis associated with injection of contaminated methylprednisolone acetate, a CSF beta-d-glucan cut-off value of 138 pg/mL provided 100% sensitivity and 98% specificity9; our patient had a level well above this threshold. Prolonged CSF culture and molecular analysis later provided evidence of potential phaeomycotic mould infection involving Cladosporium sp and/or Exophiala sp; both have been associated with cerebral abscess and meningitis in patients with immunosuppression, and our patient had several factors causing immunosuppression: malignancy, diabetes and severe weight loss. However, both of these are environmental organisms, and hence there is a possibility that these findings represented culture contamination. It is already known that certain medical conditions and drugs can produce a false-positive serum beta-d-glucan; could the same apply to CSF examination? Evidence on this is currently lacking but likely to grow substantially as this test is increasingly applied. Overall, it is not possible to exclude the possibility that fungal meningitis contributed to the patient’s clinical state, in addition to malignancy.

Learning points.

Oncological

  • Neoplastic meningitis is suggested by clinical presentation consistent with meningitis or encephalitis, in conjunction with cerebrospinal fluid (CSF) showing a high lymphocyte count, high protein and low glucose level, and negative microbiological investigations.7

  • A history of malignancy, particularly of large-cell lymphoma, gastrointestinal, breast, lung, or melanoma, increases the likelihood.

  • Iatrogenic seeding could be considered as a possible mechanism when patterns of spread—particularly CSF infiltration with a lack of local or nodal recurrence—are atypical.

  • MRI with contrast should be organised for cases suspicious malignancy. CSF examination is a valuable diagnostic tool in patients with suspected CNS malignancies, but due to its low sensitivity (initial sample positive in 45%–94% of cases in previously reported series), negative results should not be ignored, and repeat sampling should be considered.10 Likewise, its rapid transport to the laboratory and examination as soon as possible must be secured, in order to increase the chance of successful diagnosis, as CSF specimens tend to degenerate rapidly.11

Microbiological

  • With the ever-increasing availability of sensitive microbiological diagnostics, it is inevitable that infectious agents will be implicated in an increasing proportion of chronic meningitis cases, so that in cases such as this one, there may be laboratory evidence of more than one aetiology.

  • It seems clear that measurement of the beta-d-glucan marker in CSF samples should be part of the standard workup in cases of chronic meningitis, and empirical antifungal treatment should be given to cases with high levels while prolonged cultures are undertaken.

  • Future research could help to determine whether beta-d-glucan can be falsely elevated in the presence of other conditions, such as malignant meningitis.

Footnotes

Contributors: RD conceived and coordinated the project, and wrote the report. ST provided interpretation of the clinical data and redrafted the report. HP acquired the radiological images, provided an interpretation and wrote this part of the report. GT performed the cytological analysis and wrote this part of the report. CW provided an interpretation of the microbiological data, including acquiring a Public Health England report on the mycology, and wrote the microbiological aspects of the report.

Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

Competing interests: None declared.

Patient consent for publication: Next of kin consent obtained.

Provenance and peer review: Not commissioned; externally peer reviewed.

References


Articles from BMJ Case Reports are provided here courtesy of BMJ Publishing Group

RESOURCES