The diagnosis was reviewed and vasculitis was considered as a possibility. Azathioprine was withdrawn after 8 weeks due to thrombocytopenia and was replaced by mycophenylate with an initial course of pulsed intravenous methyl prednisolone 750 mg once weekly. The patient began to experience drenching night sweats, fevers, chills and fatigue. He found the first infusion of methyl prednisolone immediately very debilitating but within an hour of the second steroid infusion he started to develop a left VII nerve palsy leading to complete paralysis of the left side of the face within the next 2 days and was re-admitted to hospital for further investigations.

General examination was normal with no lymphadenopathy or hepatomegaly. Neurological examination revealed evidence of the following cranial nerve palsies: right IV, right Vii, left VII, left XII, as well as reduced sensation in the right ulnar nerve distribution. Cerebrospinal fluid (CSF) examination and repeat MRI did not reveal any abnormalities.

After 10 weeks of high-dose steroid treatment, serum ACE level had returned to normal. The CRP was slight raised at 15.5 with a normal PV 1.62.

Ten months after the initial presentation flow cytometry of the vitreous humour showed lymphomatous cells. Leucocytes from the vitreous specimen were isolated by ammonium chloride lysis. In view of the small volume of sample available, a limited four-colour flow cytometry panel was used and this identified a clonal B-cell population (CD19+, sIgK+, CD5–). The diagnosis of activated B-cell type diffuse large B-cell lymphoma was confirmed on the bone marrow sample taken 7 days later using a combination of four-colour flow cytometry supplemented by immunocytochemistry on the trephine biopsy. The composite phenotype was CD5–, CD10–, BCL2+,IRF-4+, BCL-6+/–, FOX P1+,CD38+ with a cycling fraction of 70% (figure 1). A staging CT scan revealed a 16 mm superior mediastinal mass as well as widespread nodal and extra nodal disease in the abdomen.

Peripheral nerve syndromes, including sensory, motor or sensorimotor peripheral neuropathy, mononeuritis, mononeuritis multiplex, plexopathies and radiculopathies, are unusual complications of lymphoma and other lymphoproliferative disorders where, in a small proportion of both B, and less often, T cell lymphomas, they can occur as the presenting complaint, sometimes preceding the underlying diagnosis by several years.³ Several different mechanisms leading to peripheral nerve syndromes have been demonstrated, the commonest being meningoradiculopathy (including cranial nerves) where abnormalities in either CSF or MRI are found in over 90%. Neurolymphomatosis is the commonest mechanism of purely peripheral nerve involvement where malignant cells propagate into the peripheral nervous system in a patchy fashion; this is easily missed on biopsy unless as long a segment as possible is taken from an involved sensory nerve, rather than a standard sural nerve biopsy with many cases proven only at autopsy. The predilection for certain parts of the peripheral nervous system in individual patients is not understood but it has been shown that some patients with T cell lymphoma produce particular adhesion cells that help tumour cells cross the blood–nerve barrier. Peripheral nerve damage can also result from dysimmune inflammation and haematogenous metastatic vascular occlusion as well as a result of anti-nerve monoclonal antibodies.⁴ Less commonly, vasculitis or cryoglobulinaemia can cause mononeuropathy or asymmetric neuropathy.⁴ There remains a small group of patients in whom the link between the peripheral neuropathy and the non-Hodgkin lymphoma remains unexplained, possibly secondary to an as yet unknown paraneoplastic mechanism.

In some patients, the peripheral nerve syndromes improve or disappear with treatment of the underlying cancer and may recur with relapse of the lymphoma. In one case of haematological remission, and in the absence of evidence of systemic or central nervous system lymphoma, neurological symptoms developed that were found, at autopsy, to be secondary to infiltration by tumour of the femoral nerves.³ The interesting point about the case we present is that the individual nerve palsies developed and subsequently improved while the patients’ treatment with high-dose oral steroids was unchanged. For example, the XII nerve palsy appeared after 5 weeks of high-dose oral steroids only to improve significantly overnight a few weeks later without any alteration in treatment. The IV nerve palsy then developed while on the high-dose steroids about the time azathioprine was introduced and improved without any changes in treatment. Although lymphoma is, to a degree, steroid responsive, the facial weakness developed immediately following the second pulse of high-dose intravenous methyl prednisolone.

There was no paraprotein in this case and, due to the particular nerves involved, no biopsy has been possible but the likely mechanism for nerve damage in this patient was vasculitis or vascular occlusion, but it is difficult to hypothesis as to why such lesions that have developed on steroids might improve again without increased immunosuppression or chemotherapy. The patient did not experience any intercurrent infections or other physiological stresses coincident with the fluctuation of the symptoms. The allodynia remains unexplained, but a small fibre neuropathy could explain the pain syndrome in the context of the normal nerve conduction studies, although the nerve conduction studies were done quite early in presentation and were not repeated.

Finally, the pathway to the final diagnosis also deserves some discussion. Initially, sarcoidosis was thought to be the most likely unifying diagnosis as it is the commonest cause of uveitis and is well-known to cause cranial nerve palsies, more often via granulomatous infiltration of the meninges than associated with vasculitic mononeuritis multiplex. In this case, no hard evidence for systemic sarcoidosis was found and the serum ACE that was initially raised turned out to be non-specific and misleading. Diagnosis of primary intraocular lymphoma is challenging but it is important to recognise it as early as possible. The difficulty lies in its rarity and that it presents similarly to several much commoner diseases that are not always easy to firmly diagnose in their own right as with sarcoidosis.¹ Therefore, it is important to keep lymphoma in mind when treating for a more likely alternative and to review the diagnosis early if the disease behaves atypically, as in this case, where the neurological symptoms in particular progressed in spite of treatment.