Background

The frequency of leptomeningeal metastatic spread of melanoma in stage IV patients is 22–46% [1]. Autoimmune neurologic events are a recognized complication of Ipilimumab therapy. Melanoma is associated with paraneoplastic neurologic syndromes. Determining the etiology of neurological complications in a patient with melanoma may thus present a challenge. It is important that the correct diagnosis is made because treatments are different and increasingly effective. We present a patient who developed inflammatory demyelinating polyneuropathy while taking Ipilimumab along with Vemurafenib for systemic disease, who was successfully treated with corticosteroids and discontinuation of Ipilimumab. Interestingly, he later developed clear leptomeningeal disease as part of his overall disease progression.

Case presentation

A 42 year old Caucasian male underwent wide local excision of a truncal 3.05 mm non-ulcerated melanoma with nevoid and spitzoid features. Sentinel node biopsy was negative. Two years later, a palpable left groin nodule was treated with superficial inguinal and deep pelvic lymphadenectomy followed by Interferon therapy. Interferon was tolerated poorly, with headaches and confusion, and stopped in the first month of therapy. Two and a half years later, PET/CT showed diffuse metastatic disease to the lungs, liver, right adrenal, and hilar lymph nodes. He was enrolled in a study evaluating 6 weeks of Vemurafenib followed by Ipilimumab 10 mg/kg (higher than the FDA approved dose of 3 mg/kg) intravenously every 3 weeks (NCT01673854) [2]. After 6 weeks of Vemurafenib, tumors had shrunk by 9%. Ipilimumab was initiated 8 weeks after his first dose of Vemurafenib. After the second dose of Ipilimumab, he developed Grade 1 fatigue and began having intermittent mild headaches attributed to stress at work. MRI of the brain was negative.

Six days following the third dose of Ipilimumab, he presented with bilateral hip pain 8 on a 10 point scale, subjective bilateral thigh weakness, and paresthesia of the soles of his feet. Initial examination showed normal muscle strength except for the left quadriceps (4/5). Upper extremity strength was intact. Sensory exam was not performed. MRI of the lumbar spine demonstrated diffuse enhancement surrounding the entire conus and all of the nerve roots thought most compatible with leptomeningeal carcinomatosis (Fig. 1). The remainder of the spine was not imaged because of the patient’s pain and absence of signs/symptoms suggestive of involvement above the lumbar spine. PET/CT performed the following day showed 23% increase in extra-CNS tumor measurements. Vemurafenib was resumed because of worsening of disease on Ipilimumab and previous improvement, although slight, on Vemurafenib for 6 weeks.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1

Initial MRI lumbar spine following Ipilimumab. Sagittal and axial T1w fat (left and top right) saturated post contrast and axial T2w FSE (bottom right) images demonstrating smooth, non-nodular avid enhancement of cauda equina without significant nerve root thickening

Patient’s lower extremity weakness, pain, and numbness worsened rapidly over the next week with deterioration to ECOG 3. There were still no cranial nerve nor upper extremity signs nor symptoms. Vemurafenib was held, dexamethasone was started, and daily radiation was initiated to conus and cauda equina at 3 Gy per fraction for presumed leptomeningeal disease. However, completion MRI of the spine done after 5 days of radiation, and 16 days after the initial MRI of the lumbar spine, revealed diffuse, symmetric, and smooth nerve root enhancement affecting the entire spine while sparing the entirety of the cord surface; imaging findings more compatible with inflammatory demyelinating polyneuropathy (Fig. 2). Brain MRI was normal. Radiation was stopped, neurology was consulted, and workup for inflammatory demyelinating polyneuropathy was initiated. CSF was negative for malignant cells, but showed significantly increased protein without pleocytosis (Table 1). Autoimmune testing was unremarkable (Table 1). Electrophysiological studies were done 8 weeks after the onset of the first neurological complaints, and at that point, his neurological exam showed weakness in foot dorsiflexion, knee and hip flexion bilaterally, worse on the left. He had impaired vibration in the toes and absent delayed tendon reflexes in the lower extremities.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2

Follow up post contrast cervical spine MRI. Parasagittal and axial T1w fat saturated post contrast images demonstrating smooth avid enhancement of all anterior and posterior cervical nerve roots (arrows) but no involvement of the cord surface. Both factors strongly mitigate against leptomeningeal metastatic disease

NCS were normal in the upper extremities despite abnormal cervical spine MRI but showed multiple abnormalities in the lower extremities including absent sensory response in the right sural nerve, markedly slowed conduction velocities between the ankle and below the fibular head stimulation, severely reduced compound muscle action potential (CMAP) amplitudes, and mild-to-moderate prolongation of the distal latency on the left side. Left tibial nerve motor NCS showed significantly prolonged distal latency with moderately decreased CMAP amplitudes and slowed conduction velocity. Right tibial nerve motor NCS was normal except for a mild prolongation of distal latency and mildly slowed conduction velocity. The F-wave responses of the bilateral tibial nerves were absent, indicating a more proximal conduction block. Needle electromyography of the left lower extremity was significant for denervation changes in lumbar paraspinal and tibialis anterior muscle as well as neurogenic changes in all the tested muscles of the right leg. The study was indicative of an asymmetric, subacute to early chronic and ongoing lumbar polyradiculoneuropathy with axonal involvement and demyelinating features. It is not clear why the cervical and thoracic spine findings did not cause detectable signs nor symptoms.

Vemurafenib was restarted and Dexamethasone was continued. Patient had partial response systemically and neurologic improvement. Dexamethasone was tapered and then stopped almost 6 months after initiation. It is not possible to evaluate the extent to which the dexamethasone or anti-melanoma agents contributed to this outcome.

PET/CT 6.5 months after resuming Vemurafenib showed progression of disease. Brain MRI showed multiple new foci of nodular leptomeningeal enhancement consistent with metastases. CSF cytology was concerning for metastases (Table 1). Vemurafenib was discontinued, and intrathecal IL2 and Dabrafenib therapy was initiated at another institution. Head CT after 5 months of intrathecal IL2 showed progression with the disease now predominantly dura-based, and patient elected for home hospice.