To discuss two cases of myelitis associated with transiently detectable anti-CRMP-5 autoantibodies in patients without detectable malignancy.

Collapsin-responsive mediator protein 5 (CRMP-5), a cytoplasmic regulator of neurite outgrowth expressed throughout the nervous system, can be a target of paraneoplastic autoimmune neurologic processes, which can manifest as myelitis. Presence of anti-CRMP-5 autoantibodies is strongly associated with underlying malignancy. The significance of transient elevation in CRMP-5 antibody in patients with neurologic syndromes but no malignancy is not known.

Case 1: A 44-year-old male presenting with subacute transverse myelitis and a longitudinally extensive thoracic cord lesion was found to have anti-CRMP-5 autoantibody in serum (Western blot, Mayo Clinic Laboratory). Subsequent anti-CRMP-5 autoantibody testing was negative, and three whole-body PET/CT scans failed to disclose malignancy during the three-year follow-up.

Case 2: A 65-year-old female presented with subacute myelitis associated with cervical and thoracic spinal lesions and subsequently developed optic neuritis and cranial neuropathies. Anti-CRMP-5 antibody was present in serum (Western blot, Mayo Clinic Laboratory), but not in CSF. Subsequent anti-CRMP-5 antibody tests and malignancy workup were negative. The patient demonstrated spontaneous clinical and radiographic improvement over three years of follow-up.

CRMP-5 autoantibody is associated with paraneoplastic neurologic syndromes but can be found in patients with malignancy and no neurologic disease and, rarely, in healthy people. Although our two patients with myelitis and detectable CRMP-5 autoantibody in sera meet the criteria of CRMP-5 Ab-associated paraneoplastic syndrome, they exhibited a large number of atypical features: absence of CRMP-5 Ab on immunofluorescence; negative CRMP-5 antibody on subsequent testing; no malignancy; atypical disease course with spontaneous improvement. The possibility of CRMP5-Ab being a false positive or a ‘true but unrelated’ finding must also be considered. Our cases illustrate the challenges of interpreting paraneoplastic antibodies that are not well-addressed in the current literature.