BI05 Cemiplimab toes: a new entity in the expanding spectrum of immune checkpoint inhibitor skin toxicities

Abstract Cemiplimab is a highly selective humanized monoclonal antibody targeting programmed cell death protein-1 (PD-1). It is approved for locally advanced and metastatic cutaneous squamous cell carcinoma (cSCC). It is associated with a wide range of cutaneous immune-related adverse effects (irAEs) including immunobullous, lichenoid and eczematous reactions. We present an unusual manifestation of cemiplimab skin toxicity undescribed previously – ‘cemiplimab toes’. A 47-year-old Indonesian woman started cemiplimab in October 2023 for primary cSCC in the external auditory meatus, which metastasized to parotid and cervical lymph nodes and bone. Despite extensive surgery and radiotherapy the tumour recurred locally and was inoperable. Within two to three cycles of cemiplimab, she noticed itching of her toes, which recurred with subsequent cycles of cemiplimab. However, by cycle 6, she reported the onset of marked periungual redness, swelling, itching and pain in all 10 toes, making mobilizing difficult. As pain was severe, the initial concern was of a vasculitic process and oncology started prednisolone 2 mg kg−1 per day (120 mg per day) and flucloxacillin, with symptoms improving rapidly. On dermatology review, all 10 toes demonstrated intense periungual redness, possible vesicles, and erosions gradually evolving into a more violaceous appearance with hyperkeratosis, paronychia-type changes and fissuring. Raynaud symptoms were absent, the fingers were unaffected, there were no rashes elsewhere and there was no mucosal involvement. Clinically it was difficult to distinguish between eczematous, psoriasiform, lichenoid, vasculitic, perniotic, autoimmune bullous, ischaemic and paraneoplastic processes. Bacterial swabs grew Staphylococcus aureus, and viral swabs and mycology were negative. Indirect immunofluorescence and autoimmune and vasculitis screen (including antinuclear antibodies, extended myositis panel, creatine phospho­kinase, complement, cryoglobulins and thrombophilia screen) were normal or negative. Incisional skin biopsy showed a predominantly spongiotic dermatitis with discrete epidermal psoriasiform change and no evidence of vasculitis, interface or lichenoid dermatitis or bullous processes. Direct immunofluorescence was negative. Although she responded well to oral steroids and suspending cemiplimab, her skin flared with subsequent cycles with prednisolone doses below 10 mg per day despite topical treatment with clobetasol propionate, tacrolimus and calcipotriol–betamethasone ointments, fludroxycortide tape, Fucibet cream and potassium permanganate soaks. Treatment that will allow continuation of cemiplimab while avoiding ongoing immunosuppression remains challenging. This dramatic localized skin irAE to a PD-1 checkpoint inhibitor involving only the toes has not been described previously. Clinically some features were reminiscent of ‘COVID toes’, but the underlying mechanism is unknown. This case of ‘cemiplimab toes’ is a striking reminder of the ever-evolving spectrum of cutaneous irAEs associated with immunotherapy and the challenges in diagnosing and managing these effectively to mitigate morbidity while continuing oncological treatment.