Eruptive Keratoacanthomas Associated With Pembrolizumab Therapy

Key Points

Question

Has pembrolizumab therapy ever been associated with the occurrence of eruptive keratoacanthomas (KAs)?

Findings

In this case report study of 3 patients, pembrolizumab therapy was associated with eruptive KAs. An enhanced immune response (observed on biopsy specimens) in the setting of preexisting and unidentified mutations may have induced the KAs in these cases; all were treated successfully, and all KAs resolved.

Meaning

Eruptive KAs associated with pembrolizumab therapy could be treated with noninvasive dermatologic therapy, allowing patients to continue treatment with pembrolizumab.

Abstract

Importance

To our knowledge, there have been no previous reports of eruptive keratoacanthomas (KAs) in patients receiving pembrolizumab.

Objective

To report the cases of 3 consecutive patients with pembrolizumab-induced eruptive KAs and their management.

Design, Setting, and Participants

Case report study of 3 patients from 2 centers with pembrolizumab-treated cancer who all developed eruptive KAs.

Interventions

All 3 patients had AK treatment with clobetasol ointment and intralesional triamcinolone; 2 patients also underwent open superficial cryosurgery.

Results

Three consecutive patients with cancer, 2 men and 1 woman (median age, 83 years; range 77-91 years), experienced pembrolizumab-associated eruptive KAs. All patients presented with a sudden onset of multiple lesions on sun-exposed areas of their extremities after a median of 13 months (range, 4-18 months) of pembrolizumab therapy. On lesional biopsy, a lichenoid infiltrate was observed in the underlying dermis, predominantly composed of CD3⁺ T cells, scattered CD20⁺ B cells, and relatively few PD-1⁺ (programmed cell death 1–positive) T cells, an immunophenotypic pattern also observed in other cases of anti–PD-1–induced lichenoid dermatitis. Patients were treated with clobetasol ointment and intralesional triamcinolone, alone or in combination with open superficial cryosurgery. All KAs resolved in all patients, and no new lesions occurred during close follow-up. Pembrolizumab treatment was continued without disruption in all 3 cases, and all patients had complete responses of their primary cancers.

Conclusions and Relevance

Pembrolizumab is used in advanced melanoma, advanced non–small-cell lung cancer, and in head and neck cancer. A variety of dermatologic immune-related adverse events including maculopapular eruption, lichenoid reactions, pruritus, and vitiligo have been described. This case series demonstrates that pembrolizumab therapy may also be associated with eruptive KAs with characteristic dermal inflammation, which improved with corticosteroid treatment (topical and intralesional) alone or in combination with cryosurgery, allowing patients to continue therapy with pembrolizumab.

This case report describes 3 patients who developed eruptive keratoacanthomas in association with pembrolizumab treatment and their management.

Introduction

The typical clinical course of keratoacanthoma (KA) is characterized by a cutaneous solitary lesion with rapid initial growth followed by a variable period of stability and occasionally spontaneous resolution. Eruption of multiple KAs has been described in different syndromes, including generalized eruptive KA of Grzybowski and Muir-Torre syndrome. Systemic medications such as leflunomide, transforming growth factor beta inhibitors, vismodegib, and BRAF inhibitors have been associated with this eruptive neoplasm.

Pembrolizumab is a monoclonal antibody directed against the PD-1 protein (programmed cell death 1), stimulating T-cell responses. This is mainly used for the treatment of melanoma, non–small-cell lung cancer that expresses PD-1, and head and neck squamous cell carcinoma (SCC). Pembrolizumab has been associated with a variety of dermatologic adverse events (dAEs) including maculopapular eruption, lichenoid reactions, eczema, and vitiligo. These immune-related dAEs in patients receiving anti–PD-1 therapy uncommonly result in treatment interruption, with a calculated high-grade incidence of less than 3%, but the reactions lead to clinically significant impact on patient morbidity. The development of dAEs in patients receiving anti–PD-1 therapy might be a positive prognostic factor. To our knowledge, there have been no previous reports of anti–CTLA-4 and anti–PD-1-induced eruptive KAs to date. We report the cases of 3 consecutive patients with pembrolizumab-associated eruptive KAs, including dermatologic therapeutic approach and follow-up.

Report of Cases

Three consecutive patients with pembrolizumab-associated eruptive KAs were identified, 2 men and 1 woman (median age, 83 years; range, 77-91 years). Two had previously experienced biopsy-proven lichenoid eruptions while receiving only pembrolizumab; these eruptions were successfully treated with clobetasol ointment prior to the appearance of KAs. One patient was diagnosed with metastatic SCC and was previously treated with cetuximab, methotrexate, and radiotherapy. The other 2 patients had metastatic melanoma, and 1 of these was treated for 2 months with ipilimumab and vemurafenib 6 months prior to beginning pembrolizumab therapy (Table).

Table. Clinical and Demographic Characteristics of 3 Patients Who Developed Eruptive Keratoacanthomas While Receiving Pembrolizumab.

Patient No./Sex	Underlying Cancer	Prior Therapy	Pembrolizumab Dose, Duration of Therapy	Anatomic Distribution of KA	KA Treatment (Response)	Other dAE
1/M	MM	Vemurafenib, ipilimumab	2 mg/kg, 18 mo	Extremities	Clobetasol ointment, IL-Tr, cryosurgery (CR)	Lichenoid skin reaction (developed 11 mo after pembrolizumab initiation)
2/F	MM	NA	2 mg/kg, 13 mo	Legs	Clobetasol ointment, IL-Tr, cryosurgery (CR)	NA
3/M	mSCC (primary scalp SCC)	Cetuximab, methotrexate, radiotherapy	2 mg/kg, 4 mo	Dorsal hands	Clobetasol ointment, IL-Tr (CR)	Lichenoid skin reaction (developed 2 mo after pembrolizumab initiation)

Abbreviations: CR, complete response; dAE, dermatologic adverse event; KA, keratoacanthomas; IL-Tr, intralesional triamcinolone; MM, metastatic melanoma; mSCC, metastatic squamous cell carcinoma; NA, not applicable; SCC, squamous cell carcinoma.

All 3 patients presented with a sudden onset of multiple lesions on sun-exposed areas of their extremities after a median of 13 months (range, 4-18 months) of pembrolizumab therapy. Clinical examination showed multiple tender erythematous, hyperkeratotic nodules, papules, and plaques on the extremities (Figure 1), consistent with KA.

Figure 1. Clinical Features Before and After Therapy With Clobetasol Ointment and Intralesional Triamcinolone Only (Patient 3).

A, Multiple erythematous, hyperkeratotic nodules and papules on dorsal hand surface. B, Therapy results 5 months after treatment.

All 3 patients also had biopsy results (Figure 2) consistent with KA, with lesional biopsies showing multiloculated, crateriform, keratin-filled lesions. The squamous cells had glassy-appearing cytoplasm with minimal cytologic atypia. A lichenoid infiltrate was observed in the underlying dermis of KA lesions in all patients, composed of CD3⁺ T cells with scattered CD20⁺ B cells. The T-cell infiltrate showed a normal CD4:CD8 ratio. Staining with PD-1 (antibody purchased from Leica Biosystems) showed weak positivity in rare lymphocytes. A diagnosis of hypertrophic lichen planus was considered; however, in each case, the cup-shaped architecture with glassy-appearing keratinocytes and sparse underlying infiltrate favored KA over hypertrophic lichen planus.

Figure 2. Hematoxylin-Eosin Biopsy Sections Showing a Multiloculated, Crateriform, Keratin-Filled Atypical Squamoproliferative Lesion.

A, Large keratinocytes with glassy-appearing, eosinophilic cytoplasm and minimal cytologic atypia. B, Higher magnification.

All of the patients’ KAs were treated with clobetasol ointment every other day until resolution and intralesional triamcinolone (20 mg/mL; 0.1 mL total to each nodule, repeated every 4 weeks until resolution). Patients 2 and 3 were also treated with open superficial cryosurgery (with liquid nitrogen 1.5 inches away from the lesion for 10 seconds). After 1 month of therapy, all patients had complete response (Figure 1), and at last follow-up all remained free of new lesions. Overall, the KA treatment was well tolerated and without any adverse event. Pembrolizumab therapy was continued in all the 3 patients, and a complete response of their primary cancers was observed (Table).

Discussion

Eruptive KAs are defined as multiple KAs that appear in a short period of time. In this series, we describe 3 patients undergoing pembrolizumab immunotherapy who developed sudden onset of multiple hyperkeratotic papules and plaques on sun-exposed skin with features consistent with KA on skin biopsy. Interestingly, skin biopsy in all 3 cases demonstrated a sparse lichenoid infiltrate in the underlying dermis of KAs. This infiltrate was predominantly composed of CD3⁺ T cells with scattered CD20⁺ B cells and relatively few PD-1⁺ T cells, an immunophenotypic pattern also observed in pembrolizumab-induced lichenoid dermatitis and in lichenoid mucocutaneous eruptions due to other anti–PD-1 therapies. Interestingly, in a previous report of dAEs of anti–PD-1 therapy, the rate of cutaneous SCC and actinic keratosis was similar to that expected for the age groups.

Given that pembrolizumab targets the PD-1 receptor, an immune mechanism may be responsible for the resulting dAEs. In a previous study, pembrolizumab therapy was associated with dAEs in 42% of patients. The most common dAEs associated with this therapy were maculopapular eruption, pruritus, and hypopigmentation. Most patients showed a favorable cutaneous safety profile and responded to topical corticosteroids, and survival analyses showed that patients who developed dAEs had significantly longer progression-free intervals regardless of the treatment regimen. Moreover, immune-related dAEs such as vitiligo and lichenoid reactions may be associated with durable clinical benefit and with improved survival in patients with melanoma treated with anti–PD-1 immunotherapy. The median time from initiation of pembrolizumab to onset of eruptive KA in the present patients was 13 months, which is consistent with the delayed timing to onset of other immune-related dAEs that have been reported with pembrolizumab.

Eruptive KAs pose a great challenge to treat owing to the large number of lesions and the morbidity associated with surgical and medical management; controversies remain about their management. Whenever possible, surgical treatment of a solitary KA is the standard regimen, with full-thickness excision providing both good aesthetic outcome and an optimal specimen for pathologic analysis. For eruptive KAs, especially those associated with syndromes, systemic acitretin or other retinoids are first-line therapy, either as monotherapy or combined with surgery. Several other therapies have been proposed, including cyclosporine, intralesional corticosteroids, and cryosurgery.

In the present patients, surgical excision was not a suitable option owing to the morbidity associated with excision of a large number of simultaneously erupting lesions. Systemic agents such as oral retinoid therapy posed a challenge owing to potential adverse effects, tolerability, and potential interactions in the setting of patients with metastatic cancer undergoing immunotherapies. These concerns, coupled with the inflammatory infiltrate seen on biopsy, led to conservative local immunosuppressive therapy with topical and intralesional steroids.

It is paradoxical that immunosuppressive corticosteroids can cause regression of KAs, since regression may be immune mediated, and KAs are probably more common in immunosuppressed hosts. However, intralesional and topical corticosteroids in combination with cryosurgery were used in 2 of the patients, and this treatment combination led to complete response after therapy. The third patient was treated only with intralesional and topical corticosteroids with similarly complete response. We might hypothesize that in pembrolizumab-induced KAs, an enhanced immune response (observed on biopsies) in the setting of preexisting and unidentified mutations, could induce the epithelial proliferation, and then corticosteroids would be expected to reduce KAs in these patients.

Conclusions

We report the first 3 cases to our knowledge of eruptive KAs associated with the use of pembrolizumab. In all 3 patients, eruptive KAs were controlled with conservative management, including intralesional and topical corticosteroids and cryosurgery, allowing patients to continue therapy with pembrolizumab. The inflammatory infiltrate seen on KAs biopsy specimens and the complete response to corticosteroids suggest that this may represent an immune-related dAE. Given that immune-related dAEs typically take longer to develop, patients receiving pembrolizumab will need to be monitored closely. Further research is necessary to elucidate the putative genetic and immune mechanism related to pembrolizumab-induced dAEs.