Coexistence of psoriasiform stratum corneum and underlying lichenoid features in skin biopsies of cutaneous reactions to anti-PD1.

Abstract

Immune checkpoint inhibitors have been established as an excellent option for the treatment of advanced malignancies, thus the prescription of anti-PD1/PD-L1 antibodies (nivolumab, pembrolizumab) is rapidly increasing.1 Despite the efficacy of immunotherapies, they may cause a non-specific activation of T-lymphocytes leading to immune-related adverse events including induction of autoimmune processes or flares of pre-existing conditions.2 Among these, cutaneous immune related processes are common; clinical trials show that 8%–40% of patients treated with anti-PD1 develop skin lesions.1 The most prevalent subtypes of cutaneous toxicities secondary to anti-PD1 therapy described in the literature include maculopapular rash, pruritus, eczema, and psoriasiform and lichenoid eruptions. Other less common toxicities have also been reported: alopecia, bullous pemphigoid, skin hypopigmentation/depigmentation, alopecia, toxic epidermal necrolysis, among others.1, 3 Given the increased use of these therapies, there is a need for adequate characterization and classification of the immune-related skin manifestations produced by anti-PD1, from both a clinical and a histopathological perspective. This will contribute to improve our knowledge of these dermatoses thereby contributing to provide better management of patients treated with these therapies.4 The aim of this report is to provide a clinical and histopathological description of toxicodermias secondary to anti-PD1/PD-L1 characterized by previously unreported coexistence of two dermatoses. The study included all consecutive cases that experienced an anti-PD1-induced cutaneous eruption and had a skin biopsy showing coexistence of psoriasiform stratum corneum and underlying lichenoid features from September 1, 2021 to June 31, 2022 at our center. From this series we collected: demographic parameters, clinical history of previous dermatoses, time between initiation of the immunotherapy and the development of the skin lesions, type of therapy (immunotherapy drug and dose schedule), clinical and histopathological features, follow-up, treatment, and outcome. A total of three patients were included, with a mean follow-up period of 10 months. Mean age was 70.66 years (69–73 years). One patient was female. Only one of the patients had a previous diagnosis of psoriasis. The average time between skin toxicity onset and initiation of the immunotherapy was 8.33 weeks (3–18 weeks). All patients presented pruritic lesions affecting trunk and extremities, consisting of erythematous, scaly, well-circumscribed, papules-plaques with a lichenoid hue. Two of the patients presented palmoplantar involvement. First clinical impressions were, anti-PD1-induced lichenoid dermatitis in Patient 1, a psoriasis flare with lichenoid features in Patient 2, and pityriasis rubra pilaris in Patient 3 (Figure 1A–E; Table 1). As a result of the atypical clinical features, skin biopsies were performed in all the cases, showing mixed histopathological findings of psoriatic and lichenoid dermatosis. The patients had not received previous treatment with topical or systemic therapy for the dermatosis before our clinical evaluation and skin biopsy. All skin biopsy specimens showed neutrophil-imbued parakeratotic stratum corneum reminiscent of psoriasis.5 In addition, hypogranulosis and irregular non-psoriatic acanthosis were evident in all the biopsies. All skin biopsy specimens also displayed interface dermatitis. Vacuolar degeneration with presence of few lymphocytes and sparse neutrophil exocytosis was seen in Case 1. A dense band-like lymphohistiocytic infiltrate obscuring the dermal–epidermal junction was seen in Cases 2 and 3. Squamatization of basal layer keratinocytes and keratinocyte dysmaturation were also commonly identified. Finally, one case also presented apoptotic keratinocytes (Figure 2A–I; Table 2). The component of the dermal inflammatory infiltrate was mostly lymphocytic, while neutrophils were found in the corneum layer, and in Patient 1 in the dermal–epidermal junction. In Patients 2 and 3 neutrophils were scarce, and lymphocytes predominated, mixed with few plasma cells and almost no eosinophils. Only one of our patients was properly controlled with topical therapy, while two patients required systemic drugs. Treatment with methotrexate was started in one case; the other case, with previous diagnosis of psoriasis, achieved adequate control of the skin lesions with guselkumab. None of the patients showed progression of the underlying neoplasm. According to the literature, the period between the start of anti-PD1 therapy and the appearance of skin lesions ranges from 3 weeks for psoriatic forms to 6–12 weeks for lichenoid eruptions, a similar range to our mixed-featured cases (3–21 weeks). As in our cases, when clinically psoriasiform lesions appear, palmoplantar involvement may occur. The area and morphology of these induced lesions differ from previous plaque psoriasis when personal history is present. It is noteworthy that Patient 2, who previously presented a severe psoriasis that was controlled with biological therapy, showed clear lichenoid lesions different from the previous psoriasis. As both, lichen ruber and psoriasis are driven by Th17 immune responses, PD1 immune modulation may explain the morphological change. Although some authors have also observed psoriasiform lesions with erythematous-violaceous papulo-plaques predominantly on the neckline and trunk, showing pure lichenoid histologic changes,6-8 in our series, histopathological features showed mixed psoriasiform and lichenoid changes. There was a spectrum, as Patient 1 lesions were histopathologically closer to psoriasis, though with interface dermatitis; while Patients 2 and 3 presented a more lichenoid spectrum admixed with hypogranulosis and a detached and focally parakeratotic corneum layer. Results previously reported in the literature suggest that these combined or admixed histopathological patterns may be considered as a clue for a drug-induced reaction; in addition, lichenoid features, sometimes with hints pointing to an anti-PD1 cutaneous reaction, such as spongiosis, parakeratosis, and eosinophils have been described.9, 10 Parakeratosis, also described by other authors in lichenoid anti-PD1 reactions, was observed in our cases, as well as spongiosis in Case 3, but we did not find eosinophilic infiltrates in our series.11 Although Curry et al described a psoriasis-like presentation, with presence of pustules, we found histopathological spongiform pustules but not clinical pustules.9 As anti-PD1 therapy can activate common pathways for psoriasis and lichen planus, we can hypothesize that the presence of this mixed phenotype may be underreported in the literature. Th17 involvement in psoriasis is well known but IL-17 might also have a pathogenic role in lichen planus.12 The cutaneous reaction pattern found in our patients is quite novel because we consistently observed typical findings of psoriasis on the stratum corneum (mostly parakeratosis involving neutrophils and foci of hypogranulosis) and a dermal lichen planus with band-like infiltrates was underlying it. This particular combination might be suggestive of anti-PD1 toxicodermia.13 We describe a peculiar combination of features of skin lesions, with psoriasiform stratum corneum and lichenoid epidermal involvement, in three patients with anti-PD1-related dermatosis. With this peculiar pattern, we contribute to increase the gamut of anti-PD1-associated cutaneous reactions. An accurate characterization of immunotherapy-associated skin lesions is crucial for dermatopathologists and dermatologists in order to achieve more specific diagnoses and better management. The authors declare no conflict on interest. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.