Erosive Mucocutaneous Eruptions in Patients Treated With Immune Checkpoint Inhibitors: A Case Series

Immune checkpoint inhibitor (ICI)-related erosive mucocutaneous eruption (IEME) diseases include Stevens-Johnson syndrome (SJS), toxic epidermal necrolysis (TEN), erythema multiforme (EM), and erosive lichenoid reactions [1, 2]. The hypothesized cause of IEME is reduced immune tolerance from ICIs, which may increase susceptibility to adverse reactions [3]. We aimed to describe the clinical characteristics and course of IEME [4]. We retrospectively searched our institution's electronic health record for adult patients (> 18 years) who received FDA-approved ICIs and had a diagnosis of TEN, SJS, or EM (by ICD-9 and ICD-10 codes) from 2013 through August 31, 2023. We assessed demographic information, medical history, available skin biopsy results, and clinical course. The Mayo Clinic Institutional Review Board approved this study. Of 12 patients who met the criteria, 7 had mucocutaneous eruptions attributed to ICIs (i.e., IEME). Five patients had reactions attributed to concomitant medications or other causes and were not further included. In the seven patients with IEME (mean age, 69.3 years; four women and three men), three patients had SJS/TEN and four had EM (Tables 1 and 2). The mean time from ICI initiation to rash onset was 155.1 days (Table 1), the longest with atezolizumab (mean, 326.5 days; n = 2) and the shortest with pembrolizumab (mean, 45.8 days; n = 4). Pembrolizumab, a PD-1 inhibitor, was more frequently associated with EM-like eruptions than with SJS/TEN (Table 2). Concomitant medications included trimethoprim-sulfamethoxazole (n = 2), cephalosporins (n = 1), and amoxicillin (n = 1). All available skin biopsies (n = 6) were characterized as interface dermatitis, showing vacuolar interface inflammation and/or dyskeratotic epidermal keratinocytes. The median (range) hospital stay for an adverse drug reaction was 4.5 (2–8) days (n = 6). Treatment ranged from no intervention to systemic corticosteroids and cyclosporine; healing times varied from 1 week to 1 year. The onset of rash varied, from during treatment to 3 months after discontinuation—most within the first month after discontinuation. The rash typically manifested as skin pain, sometimes preceded by itching and milder symptoms, and rash morphology was diverse (Table 1). Although severe mucocutaneous eruptions may sometimes be categorized as severe cutaneous adverse reactions (SCARs), our 3 SJS/TEN cases demonstrated distinct features, including delayed onset, mild initial morphology, rare ocular involvement, a benign clinical course, and favorable treatment response. This suggests that IEME may diverge from classic SJS/TEN [2]. The association of IEME with concomitant medications, such as trimethoprim-sulfamethoxazole (Patients SJS #1 and EM #4) and amoxicillin (Patient EM #5), also suggests a two-hit mechanism, in which ICIs lower immune tolerance, amplifying sensitivity to subsequent drug exposures [3]. This highlights the need for caution with coadministered medications and further investigation into IEME pathogenesis. Histopathologic evaluation in IEME generally demonstrates a lichenoid interface, often with full-thickness epidermal necrosis [2, 3], as in our cases [5]. The 5 patients not included in the main analysis underscore the potential for a two-hit mechanism and a heightened risk of SCAR in the context of ICI therapy, even months or years after discontinuation. Notably, ICIs may increase the positive predictive value of reactions among high-notoriety drugs, while also potentially increasing the risk of rash, of varying severity, associated with drugs not as commonly associated with delayed rash. These cases suggest that, with appropriate management, ICIs may be safely continued or resumed after resolution of cutaneous reactions in certain patients. Study limitations include a small sample size, a retrospective descriptive nature, potential selection bias, the inability to establish causation, potential confounding, and the absence of a control group. In this cohort, IEME distinctly presented with a delayed onset, milder initial morphologic presentation/clinical course, and more favorable outcomes compared with classic SJS/TEN. Patients taking ICIs require close observation and urgent consultation for SJS/TEN-like reactions, which are a rare but serious complication. The Scientific Publications staff at Mayo Clinic provided editorial consultation, proofreading, and administrative and clerical support. Declaration of AI-assisted technologies in the writing process: ChatGPT, a generative AI tool, was used sparingly to refine original content for minor grammatical corrections and clarity. The writing, analysis, and interpretation of the data were conducted primarily by the authors, who take full responsibility for the final content of the manuscript. This study was supported by internal funds from Mayo Clinic. IRB Approval Status: Reviewed and approved by the Mayo Clinic Institutional Review Board; ID # 22-012372. Dr. Alavi serves on the Board of Directors for the Hidradenitis Suppurativa Foundation; served as consultant/advisor for Abbvie, Almirall, Boehringer Ingelheim, Incyte, InflaRx, LEO Pharma, Novartis, Sanofi, and UCB; and served as principal investigator for Boehringer Ingelheim and Processa. The data that support the findings of this study are available from the corresponding author upon reasonable request.