Cutaneous Granulomatous Reactions in Patients Treated with Immune Checkpoint Inhibitors: A Systematic Review

To the Editor:

The development of immune checkpoint inhibitors (ICIs) represents a significant therapeutic advancement in medical oncology. However, adverse events of non-infectious granulomatous dermatoses (GD) have emerged with their increasing use. Cutaneous GD encompasses a wide range of skin diseases characterized by granuloma formation and histiocyte accumulation.1 This systematic review evaluates all reports of ICI-associated non-infectious cutaneous GD to improve our understanding of the clinical implications.

MEDLINE and EMBASE were searched on March 1, 2024, in accordance with PRISMA guidelines (Supplemental Figure S1). The quality of evidence was evaluated using the Oxford Centre for Evidence-Based Medicine 2011 Levels of Evidence. ²

Eighty-two articles were included (n = 128 patients). The median age was 60 years (range: 26–88 years), with a slight female predominance (n = 57/97, 58.8%). Patients were treated with programmed death-1 (n = 58/128, 45.3%), programmed death-ligand 1 (n = 6/128, 4.7%), and cytotoxic T-lymphocyte-associated protein 4 inhibitors (n = 22/128, 17.2%) (Supplemental Table S1).

Melanomas were the most common tumor type treated with immunotherapy (n = 66/101, 65.3%). The GDs observed were as follows: sarcoid-like granulomas (n = 104/128, 81.3%), interstitial granulomatous dermatitis (n = 6/128, 4.7%), granuloma annulare (n = 10/128, 7.8%), granulomatous panniculitis (n = 7/128, 5.5%), and foreign-body granuloma (n = 1/128, 0.8%). Excluding granuloma annulare and patients with unreported data, extracutaneous involvement was observed in 65.2% (58/89), including mediastinal/hilar lymphadenopathy (n = 47/89, 52.8%) and lung pathology (n = 32/89, 36.0%). Interestingly, among these 58 patients with extracutaneous involvement, symptoms related to systemic GD, such as dyspnea or myalgia, were observed in only 31.0%. However, radiological manifestations of systemic involvement such as PET-positive lymphadenopathy prompted immunotherapy discontinuation in 53.4% (n = 31/58) of these patients. Conversely, immunotherapy cessation occurred in only 35.5% (n = 11/31) of patients with isolated skin symptoms. Where reported (n = 96), GD lesions were predominantly treated with systemic steroids (n = 42/96, 43.8%) and topical steroids (n = 33/96, 34.4%); most patients (n = 74/85, 87.1%) experienced partial-to-complete skin clearance irrespective of the treatment method. Among 71 patients with reported tumor outcomes, 25.3%, 21.1%, and 23.9% experienced remission, partial tumor response, and stable disease, respectively. Disease progression was observed in 30.0% of patients.

ICI-associated GD is likely driven by a hyperactive immune response involving a marked increase in proliferation of T-helper-1 (Th1) cells. ³ Melanomas are also highly immunogenic, and the ICI therapy-enhanced destruction of tumor cells may expose additional neoantigens. ⁴ Thus, the increased number and function of Th1 cells, compounded by elevated neoantigen exposure, may promote a Th1 response and cytokine milieu favorable to developing GD lesions during ICI therapy.^3,4

There is limited consensus on the diagnosis and management of ICI-associated GD despite the diagnostic pitfalls of extracutaneous manifestations, which may clinically and radiologically mimic cancer recurrence or metastasis and thus lead to inappropriate immunotherapy interruptions. Using timely biopsy to confirm the etiology of radiological findings is imperative for appropriate management. Additionally, previous studies have noted that sarcoidosis or granulomatous reactions may be an indicator of an effective anti-tumor response. ⁵ Given the primarily positive tumor responses observed in our own study, continuation of immunotherapy should be considered in patients with limited systemic symptoms.

Limitations of this review include small sample sizes and lack of high-quality randomized controlled trials. Regardless, our study highlights that clinicians should maintain a high degree of suspicion for these adverse reactions, which may mimic metastatic disease. Our findings offer helpful management considerations for dermatologists and dermatopathologists, who can substantially aid in distinguishing between these 2 disease entities to optimize patient outcomes.