Treatment of Immune Checkpoint Inhibitor-Mediated Psoriasis: A Systematic Review

To the Editor:

Immune checkpoint inhibitors (ICIs) have revolutionized the treatment of advanced/metastatic cancers.¹ Since ICIs augment Th1/Th17 cell activity and IL-17A secretion, ICI-induced de novo psoriasis and exacerbations of pre-existing psoriasis of diverse phenotypes have been widely reported.^1,2 While many efficacious treatments exist for idiopathic psoriasis, recommendations for managing ICI-mediated psoriasis are insufficient and notably absent from the National Comprehensive Cancer Network guidelines for treating immune-related adverse events (irAEs).³ To support future recommendations, we systemically reviewed the treatment of ICI-mediated psoriasis.

Following PRISMA guidelines, studies reporting psoriasis amongst ICI-treated individuals were identified in PubMed and Embase in December 2020 (PROSPERO ID: CRD42021226337). Search strategy/criteria are provided (Figure S1, Supplement).

Among 1,393 references, 60 studies were included (Figure S2, Supplement), yielding 242 patients with individual/near-individual-level data (mean age 67.3 years; 77.3% male; 82.8% anti-PD-1 monotherapy) (Table S1, Supplement). This cohort included 55.0% (120/218) de novo psoriasis and 45.0% (98/218) pre-existing disease. The mean pre-toxicity ICI cycles (overall n=8.8) were fewer in cases of pre-existing psoriasis (n=6.4) versus de novo psoriasis (n=9.9). Psoriasis vulgaris was the most reported phenotype (70.6%), followed by palmoplantar (38.3%), nail (28.6%), and guttate (21.4%) psoriasis.

Topical steroids were the most frequently delivered treatment (181/218, 83.0%), which most patients received as monotherapy (111/218, 50.9%). Other agents included acitretin (39/219, 17.9%), systemic steroids (35/218, 16.1%), phototherapy (20/218, 9.2%), and methotrexate (14/218, 6.4%). Biologics-treated cases (9/218, 4.1%) are detailed separately (Table S4, Supplement).

Patients who received topical steroid monotherapy had a reduced rate of permanent ICI discontinuation (12.9% vs 31.5%, P=0.002). Patients who received systemic steroids more frequently achieved complete/partial response compared to those without systemic steroids (100.0% vs 82.9%, P=0.017), but were more likely to have ICIs permanently discontinued (59.3% vs 22.6%, P<0.001). Outcomes per-treatment are provided (Table S2, Supplement). Overall, most patients experienced a partial (113/197, 57.4%) or complete (59/197, 29.9%) response. Most patients had ICI treatment continued without interruption (119/205, 58.0%), but 47 (22.9%) had ICIs permanently discontinued.

In this systematic review of ICI-mediated psoriasis treatment, we identified a reduced rate of ICI discontinuation among patients treated with topical steroid monotherapy, likely confounded by toxicity severity. All systemic steroid-treated patients demonstrated response, but most ultimately experienced permanent ICI discontinuation. While steroids are frequently recommended and utilized by oncologists for managing irAEs across organ systems,³ we emphasize that systemic corticosteroids are not recommended in the American Academy of Dermatology-National Psoriasis Foundation management guidelines for idiopathic psoriasis, given lack of durable efficacy without significant side-effects and flares upon tapering.⁴ Systemic steroids may additionally dampen the ICI-driven anti-tumor effect.^1,2 Our clinical opinion favors using non-immunosuppressive options (i.e. acitretin, phototherapy) or highly specific psoriasis-directed biologics for cases refractory to topical steroid monotherapy, given that reduced tumor expression of key psoriasis pathway-mediators (TNF, IL23A, IL17A) does not affect overall survival across most tumor types.⁵

Limitations include utilizing predominantly observational data, invoking reporting bias and publication bias. Despite our male skew, the predominance of melanoma/lung cancer and anti-PD-1 monotherapy is consistent with the general ICI-treated population as previously reported.^1,2 These data support the establishment of consensus-driven expert recommendations for the treatment of ICI-mediated psoriasis.

Supplementary Material

Supplement 1

Supplemental Table S1: Individual Characteristics, Clinical Features, Treatments, and Treatment Outcomes of Reported Cases of Immune Checkpoint Inhibitor-Mediated Psoriasis

^a Percentages are over the total number of patients with available data; patients with unknown data are not included in the denominator.

^b Percentages are over the total number of patients included in the systematic review (242).

^c “Other” cancer types were renal cell carcinoma (n=9), hepatocellular carcinoma (n=4), lymphoma (total n=4; Hodgkin’s lymphoma, n=2; diffuse large B-cell lymphoma, n=1; cutaneous T-cell lymphoma, n=1), gastric/esophageal (n=3), sarcoma, any (n=3), breast (n=1), Merkel cell carcinoma (n=1), and ovarian (n=1).

^d At time of publication.

^e Three total cases of linear psoriasis (1), rupioid psoriasis (1), and follicular psoriasis (1) were identified in the literature.

^f Note that individual patients can belong to multiple treatment groups.

^g Biologics included adalimumab (1); etanercept (1); guselkumab (2); infliximab (1); secukinumab (2); ustekinumab (2); see Supplemental Table S4.

^h All three cases of sulfasalazine treatment were for psoriatic arthritis.

ⁱ Patients received no therapy (4); cyclosporin A/cyclosporine (2); etoposide (1); melphalan and autologous stem cell transplant (1).

^j “Complete response” was defined by the presence of specific language or clinical images clearly describing clearance, rather than improvement or control, of lesions. In studies with tiered responses “no response; moderate response; excellent response,” the second tier was considered “partial response,” and the third tier was considered “complete response.” In studies with binary responses “none; controlled,” the second tier was considered “partial response.”

^k A total of 24 (9.9%) individuals were included from studies in which treatment agents were reported with associated treatment efficacy and immunotherapy outcomes reported in aggregate, but for which individual-level agents received is unavailable. 14 individuals received “non-systemic therapies only.” 6 individuals received “unspecified immunomodulators.” 4 individuals received “Unknown.”

Abbreviations: CTLA-4 = cytotoxic T-lymphocyte-associated protein 4. HNC = head and neck carcinoma. ICI = immune checkpoint inhibitor. NB-UVB = narrow band ultraviolet light B. NSCLC = non-small cell lung cancer. PD-1 = Programmed cell death protein 1. PD-L1 = Programmed cell death protein-ligand 1. PsO = psoriasis. PUVA = psoralen and ultraviolet light A. SCLC = small cell lung cancer.

Supplemental Table S2: Treatment Efficacy and ICI Continuation Outcomes by Psoriasis-Directed Treatment Received

^a The total number of patients with specific reported agents is 218. Patients for whom the specific treatment delivered was not reported or unknown (n=24, reported in Table 1) are not included in this denominator.

^b Note that these percentages are calculated with patients with unavailable data “N/A” removed from the denominator.

^c Biologics included adalimumab (1); etanercept (1); guselkumab (2); infliximab (1); secukinumab (2); and ustekinumab (2); see Supplemental Table S4.

^d All three cases of sulfasalazine treatment were for psoriatic arthritis.

^e “ICI continued” and “ICI temporarily held and restarted” were combined into a single outcome group due to limited granularity of reviewed data.

^f A single patient who received topical steroids and apremilast had ICI therapy discontinued prior to onset of psoriasiform toxicity.

Abbreviations: ICI = immune checkpoint inhibitor. N/A = not available. NB-UVB = narrow band ultraviolet light B. PUVA = psoralen and ultraviolet light A.

Supplemental Table S3: Study Characteristics

^a One systematic review met inclusion criteria; all cited cases that met inclusion criteria were already included through independent review of primary literature.

Supplemental Table S4: Summary of Reported Patients who Received Biologics

^a The retrospective cohort studies from Nikolaou 2020 and Phillips 2019 report age, sex, cancer type, ICI agent, and number of cycles received in aggregate, and thus individual-level data is unavailable for these subjects.

^b Specific ICI agent was not reported at the individual level in the Phillips 2019 study.

^c The Nikolaou 2020 study reports the novelty of the psoriasiform toxicity in aggregate, and thus individual-level data is unavailable for these subjects.

^d Novelty of the psoriasiform toxicity is not reported in the Phillips 2019 study.

^e Supplemental data for these patients in the Nikolaou 2020 study categorized immunotherapy outcomes in binary as discontinued or not discontinued.

^f Both cases were described in the Phillips 2019 study as “corticosteroid-refractory or dependent.”

^g Both cases were described as “moderate-significant response,” categorized in this systematic review as a partial response.

Abbreviations: BMR = best malignancy response. CONT = continued. CR = complete response. D/C = discontinued. ICI = immune checkpoint inhibitor. IL = interleukin. N/A = not available. NB-UVB = narrow band ultraviolet light B. NR = not reported. NSCLC = non-small cell lung cancer. PR = partial response. Q = every. TNF = tumor necrosis factor.

Supplemental Figure S1: Search Strategy

Supplemental Figure S2: PRISMA Study Flow Diagram

Abbreviations: ICI = immune checkpoint inhibitor.

Supplemental Figure S3: Treatment Efficacy by Treatment Group

Abbreviations: CR = complete response. NR = no response. PR = partial response.

Supplemental Figure S4: Immune Checkpoint Inhibitor Continuation by Treatment Group

Abbreviations: CR = complete response. NR = no response. PR = partial response.

Supplement 2

Supplemental Table S5: Complete Data Extracted from 60 Included Studies

Supplemental Table S6: Complete List of 60 Included Studies