Safety of Immune Checkpoint Inhibitors in Cancer Patients With Preexisting Autoimmune Vasculitis

Abstract

Objective

Immune checkpoint inhibitors (ICIs) are effective cancer therapies but often cause serious immune‐related adverse events (irAEs). Patients with preexisting autoimmune diseases, including vasculitis, are excluded from trials. We aimed to evaluate the frequency, severity, and outcomes of vasculitis flares and irAEs in this population.

Methods

We performed a retrospective review of cancer patients with prior vasculitis treated with ICIs at our institution and conducted a literature search for additional cases. Data included baseline features, vasculitis flares, irAEs, treatments, and tumor response.

Results

Twenty‐five patients were identified (16 from our institution and 9 from the literature). Median age was 71 years; 13 (52%) were female. Vasculitis types included giant cell arteritis (GCA, n = 9), granulomatosis with polyangiitis (GPA, n = 8), eosinophilic granulomatosis with polyangiitis (EGPA, n = 2), leukocytoclastic vasculitis (n = 2), other cutaneous vasculitis (n = 2), and Henoch–Schönlein purpura and Behçet disease (1 each). Cancer types were diverse. Most patients received anti–programmed death‐1 monotherapy (n = 17). Eight patients (31%) experienced a vasculitis flare after ICI initiation (2/9 GCA, 4/8 GPA, 2/4 cutaneous vasculitis). Flares occurred after a median of seven weeks. Treatment included glucocorticoids in seven patients, combined with biologic or cytotoxic agents depending on vasculitis type. Seven of eight flares resolved, and four patients continued ICI therapy. Three additional patients developed de novo irAEs: severe hepatitis (n = 1), grade 3 colitis (n = 1), and grade 3 autoimmune hemolytic anemia (n = 1), all improving with treatment. One patient with GCA died due to a flare; no deaths were attributed to de novo irAEs. Overall, more than one‐third of patients achieved a favorable tumor response.

Conclusion

Nearly one‐third of patients with preexisting vasculitis experienced a disease flare during ICI therapy, with one fatal case. For most, outcomes were favorable with effective flare management. Preexisting vasculitis should not be considered an absolute contraindication for cancer immunotherapy with ICI.

INTRODUCTION

Immune checkpoint inhibitors (ICIs) have revolutionized cancer therapy, becoming the standard of care in the treatment of various malignancies both in the metastatic and adjuvant settings. These therapies target key immune checkpoints, such as CTLA‐4, programmed cell death 1 (PD‐1), programmed death ligand 1 (PD‐L1), and lymphocyte activation gene 3 (LAG‐3), thereby enhancing the immune system's ability to fight cancer. ¹ Although ICIs have demonstrated significant clinical benefits, they are also associated with a wide range of immune‐related adverse events (irAEs) that can affect every organ or system. ² Therefore, patients with preexisting autoimmune diseases have been excluded from the original clinical trials that led to ICIs’ approval, and to date, clinicians remain hesitant to use ICIs in this special cancer population due to concerns about flares of the underlying autoimmune disease and an increased risk of severe irAEs.

Vasculitis are a group of heterogeneous autoimmune diseases characterized by inflammation of the blood vessels, with varying phenotypes and severity depending on the type and size of the vessels involved. ³ , ⁴ Given the variation in the clinical presentation and pathogenesis of vasculitis, classification criteria have been developed based on the size of the vessels involved, small, medium, or large, and also on the presence of antineutrophilic cytoplasmic antibody (ANCA). ⁵ , ⁶ Small‐vessel vasculitis most commonly cause rash and ulcers, but can also be severe and compromise vital organs such as the kidney. Vasculitis that affects medium to large vessels often results in inflammation and ischemia of vital organs requiring prolonged immunosuppression. Conceivably, patients with vasculitis may experience significant flares following ICI therapy, because of the heightened immune and inflammatory state caused by these agents. ⁷ There are limited data on the frequency, severity, management, and outcomes of vasculitis exacerbations in patients undergoing ICI therapy. Most studies have thus far described de novo vasculitis cases after ICI therapy with reports including conditions such as giant cell arteritis (GCA), aortitis, primary angiitis of the central nervous system, isolated vasculitis neuropathy, Behçet disease, IgA vasculitis, antineutrophilic cytoplasmic antibody–associated vasculitis, cutaneous small‐vessel vasculitis, and cryoglobulinemic vasculitis. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ However, there are scarce data on patients with preexisting vasculitis who have undergone ICI therapy. Our aim was to conduct a comprehensive review of the clinical outcomes of patients with preexisting vasculitis receiving ICI, including both patients from our center and patients reported in the literature. To our knowledge, this is the largest cases series describing clinical outcomes in this population.

MATERIALS AND METHODS

Study design and participants

We performed a retrospective review of electronic health records of patients with cancer and a history of autoimmune vasculitis who received ICI at The University of Texas MD Anderson Cancer Center from 2004 to 2024. We included patients who had any type of vasculitis confirmed by a rheumatologist, a clinical history with signs and symptoms of vasculitis, and appropriate treatment history. Eligible patients must also have received at least one dose of an ICI currently approved by the US Food and Drug Administration, including ipilimumab, tremelimumab, nivolumab, pembrolizumab, dostarlimab, cemiplimab, retifanlimab, toripalimab, tislelizumab, atezolizumab, avelumab, durvalumab, or relatlimab. The diagnosis of vasculitis was reviewed by evaluating each patient's electronic health record by two independent reviewers.

In addition, we conducted a scoping review with a comprehensive literature search to identify and summarize cases of preexisting autoimmune vasculitis in patients receiving ICI therapy. The search was conducted on Medline from its inception to December 2024. The search employed specific terms related to ICIs, including currently approved ones such as “anti‐CTLA4,” “anti‐PD1,” “anti‐PDL1,” “anti‐LAG‐3,” and all 13 drugs mentioned earlier. Additionally, various terms associated with autoimmune vasculitis were included, such as “Giant Cell Arteritis,” “Temporal Arteritis,” “Takayasu Arteritis,” “Polyarteritis Nodosa,” “Kawasaki Disease,” “Granulomatosis with Polyangiitis,” “Wegener's Granulomatosis,” “Microscopic Polyangiitis,” “Eosinophilic Granulomatosis with Polyangiitis,” “Churg‐Strauss Syndrome,” “Henoch‐Schönlein Purpura,” “Cryoglobulinemic Vasculitis,” “Systemic Lupus Erythematosus Vasculitis,” “Behçet's Disease,” “Cogan's Syndrome,” “Cutaneous Leukocytoclastic Angiitis.”

Data collection and analysis

For both the MD Anderson and the published case reports, we extracted data on patient demographics, oncologic history, ICI details (drug, dosage, timing, and cycles), and type of autoimmune vasculitis classified according to the 2012 Revised Chapel‐Hill Nomenclature. ⁵ Characteristics of preexisting vasculitis (organ involvement, biopsy results, associated antibodies, and prior treatment), status of vasculitis at ICI initiation (active vs inactive disease, and on vs off immunosuppressive therapy), vasculitis flares if any (timing, clinical phenotypes, and management strategies), de novo irAEs (incidence, type, severity, and definition based on the 2023 Society for Immunotherapy of Cancer), ¹⁷ ICI rechallenge when reported, and overall clinical outcomes (any adverse event and tumor response).

We used descriptive statistics to summarize the data, including medians and ranges for continuous variables and frequencies and percentages for categorical variables. The study was approved by the Institutional Review Board at MD Anderson's Cancer Center (PA15‐0071). Informed consent was not required.

RESULTS

Patient characteristics

A total of 23,332 patients received ICI at MD Anderson between January 1, 2004, and June 30, 2024. Among these, 16 (0.07%) patients with preexisting vasculitis were identified. Additionally, a literature search identified 9 more patients with preexisting vasculitis. Therefore, a total of 25 patients are described here.

Patients’ demographics and baseline clinical characteristics from our institution are presented in Table 1, whereas corresponding data from literature‐reported cases are summarized in Table 2. Eight (50%) of the patients from MD Anderson were female with a median age of 74 years (range 55–85). Five (56%) of the patients reported in the literature were female with a median age of 67 years (range 48–83). In the entire series of 25 patients, the most common cancer type was melanoma (n = 7, 28%). Most patients (n = 20, 91%) had stage III to IV advanced cancer, and single‐agent anti‐PD‐1 was the most frequently used ICI (n = 17, 68%). Anti‐PD‐L1 was used in five patients (20%), whereas the remaining three patients (12%) received anti‐CTLA‐4, either alone or in combination with anti‐PD‐1. Vasculitis diagnoses were nine (34.6%) GCA, eight (30.8%) granulomatosis with polyangiitis (GPA), two eosinophilic granulomatosis with polyangiitis (EGPA), two leukocytoclastic vasculitis, two other skin vasculitis, one Henoch–Schönlein purpura, and one Behçet disease. Only three patients, one with GPA, one with cutaneous vasculitis and one with Behçet disease, had active vasculitis when ICI was initiated. Nine patients were receiving treatment for vasculitis at ICI initiation, including prednisone (n = 5, mean dose 7.6 mg daily), rituximab (n = 3), tocilizumab and methotrexate (n = 1 patient each). The median follow‐up time since initiation of ICI therapy was 1.1 years (range 0.1–3.8).

Table 1.

Baseline characteristics of patients at MD Anderson Cancer Center with preexisting vasculitis who received ICIs (n = 16)^*

Case no.	Race	Vasculitis type	Disease duration, y	Positive antibodies	Previous vasculitis treatment	Cancer type	Stage	ICI therapy
1	White	GCA	4.4	ANA	Glucocorticoids	Mediastinum	IV	Atezolizumab
2	White	GCA	6.2	None	Glucocorticoids	Melanoma	III	Nivolumab + Ipilimumab
3	White	GCA	4.8	ANA	Glucocorticoids	Colon	IV	Pembrolizumab
4	White	GCA	0.9	None	Glucocorticoids	Melanoma	IV	Pembrolizumab
5	White	GCA	9.9	None	Glucocorticoids, methotrexate, tocilizumab	Lung	IV	Durvalumab
6	White	GCA	10.5	ANA	Glucocorticoids	Thyroid	IV	Pembrolizumab
7	White	GCA	0.5		Glucocorticoids	Thyroid	IV	Pembrolizumab
8	White	GCA	1.8		Glucocorticoids	Sarcoma	IV	Pembrolizumab
9	White	GPA	6.7	C‐ANCA, PR3	Glucocorticoids, cyclophosphamide, rituximab, hydroxychloroquine	Melanoma	IV	Pembrolizumab
10	White	GPA	17.6	ANA ANCA	Glucocorticoids, cyclophosphamide	Bladder	IV	Pembrolizumab
11	White	GPA	24.7	ANCA PR3	Glucocorticoids, cyclophosphamide	Endometrium	II	Pembrolizumab
12	White	LCV	2.6	RF	Glucocorticoids, dapsone	Cervix	I	Pembrolizumab
13	Hispanic	LCV	19.2		None	Liver	III	Atezolizumab
14	Hispanic	OSV	3.6	None	Glucocorticoids, mycophenolate mofetil, methotrexate	CLL	NA	Nivolumab
15	White	OSV	25.4	None	Glucocorticoids	Lung	III	Atezolizumab
16	White	HSP	2.1		Glucocorticoids	Liver	III	Pembrolizumab

^*ANA, antinuclear antibody; C‐ANCA, cytoplasmic antineutrophil cytoplasmic antibody; CLL, chronic lymphocytic leukemia; GCA, giant cell arteritis; GPA, granulomatosis with polyangiitis; HSP, Henoch–Schönlein purpura; ICI, immune checkpoint inhibitor; LCV, leukocytoclastic vasculitis; NA, not applicable; OSV, other skin vasculitis; PR3, proteinase 3 antibody; RF, rheumatoid factor.

Table 2.

Baseline characteristics of case reports in the literature with preexisting vasculitis who received ICIs (n = 9)^*

Case no. (ref)	Race	Vasculitis type	Disease duration, y	Positive antibodies	Previous vasculitis treatment	Cancer type	Stage	ICI therapy
17 (Nabel et al 19 )	NR	GPA	10	PR3	Glucocorticoids, methotrexate, cyclophosphamide	Bladder	IV	Pembrolizumab
18 (Aqeel et al 20 )	White	GPA	NR	ANCA PR3	NR	NSCLC	NR	Pembrolizumab
19 (Hung et al 21 )	NR	GPA a	3	ANCA PR3	None	Melanoma	III	Pembrolizumab
20 (Wulfken et al 46 )	NR	GPA	23	ANCA PR3	Glucocorticoids, rituximab, cyclophosphamide, infliximab, azathioprine, methotrexate	Melanoma	IV	Nivolumab
21 (Yamada et al 47 )	NR	GPA	4	ANCA MPO	Glucocorticoids, methotrexate	NSCLC	III	Pembrolizumab
22 (Ramos et al 22 )	NR	EGPA	0.9	ANCA MPO	Glucocorticoids, rituximab, cyclophosphamide, plasmapheresis	Endometrial	NR	Pembrolizumab
23 (Maul et al 23 )	NR	EGPA	6	NR	Glucocorticoids, azathioprine	Melanoma	IV	Ipilimumab
24 (Pedersen et al 24 )	NR	Behçet disease	5	NR	Methotrexate, hydroxychloroquine	Melanoma	IV	Ipilimumab
25 (Shanklin et al 18 )	White	GCA	14	NR	Glucocorticoids, methotrexate	Bladder	IV	Atezolizumab

^*ANCA MPO, antineutrophil cytoplasmic antibody – myeloperoxidase subtype; ANCA PR3, antineutrophil cytoplasmic antibody – proteinase 3 subtype; EGPA, eosinophilic granulomatosis with polyangiitis; GCA, giant cell arteritis; GPA, granulomatosis with polyangiitis; ICI, immune checkpoint inhibitor; NR, not reported; NSCLC, non–small cell lung cancer.

^a Sinus‐limited case.

Safety of ICI therapy in patients with preexisting vasculitis

Overall, 11 (44%) patients experienced flares and/or irAEs after ICI initiation (Table 3). Among them, 8 (32%) had a vasculitis flare, including 4 with GPA, 2 with GCA, and 2 with skin‐limited vasculitis, whereas 3 (12%) developed de novo irAEs. Of the three patients with active vasculitis at ICI initiation, two experienced a flare. Among 16 patients not receiving baseline immunosuppression for vasculitis, 7 (44%) experienced a vasculitis flare, and 1 (6%) developed de novo irAEs. In contrast, among nine patients receiving baseline immunosuppression, only one (11%) experienced a vasculitis flare, whereas two (22%) developed other irAEs.

Table 3.

Clinical outcomes of patients with vasculitis who received ICIs (n = 26)^*

Case no. (ref)	Disease activity at ICI initiation	Vasculitis treatment at ICI initiation	Vasculitis flare	Flare treatment	Flare outcome	Time to flare, wk	Other irAEs	Cancer outcome	Follow‐up time, y
Giant cell arteritis (n = 9)
1	Inactive	None	Yes	Glucocorticoids + sarilumab	Controlled	34	No	Remission	1.0
2	Inactive	Glucocorticoids	No	NA	NA	NA	Grade 4 hepatitis	Remission	3.1
3	Inactive	None	No	NA	NA	NA	Grade 3 AHA	Progression	0.1
4	Inactive	Glucocorticoids	No	NA	NA	NA	No	Progression	1.8
5	Inactive	Tocilizumab	No	NA	NA	NA	No	Remission	0.5
6	Inactive	None	No	NA	NA	NA	No	Progression	1.1
7	Inactive	None	No	NA	NA	NA	No	Stable	3.8
8	Inactive	None	No	NA	NA	NA	No	Progression	2.3
25 (Shanklin et al 18 )	Inactive	None	Yes	Glucocorticoids + cyclophosphamide	Poor	30.4	No	NA	0.1
Total, n (%)		3 (33)	2 (22)				2 (22)
Median						32			1.1
Granulomatosis with polyangiitis (n = 8)
9	Inactive	None	Yes	Glucocorticoids + rituximab	Controlled	1.3	No	Progression	0.6
10	Inactive	None	No	NA	NA	NA	No	Progression	1.7
11	Inactive	Rituximab	No	NA	NA	NA	No	Partial response	1.2
17 (Nabel et al 19 )	Inactive	None	Yes	Glucocorticoids + rituximab	Controlled	1	NA	Stable disease	0.2
18 (Aqeel et al 20 )	NA	NA	Yes	Glucocorticoids + rituximab	Controlled	4	NA	Progression	0.4
19 (Hung et al 21 )	Active	None	Yes	Glucocorticoids + rituximab, infliximab	Controlled	3	NA	NA	NA
20 (Wulfken et al 46 )	Inactive	Rituximab	No	No	NA	NA	No	Partial response	1.2
21 (Yamada et al 47 )	Inactive	Glucocorticoids + methotrexate	No	No	NA	NA	No	Partial response	0.5
Total, n (%)		3/7 (43)	4 (50)				0
Median						2.2			0.6
Cutaneous vasculitis (n = 4)
12	Active	None	Yes	Supportive	Controlled	27	No	Progression	1.1
13	Inactive	None	No	NA	NA	NA	No	Progression	1.9
14	Inactive	Glucocorticoids	Yes	Glucocorticoids	Controlled	10.1	No	Remission	1.3
15	Inactive	None	No	NA	NA	NA	No	Remission	2.0
Total, n (%)		1 (25)	2 (50)				0
Median						18.5			1.6
Other vasculitis (n = 4)
Eosinophilic granulomatosis with polyangiitis
22 (Ramos et al 22 )	Inactive	Rituximab	No	No	NA	NA	No	Progression	0.3
23 (Maul et al 23 )	Inactive	Glucocorticoids	No	No	NA	NA	Grade 3 colitis	Partial response	2.1
Behçet disease
24 (Pedersen et al 24 )	Active	None	No	No	NA	NA	No	Progression	0.3
Henoch–Schönlein purpura
16	Inactive	None	No	NA	NA	NA	No	Progression	2.1
Total, n (%)		2 (40)	0				1 (25)
Median									1.2

^*AHA, autoimmune hemolytic anemia; ICI, immune checkpoint inhibitor; irAEs, immune relates adverse events; NA, not applicable.

Vasculitis flares occurred in 6 of 17 patients (35%) who received anti‐PD‐1 therapy, 2 of 5 (40%) who received anti‐PD‐L1, and none of the 3 patients who received anti‐CTLA‐4, either as a single agent or in combination with anti‐PD‐1. De novo irAEs occurred in one patient (6%) who received anti‐PD‐1 therapy and two others (67%) who received anti‐CTLA‐4.

Of the 11 patients who experienced flares and/or irAEs, 10 (91%) required additional immunosuppression or an increase in baseline immunosuppression. Cancer outcomes were available for 23 patients (excluding 2 patients from the literature with missing data). Among them, 8 (35%) achieved partial or complete tumor response, 2 (9%) had stable disease, and 13 (57%) experienced cancer progression. When stratified by immunosuppressive treatment, five of nine patients (56%) who received immunosuppression achieved remission or remained stable, whereas four (44%) had cancer progression. In contrast, of the 14 patients who did not require immunosuppression, 5 (36%) achieved a response or had stable disease, whereas 9 (64%) had cancer progression.

One patient with GCA reported in the literature died following a vasculitis flare (see below). No deaths were attributed to de novo irAEs. However, 10 patients (42%) died due to cancer progression.

Clinical outcomes according to the type of vasculitis subtype are summarized in Table 3 and below.

GCA (n = 9)

Three patients (33%) were receiving vasculitis maintenance treatment (glucocorticoids, tocilizumab) before ICI, and none had active disease at ICI initiation. Two patients (22%) experienced a flare within 32 weeks of ICI initiation.

Case 1 (MD Anderson)

A patient in their 90s was diagnosed with polymyalgia rheumatica (PMR) with temporal artery biopsy‐proven GCA in 2019. High‐dose steroids started and tapered until May 2023. In December 2022, the patient was diagnosed with metastatic mediastinal carcinoma starting treatment with atezolizumab in October 2023. At this time, the patient was asymptomatic and was not on therapy for vasculitis. In December 2023, proximal muscle stiffness and headaches developed, and a temporal artery ultrasound in January 2024 showed signs consistent with a vasculitis flare. Prednisone was initiated, followed by sarilumab 200 mg every 2 weeks. As of October 2024, the patient remains asymptomatic with respect to their PMR and GCA, on a stable dose of prednisone 5 mg daily and sarilumab. The underlying cancer remains stable, and treatment with atezolizumab is ongoing.

Case 25 (literature case report)

A patient in their 70s had bladder cancer and GCA. ¹⁸ Fourteen years before initiating ICI therapy, the patient was diagnosed with GCA and treated with methotrexate and prednisone. Methotrexate had been discontinued five years after the cancer diagnosis. The vasculitis had been inactive for several years before ICI initiation, on low dose prednisone which was discontinued when starting ICI therapy. A GCA‐related stroke occurred after seven months of treatment with atezolizumab. The patient was treated with glucocorticoid pulses and cyclophosphamide (doses unspecified). Despite treatment, his condition deteriorated, and the patient died within two weeks, and death was attributed to the vasculitis.

De novo irAEs occurred in two patients (22%) with preexisting GCA after initiating ICI treatment. One patient developed grade 4 hepatitis, and the other patient had grade 3 autoimmune hemolytic anemia. Both patients were without disease activity when initiating ICI flares were managed with discontinuation of ICIs and initiating immunosuppression—glucocorticoids for hepatitis, and glucocorticoids plus rituximab for hemolytic anemia. Cancer outcomes were reported for eight patients; four (50%) achieved cancer remission or stable disease, whereas the other four had cancer progression.

Granulomatosis with polyangiitis (n = 8)

Three patients (43%) were receiving maintenance treatment for vasculitis (glucocorticoids, methotrexate, and rituximab) before initiating ICI therapy, and only one patient had active disease (sinusitis) at the time of ICI initiation. Four patients (50%) experienced a vasculitis flare. These flares occurred a median of 2.2 weeks following ICI initiation. none of the patients experienced de novo irAEs.

Case 9 (MD Anderson)

A patient in their 50s had a history of GPA, with prolonged respiratory symptoms, sinusitis, and eye inflammation, diagnosed in 2014 with biopsy of a lung nodule. The patient had received rituximab, cyclophosphamide and mycophenolate mofetil, with good response, and occasional intermittent flares treated with prednisone, and maintenance hydroxychloroquine for several years. In June 2016, melanoma of the scalp was diagnosed, with recurrence and lung metastases in July 2020. In August 2020, treatment with pembrolizumab was started; at that time the vasculitis was inactive and did not require treatment. Two months later, the patient was hospitalized with pleural effusion attributed to GPA flare, treated with prednisone, which resolved. However, renal function worsened, requiring dialysis and leading to permanent discontinuation of immunotherapy. Kidney biopsy showed 70% of glomeruli with sclerosis. The patient subsequently developed sinusitis and received rituximab. Afterward, acute unilateral ischemic optic neuritis developed with an MRI showing diffuse atrophy of the left optic nerve and was treated with pulse steroids. The vasculitis was eventually controlled with treatment and prednisone taper, but the patient died from cancer progression.

Case 17 (literature case report)

A patient in their 50s with bladder cancer and GPA had a disease flare with arthralgia, fatigue and cavitation of a pulmonary nodule after one dose of pembrolizumab. ¹⁹ The vasculitis had been inactive for several years and the patient was not receiving immunosuppression. The flare was treated initially with prednisone 60 mg daily. However, upon tapering the prednisone to 20 mg daily, recurrent reactive airway syndrome developed, requiring two doses of 1,000 mg of rituximab with resolution of symptoms.

Case 18 (literature case report)

A patient in their 60s had non–small cell lung cancer and GPA. ²⁰ A flare with acute kidney injury with proteinuria and hematuria occurred after one dose of pembrolizumab. Vasculitis activity and baseline treatment before ICI initiation were not reported. The flare was treated with glucocorticoids and four weekly doses of rituximab, with improvement.

Case 19 (literature case report)

A patient in their 60s with melanoma and prior GPA limited to sinus disease received pembrolizumab. ²¹ The patient experienced progression of the sinus‐limited GPA to a systemic form, with a flare characterized by headache, arthralgias, purpuric rash, proteinuria, and hemoptysis. At the time of initiation of immunotherapy, the patient had active GPA (chronic sinusitis) without baseline treatment. Treatment with pulses of methylprednisolone, rituximab, and a subsequent dose of infliximab, led to improvement of symptoms.

Four of the seven patients (57%) with available tumor outcome data had a partial tumor response or remained stable, and three (43%) had progressive disease.

Cutaneous vasculitis (n = 4)

Only one patient was receiving glucocorticoids (unspecified dose) for vasculitis maintenance before ICI, and one had active disease at the start of treatment. Two patients (50%) experienced a disease flare at a median of 18.5 weeks, and none developed de novo irAEs.

Case 12 (MD Anderson)

A patient in their 50s had leukocytoclastic vasculitis diagnosed in 2020 and well‐controlled HIV infection. Periodic vasculitis flares typically resolved with topical glucocorticoids, and dapsone was sometimes required. The patient was diagnosed with cervical squamous cell carcinoma in 2001, and remained disease‐free until recurrence, starting pembrolizumab in April 2023. The vasculitis was considered active at the time of ICI initiation, with a mild flare two weeks earlier that was treated with topical steroids only, without systemic immunosuppressive therapy. Two cutaneous vasculitis flares occurred in November 2023 and January 2024, which were self‐limited, only requiring topical steroids and intermittent dapsone. The patient was able to continue ICI therapy but died from infection in a pelvic mass in October 2024.

Case 14 (MD Anderson)

A patient in their 60s with cutaneous vasculitis diagnosed in 2013 was initially treated with prednisone and methotrexate and later on with mycophenolate mofetil. The patient subsequently remained stable on prednisone alone (2.5 mg daily). In 2013 the patient was diagnosed with chronic lymphocytic leukemia with Richter's transformation into lymphoma in 2017. Treatment with ibrutinib and nivolumab was initiated. The vasculitis had been controlled with low dose prednisone without flares for at least one year before ICI initiation. After five doses of nivolumab, skin vasculitis flares occurred and were treated with prednisone, up to 10 mg daily, with improvement. The patient eventually underwent stem cell transplantation.

Three patients had tumor progression or recurrence, and one remained disease‐free until last follow‐up.

Other vasculitis types (n = 4)

Four additional patients with various vasculitis were identified.

EGPA (literature case report)

A patient in their 80s with endometrial cancer and EGPA in remission, taking maintenance rituximab, was treated with pembrolizumab. No flare or de novo irAEs occurred. ²² Another patient in their 60s with EGPA in remission, stable on glucocorticoids, had melanoma and was treated with ipilimumab, and later on pembrolizumab. No vasculitis flares occurred but the patient developed irAE colitis although receiving ipilimumab. ²³

Behçet disease (literature case report)

A patient in their 40s with melanoma and active Behçet disease (oral ulcers) not receiving immunosuppression when starting ipilimumab did not develop flares or irAEs. ²⁴

Henoch–Schönlein purpura (MD Anderson)

A patient in their 60s with a history of Henoch–Schönlein purpura, inactive and without treatment, and liver cancer received pembrolizumab without vasculitis flare or irAEs.

Rechallenge with ICI therapy in patients with preexisting vasculitis after experiencing adverse events

Four of the patients at MD Anderson who developed a vasculitis flare (two with cutaneous vasculitis, one with GCA, and one with GPA), and one with colitis–irAE were rechallenged with ICI. None had a new flare or irAEs.

DISCUSSION

To our knowledge, this is the largest case series reporting the outcomes of patients with preexisting vasculitis who received ICI. Our findings show that about a third of the patients had a vasculitis flare following ICI therapy. Flares responded to therapy and generally did not require discontinuation of ICI.

The therapeutic success of ICIs in treatment of cancer stem from their ability to reinvigorate immune activation, primarily enhancing Th1 and CD8⁺ T cell responses essential for antitumor immunity. ²⁵ , ²⁶ However, this robust immune activation can also trigger inflammatory and autoimmune adverse events, posing significant risks for patients with preexisting autoimmune diseases. ⁷ Given the shared immune pathways between autoimmune diseases, including vasculitis, and ICI‐induced immune responses, particularly the activation of Th1, Th17, and CD8⁺ T cells, ²⁷ , ²⁸ , ²⁹ , ³⁰ concerns remain regarding the potential for disease flares or de novo irAEs in these patients following ICI initiation. Vasculitis can cause severe morbidity and mortality due to its potential to cause widespread inflammation and organ damage. Patients with preexisting vasculitis may be especially vulnerable to immune activation triggered by ICIs. Notably, the observation of expanded CD8⁺ T cell clones and increased infiltration of affected organs in patients experiencing irAEs highlight a mechanistic overlap between ICI‐induced toxicities and vasculitis exacerbation. ²⁵ , ²⁸ , ³⁰ , ³¹ , ³² , ³³ Moreover, the presence of disease‐specific autoantibodies in some patients with irAEs suggests a potential role for B cells in immune dysregulation, further complicating the potential for ICI‐related toxicities in this population. ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ These uncertainties contribute to hesitation in prescribing ICIs to patients with preexisting autoimmune diseases, particularly those with vasculitis, underscoring the need for further investigation into immune dynamics to refine risk stratification and optimize treatment strategies.

Despite prior reports on the safety of ICI therapy in patients with preexisting autoimmune diseases, data specific to those with vasculitis remain limited. ⁷ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ Our group previously conducted the first systematic literature review on this population, analyzing 123 patients from 49 publications. ⁷ This study found that 75% of patients experienced adverse events following ICI initiation, with 41% developing flares of their underlying autoimmune disease, 31% developing de novo irAEs, and 11% experiencing both adverse events. However, only four patients had preexisting vasculitis, limiting the ability to draw definitive conclusions. Subsequent case reports have provided limited insights, documenting both patients who experienced adverse events and those who tolerated ICIs without complications. ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ⁴⁶ , ⁴⁷ However, many of these reports lack critical information, including baseline autoimmune profile, ¹⁸ , ²³ , ²⁴ vasculitis disease activity and prior treatment, maintenance immunosuppressive therapy at ICI initiation, ²⁰ and cancer outcomes. ¹⁸ , ²¹ Therefore, our understanding of ICIs safety in this population remains significantly limited. To date, the true incidence of disease flares or de novo irAEs, and their relationship with key factors such as ICI type, vasculitis subtypes and disease activity, and immunosuppressive treatments at ICI initiation, remain unknown. Additionally, the impact of these adverse events and the need for escalation of baseline immunosuppression to manage toxicities on oncologic response and overall survival is unclear. More importantly, no data are currently available on the outcome of ICI rechallenge in this population, leaving a critical gap in understanding the long‐term feasibility of ICI therapy in patients with preexisting vasculitis.

In our study, which integrates institutional experience with a comprehensive literature review, we found close to one‐third of patients experienced vasculitis flares, a rate comparable to what has been reported in previous studies of patients with other preexisting autoimmune diseases. Notably, flares occurred in approximately one‐fourth of patients with GCA, half of those with GPA, and half of those with skin‐limited vasculitis. Interestingly, de novo irAEs were observed in 12% of patients, a rate similar to the reported 14.9% incidence of all‐grade and grade ≥3 irAEs in a recent systematic review. ⁴⁸ These findings suggest that much of the risk in patients with preexisting vasculitis relates to autoimmune disease exacerbation rather than de novo irAEs per se. Although patients with active vasculitis at ICI initiation were more likely to have a flare, the number of patients is too small for robust subgroup comparisons.

Truly, one might speculate that patients receiving anti‐CTLA‐4 would be more prone to de novo irAEs than those treated with anti‐PD‐1/PD‐L1, a pattern observed in ICI clinical trials. ⁴⁹ Nevertheless, our series show that vasculitis flares occurred most frequently with anti‐PD‐1/PD‐L1 therapies, and no flares were observed with ipilimumab. Given the small sample size and the retrospective nature of our study, we cannot draw definitive conclusion regarding ICI class‐specific risk for vasculitis flares, highlighting the need for prospective studies.

Notably, patients not receiving maintenance immunosuppression for their vasculitis at ICI initiation had a higher incidence of adverse events compared to those receiving therapy (50% vs 33%). Most patients who developed adverse events required initiation or escalation of immunosuppressive therapy, with good response. However, one patient reported in the literature had a stroke, which was attributed to GCA, and died from this complication. This highlights the importance of careful management of preexisting vasculitis during ICI therapy, as complications when they happen can be dire.

Overall, cancer progression remained high (52%). Although immunosuppressive therapy may reduce the risk and improve the outcome of adverse events, it could result in potential attenuation of ICI efficacy. Clinical trials are needed to determine the optimal immunosuppressive regimen that can effectively control the underlying autoimmune disease and mitigate de novo irAEs without compromising cancer outcomes. Ultimately, 44% of patients died because of cancer‐related causes, one death was attributed to a vasculitis flare, and no deaths were due to de novo irAEs.

Previous studies in patients with preexisting autoimmune diseases treated with ICIs reported permanent ICI discontinuation in 7% to 21% of those who experienced adverse events. ⁴¹ , ⁵⁰ , ⁵¹ In preexisting vasculitis, previously reported cases were either not rechallenged with ICI therapy or lacked information on rechallenge outcomes. In our center, four patients who had vasculitis flares and one who developed colitis after ICI therapy were successfully rechallenged without further adverse events. This suggests that ICI therapy may still be a viable option after initial toxicity, but agent selection may play a role in mitigating rechallenge risks. This warrants further investigation in clinical trials.

Our study is limited by its descriptive design and small sample size, which prevents more definite conclusions about safety in patients with specific types of vasculitis. Most of the patients did not have active vasculitis when initiating ICI, so we cannot compare the incidence of flares between those with active or inactive disease. As the study was retrospective, all the evaluations of disease activity or flares were based on the clinical notes (MD Anderson) or published descriptions of cases (literature). No standard measures of disease activity were reported for any of the cases. Finally, the rationale for treatment was not always clear. In two cases from the literature, a patient with GCA was treated with infliximab and one with GCA with cyclophosphamide, which are not recommended therapies for these conditions. ¹⁸ , ²¹ , ⁵² , ⁵³ These limitations highlight the need for prospective collaborative studies. Nevertheless, our report is, to our knowledge, the largest case series of patients with preexisting vasculitis, providing a summary of the best available evidence at this time to support clinical decision‐making. Although prospective studies and collaborative registries are necessary, the rarity of vasculitis means such efforts may take years. In the meantime, our findings offer useful context to guide current practice.

Our study provides the most comprehensive report to date, demonstrating that although nearly half of patients experience vasculitis flares or de novo irAEs, these events are generally manageable with immunosuppressive therapy. Moreover, successful rechallenge in few cases suggests that permanent ICI discontinuation may not always be necessary following toxicity. However, important questions remain regarding the optimal approach of immunosuppression that balances autoimmune disease control with ICI efficacy. Additionally, mechanistic studies are necessary to elucidate the immune pathways driving ICI‐related toxicity in patients with systemic autoimmune disease, such as vasculitis, ultimately guiding personalized treatment approaches to improve long‐term outcomes in this high‐risk population.

AUTHOR CONTRIBUTIONS

All authors contributed to at least one of the following manuscript preparation roles: conceptualization AND/OR methodology, software, investigation, formal analysis, data curation, visualization, and validation AND drafting or reviewing/editing the final draft. As corresponding author, Dr Suarez‐Almazor confirms that all authors have provided the final approval of the version to be published and takes responsibility for the affirmations regarding article submission (eg, not under consideration by another journal), the integrity of the data presented, and the statements regarding compliance with institutional review board/Declaration of Helsinki requirements.

Supporting information

Disclosure Form: