A thorough search was performed on various databases such as PubMed, Embase, and Web of Science Core Collection (WoS CC) using subject terms and keywords pertaining to neoplasms, immune checkpoint inhibitors, bullous pemphigoid, and case reports. The date for the literature search was October 20, 2025. The objective was to locate all pertinent English-language articles published since the establishment of each database. The specific search strategy can be found in the Supplementary Materials. This study has been registered with International Prospective Register of Systematic Reviews (PROSPERO) under registration number CRD420251172315. The detailed study flow diagram is presented in Figure 1.

EndNote was utilized for literature management. Following the deduplication of records from three databases, two investigators (CL and CYJ) independently screened titles and abstracts and performed quality assessments of the case reports. Inclusion criteria were as follows: research on tumor-associated ICI-BP must include published articles with full-text availability, specifically focusing on case reports or case series. The diagnosis must satisfy all of the following criteria. 1) Clinical evidence: The presence of rashes must be consistent with BP characteristics, such as pruritic wheals, vesicles, and bullae. 2) Histopathological evidence: Biopsy results must demonstrate subepidermal fissures, typically accompanied by eosinophilic infiltration. 3) Immunopathological evidence requires at least one of the following criteria: a) direct immunofluorescence must show linear deposition of IgG and/or C3 along the basement membrane. b) Indirect immunofluorescence must detect anti-basement membrane band antibodies in serum. c) Serological testing must identify circulating autoantibodies against BP180 and/or BP230 using ELISA or immunoblotting. Exclusion criteria included reviews, systematic reviews, books, editorials, or conference abstracts, cases with concurrent inflammatory or blistering dermatoses that could confound BP diagnosis (e.g., lichen planus, Grover’s disease, or psoriasis), BP onset before ICI therapy, and insufficient key information like tumor type or ICI agent used.

The study selection was independently and redundantly conducted by three reviewers (CL, CYJ, and ZSQ). The evaluation criteria comprised the following: 1) demographic information, 2) a documented history of ICI use, 3) a precise timeline detailing the onset of BP symptoms and diagnosis, 4) BP diagnosis confirmed through direct immunofluorescence (DIF) and/or serological methods, 5) a comprehensive description of the intervention or treatment outcomes, 6) a thorough account of clinical status following the intervention, 7) identification and description of adverse reactions (harm) or unforeseen events, and 8) a clear presentation of results. Studies that fulfilled all criteria were included in the analysis. In cases of disagreement, the third evaluator rendered the final decision. A detailed flowchart of the study is presented in Figure 1.

Three investigators (CL, CYJ, and ZSQ) independently reviewed the full texts of all included studies and performed data extraction using a standardized form. This process was followed by a cross-verification step to ensure accuracy. The extracted data encompassed the following domains: publication details (title, first author, and publication year), patient demographics and baseline characteristics (sex, age, past medical history, cancer type, and cancer stage), ICI treatment information (ICI type, specific agent, line of therapy, start date, concomitant anti-tumor therapies, and prior ICI exposure including types and agents), characteristics of ICI-BP occurrence (time of initial skin symptoms, description of presentation, time of BP diagnosis, mucosal involvement, disease severity, and time from ICI initiation to BP onset), management of ICI-BP (initial treatment, initial treatment response, need for treatment escalation, all BP-related treatments received, effective treatment modalities, and time from BP diagnosis to disease control), and other relevant outcomes (ICI discontinuation status at BP onset, impact of BP on ICI therapy, tumor status during BP treatment and at final follow-up, ICI rechallenge attempts, ICI-BP recurrence, and other concurrent immune-related adverse events). All extracted data were compiled and managed in a centralized online Excel database accessible to all co-authors for review and reference.

Descriptive statistics summarized the demographic and clinical characteristics of the cases. Quantitative variables were reported as median and range (minimum to maximum values), and categorical variables were presented as counts and percentages. The Mann–Whitney U test was employed to compare continuous variables across groups, while the chi-square test or Fisher’s exact test was utilized for comparing categorical variables between groups. All analyses were conducted using the SPSS 26.0 software, and a p-value <0.05 was considered statistically significant.

Table 1 presents the key demographics of the participants. The majority was male (87/116, 75.00%), with a median age of 71.5 years (range, 24–90). Individuals aged 60 years or older made up 87.93% of the group. The most common primary tumor types were melanoma (41.38%), lung cancer (21.56%), renal cell carcinoma (14.66%), and urothelial cancer (4.31%).

The statistical analysis of treatment lines indicated that first-line therapy was the most prevalent, accounting for 50 cases (43.10%), followed by second-line therapy with 17 cases (14.66%), third-line and above with eight cases (6.90%), second-line therapy again with 13 cases (10.0%), and adjuvant therapy with seven cases (6.03%).

The most frequently utilized ICI type was anti-PD-1, representing 84.48% (98/116), followed by combination therapies in 10 cases (8.62%), which included anti-PD-1 and anti-CTLA-4 in seven cases (10–16), anti-PD-L1 and anti-CTLA-4 in one case (17), anti-PD-1 and anti-LAG3 in one case (14), and anti-PD-1 and anti-ILT4 in one case (18). Other ICI types included anti-PD-L1 in five cases (4.31%) (19–23), anti-CTLA-4 in two cases (1.72%) (24, 25), and anti-PD-1/TGF-β in one case (26). The ICI regimen comprised nivolumab in 49 cases (42.24%), pembrolizumab in 41 cases (35.34%), ipilimumab plus nivolumab in seven cases (6.03%), atezolizumab in four cases (3.45%), cemiplimab in three cases (2.59%), sintilimab in three cases (2.59%), ipilimumab in two cases (1.72%), and durvalumab combined with tremelimumab, nivolumab with relatlimab, pembrolizumab with MK4830, bintrafusp alfa, karelizumab, durvalumab, and tislelizumab in one case each (0.86%).

ICI therapy history was present in 12.93% of cases, with prior use of ipilimumab in seven cases (23, 27–31), pembrolizumab in three cases (24, 26, 32), nivolumab in two cases (25, 33), a combination of ipilimumab and nivolumab in two cases (10, 34, 35), and a combination of ipilimumab and pembrolizumab in one case.

Of the total cases, 13.79% (16/116) had finished ICI treatment when diagnosed with BP. Concurrent use of other tumor therapies represented 17.24% of cases, with 6.03% combined with targeted therapy (12, 16, 36–40), 5.17% with radiotherapy (32, 35, 41–44) or chemotherapy (16, 45–49), and 0.86% with a combination of targeted therapy and endocrine therapy (16). The most common pre-existing medical conditions were hypertension (12.93%) and type 2 diabetes (8.62%). A history of dermatological conditions was documented in 3.45% of cases, including psoriasis in two cases and Grover’s disease in two cases. Additionally, three cases had previous autoimmune diseases, such as Crohn’s disease, autoimmune hypophysitis, and autoimmune thyroid disease.

Pruritus occurred in 67.24% of patients, erythema or rash in 80.17%, and mucosal involvement in 19.83% (50). The severity of BP was categorized according to Common Terminology Criteria for Adverse Events (CTCAE) 5.0, with mild cases (grades 1–2) accounting for 47.41% and moderate-to-severe cases (grades 3–4) for 44.83%.

An analysis of data from 83 cases showed that the median onset time of ICI-BP-related skin symptoms was 21.4 weeks (range, 0–283 weeks), with an average of 41.21 weeks (potentially skewed rightward due to some larger values inflating the mean). Some symptoms manifested within 1 week of ICI initiation, while others appeared more than 36 months after treatment. Nevertheless, diagnosis often faced delays, with a median delay of 4.65 weeks (range, 0–115.7 weeks) and an average delay of 12.38 weeks.

In addition to BP, 26 cases reported other concurrent adverse reactions, including dysthyroidism (3.45%), vitiligo (3.45%), and diarrhea (3.45%), which are common immunotherapy-related adverse effects.

Histopathological evidence of BP was a prerequisite for inclusion. Moreover, 92.23% of cases were diagnosed via DIF, while 18.97% relied on indirect immunofluorescence (IIF) for diagnosis. Among these cases, 53.45% tested positive for BP180 and 10.34% for BP230.

In the selection of treatment regimens, 94.83% of patients were administered corticosteroids, in either systemic or topical forms. Among these patients, 83.62% received systemic corticosteroid therapy (CS), while 74.14% were treated with topical CS. Additionally, 31.03% were prescribed tetracycline antibiotics (41, 51, 52) and is often used in combination with nicotinamide (53, 54), and 25.86% received biological therapies, which included omalizumab (4) (47, 55–57), rituximab (11) (42, 58–60), dupilumab (17) (61–63), and infliximab (1) (43). Furthermore, 16.38% of patients were given immunosuppressants (64, 65), 15.52% received nicotinamide, 10.34% were treated with dapsone (13, 15, 27, 34, 47, 66–71), 9.48% received antihistamines (72, 73), 5.17% were administered intravenous immunoglobulin (IVIG) (46, 47, 69, 74–76), and 17.24% underwent plasma exchange (55, 77).

Of the patients, 54.31% experienced treatment escalation. We recorded the efficacy of all medications administered throughout the treatment course. Oral glucocorticoids constituted the fundamental basis of therapy and were the most frequently reported effective treatment regimen (n = 82). Topical potent or ultrapotent glucocorticoids (n = 36) were often used in conjunction with them for primary lesion management. When reducing glucocorticoid dosage or as an alternative therapy, tetracyclines (doxycycline/minocycline, n = 20) combined with nicotinamide (n = 8) emerged as the most commonly adopted regimens, reflecting a preference for non-steroidal immunomodulatory strategies. Traditional methotrexate (n = 11) also served as a common agent for glucocorticoid tapering. The treatment escalation pathway was evident in moderate-to-severe, refractory, or glucocorticoid-dependent cases. Dupilumab (n = 16) was the most frequently reported biologic agent, highlighting its central role in the current management of refractory ICI-BP. Rituximab (n = 9) was utilized for more severe or treatment-resistant cases. A minority of critical cases received intravenous immunoglobulin (n = 4) or plasma exchange (n = 2). Notably, intravenous methylprednisolone (n = 9) was administered to some patients requiring rapid disease control. Other immunosuppressants, such as mycophenolate mofetil and cyclosporine, along with omalizumab, were also reported sporadically.

Following treatment with ICI-BP, 49.14% of patients achieved a complete response (CR) (78–81), while 43.10% attained a partial response (PR) (82, 83), resulting in a disease control rate of 92.24%. The median duration from the initiation of treatment to disease control was 6 weeks (range, 1–167 weeks), with a mean duration of 12.57 weeks.

At the time of the occurrence of BP, 13.79% of patients had completed ICI treatment. Among those affected by BP, 17.24% continued ICI without interruption, 2.59% maintained ICI until achieving tumor response, 15.52% interrupted treatment and later resumed, and 45.56% interrupted and did not resume ICI. Eighteen patients restarted ICI treatment after experiencing BP, and seven of these patients subsequently experienced a recurrence of BP following the ICI restart.

We conducted a statistical analysis of tumor progression throughout the follow-up period. Among the 55 cases with reported tumor outcomes, the disease control rate was 67.27%, comprising 14 cases of CR (84–87), five cases of PR (88), and 18 cases of stable disease (SD) (89, 90). Additionally, there were seven cases of progressive disease (PD) and 11 cases resulting in death.

Our study cohort validates common features of ICI-BP patients, including a prevalence among elderly male patients, often associated with melanoma and lung cancer, primarily induced by anti-PD-1 agents. This demographic and cancer type distribution closely mirrors actual ICI usage trends and findings from an earlier analysis of the FDA Adverse Event Reporting System (FAERS) database (99). Notably, while the majority of incidents occurred during initial treatment, occurrences were spread throughout all stages of therapy, highlighting the potential for BP to emerge as a side effect at any phase of ICI treatment, underscoring the need for ongoing vigilance among healthcare providers.

Regarding the temporal relationship between ICI therapy and BP onset, our findings reveal a median time of 21.4 weeks post-ICI initiation for BP onset, with a broad range from 0 to 283 weeks. This variability highlights the occurrence of ICI-BP as either an “early” event, sometimes within 1 week of drug administration, or a “delayed” immune-related adverse event. Additionally, 13.85% of patients received a BP diagnosis only after completing ICI therapy, indicating persistent and potentially irreversible immune system overactivation. This emphasizes the crucial need to investigate prior medication history before treatment initiation and to conduct prolonged post-discontinuation surveillance.

A critical discovery is the notable and common delay in diagnosing ICI-BP, with a median of 4.7 weeks and a mean of 12.4 weeks. This timeframe is briefer than the usual diagnostic delay observed in classic BP (100), possibly due to the vigilant monitoring of patients on ICIs, leading to prompt attention to any related skin manifestations. Conversely, extended diagnostic delays over 4 months in classic BP frequently pertain to lesions localized to a single body area. This discovery conveys a significant clinical alert: in patients undergoing ICI treatment, especially elderly men, the presence of persistent, treatment-resistant itching or unusual skin rash should strongly indicate prodromal BP, necessitating prompt investigations like skin biopsy for a conclusive diagnosis. The average diagnostic delay of 94.74 days necessitates reduction, as minimizing this timeframe is essential for promptly commencing tailored therapy, enhancing patient well-being, and potentially impacting prognosis.

Additionally, 44.83% of cases in this cohort were classified as CTCAE grades 3–4, with 19.83% exhibiting mucosal involvement, underscoring that ICI-BP is frequently a moderate-to-severe condition necessitating systemic immunosuppressive treatment. Hypertension and diabetes were the prevailing pre-existing comorbidities (71, 95). Chronic metabolic diseases often coincide with chronic, low-grade inflammatory conditions, potentially impacting immune system homeostasis (101, 102). This interplay between metabolic inflammation and autoimmune responses triggered by ICI may influence BP susceptibility or severity, necessitating further mechanistic exploration. Although the prevalence of patients with documented dermatological conditions (7.76%, mainly psoriasis) and autoimmune diseases (2.59%) in this cohort was modest, their presence carries significant clinical implications. These patients already demonstrate immune dysregulation (103–105), and ICI treatment could exacerbate immune tolerance disruption, leading to new autoimmune manifestations like BP. This underscores the critical importance of conducting a comprehensive medical history review (especially concerning dermatological and autoimmune disorders) before commencing ICI therapy to identify high-risk groups and facilitate enhanced monitoring.

This study offers a comprehensive analysis of the treatment patterns, clinical progression, and influence on oncologic results of ICI-BP, uncovering intricate management approaches and eventual outcomes of this immune-related adverse event.

This study, through stratified analysis, provides an in-depth revelation of the heterogeneity in ICI-BP, offering key evidence for clinical risk prediction and individualized management.