ABSTRACT
There is an unmet need to predict immune‐related adverse events (irAEs) in patients receiving immune checkpoint inhibitors. This study aimed to examine whether eosinophils can predict irAEs in metastatic urothelial carcinoma (mUC). In this multicenter cohort study, we retrospectively examined eosinophil before (baseline sample) and 3 weeks after treatment (3‐week sample) in 124 patients with mUC treated with pembrolizumab between January 2018 and October 2024. We divided the patients into two groups depending on whether they had experienced an irAE (irAE group) or not (non‐irAE group). The variation in eosinophil proportion from baseline to 3 weeks was significantly higher in the irAE group than in the non‐irAE group (non‐irAE, 2.4%–3.1%; irAE, 2.2%–4.3%; p < 0.05). Notably, the variation in eosinophils from baseline to 3 weeks in patients with bladder cancer was significantly higher in the irAE group than in the non‐irAE group (non‐irAE, 2.5%–2.8%; irAE, 1.9%–4.1%; p < 0.05), but not in upper urinary tract urothelial cancer (non‐irAE, 2.2%–3.7%; irAE, 2.8%–4.4%; p = 0.97). The optimal cutoff value for eosinophils in patients with bladder cancer against the occurrence of any‐grade irAEs was 5.0% (area under the curve = 0.655). Multivariable analyses showed that eosinophil levels ≥ 5.0% increased the risk of any‐grade irAEs in patients with bladder cancer (odds ratio 3.61, 95% confidence interval, 1.00–13.0). An eosinophil proportion ≥ 5.0% may be an effective biomarker for evaluating pembrolizumab‐induced irAEs in patients with mUC, particularly those with bladder cancer.
Keywords: bladder cancer, eosinophil, immune‐related adverse event, pembrolizumab, urothelial carcinoma
1. Introduction
Metastatic urothelial carcinoma (mUC) has a poor prognosis and a 5‐year survival rate of approximately 8% [1]. To date, mainly platinum‐based chemotherapy has been used for mUC; however, the overall survival (OS) of patients receiving platinum‐based chemotherapy is approximately 1 year [2]. In this context, pembrolizumab has markedly changed the treatment strategies and improved the prognosis of patients with mUC [3, 4]. For instance, the KEYNOTE‐045 study demonstrated that OS in patients who received pembrolizumab was significantly longer than that in the chemotherapy group (median OS, 10.3 and 7.4 months, respectively) [3]. Additionally, the EV‐302 trial demonstrated that OS in patients who received pembrolizumab combined with enfortumab vedotin therapy was significantly longer than that in patients who received platinum‐based chemotherapy (median OS, 31.5 and 16.1 months, respectively), although the control arm included avelumab [4].
Immune‐related adverse events (irAEs) are caused by immune checkpoint inhibitors, including pembrolizumab, and severe irAEs can lead to fatalities. In practice, 60.9% of patients in the KEYNOTE‐045 study experienced any grade of irAEs, and 15.0% experienced severe irAEs [3]. As irAEs occur when the immune system becomes nonspecifically activated and can affect most organ systems, it is difficult for physicians to predict when and in which organ systems irAEs will occur. Therefore, establishing effective biomarkers for irAEs is necessary.
Several studies have identified biomarkers for the onset of irAEs in various cancers [5, 6, 7, 8]. We previously clarified the association between eosinophils and irAE occurrence [6, 7, 8]. However, whether eosinophils are associated with irAE occurrence in patients with mUC treated with pembrolizumab remains unclear. Therefore, we investigated whether eosinophils are effective biomarkers for irAE occurrence in patients with mUC treated with pembrolizumab.
2. Materials and Methods
2.1. Patients and Treatment
We retrospectively collected clinical data from 124 patients who had been diagnosed with mUC. Patients with mUC who had undergone ≥ 1 course of treatment with pembrolizumab monotherapy (200 mg/body on Day 1 every 3 weeks) between January 2018 and October 2024 at six hospitals in Japan (Nagoya City University Hospital, Anjo Kosei Hospital, Konan Kose Hospital, Gamagori Hospital, Inabe General Hospital, and Daido Hospital) were included in the study. Experienced pathologists determined the histological subtypes. All data, including demographic information, pathological type, and peripheral blood parameters, were obtained from the patients' medical records. This study was approved by an appropriate Ethics Committee (approval number 60‐18‐0060) and was conducted in accordance with the tenets of the Declaration of Helsinki (updated in 2024). In addition, this study was conducted in accordance with Japan's Ethical Guidelines for Life Science and Medical Research Involving Human Subjects. Given the retrospective nature of this study, the requirement for patient consent was waived. Nonetheless, an opt‐out procedure was adopted in lieu of written informed consent. An opt‐out notice clearly stating the study's purpose, data to be used (such as age, sex, and blood test results), anonymization of information, and contact point was posted on the website of the authors' affiliated institution. Participants were guaranteed the opportunity to refuse the use of their data for this study through the contact point.
2.2. Immune‐Related Adverse Events
Patients treated with pembrolizumab were divided into two groups—irAE and non‐irAE—depending on whether they had experienced an irAE. IrAEs were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0.
2.3. Data Collection
To assess whether eosinophil is an effective biomarker, we examined eosinophil proportion and count in blood samples collected before one course of treatment with pembrolizumab (baseline sample) and 21 days after treatment with one course of pembrolizumab (3‐week sample). The timing of eosinophil proportion and count measurement was determined based on a previous study [7] and the timing of blood sampling in daily clinical practice.
2.4. Number of Patients Analyzed in the Current Study
Table S1 shows the number of patients in the non‐irAE and irAE groups who were analyzed for eosinophil proportion at baseline and 3 weeks. We excluded 38 patients in the non‐irAE and irAE groups owing to missing eosinophil data from 3‐week samples. Table S2 shows the number of patients in the non‐irAE and irAE groups according to primary tumor site, analyzed for eosinophil proportion at baseline and 3 weeks. Based on this, we excluded seven patients: six diagnosed at both sites and one at the prostate site. Univariable and multivariable logistic regression analyses were performed using 3‐week samples of patients with bladder cancer.
2.5. Statistical Analysis
Statistical significance was set at p < 0.05. The changes in eosinophil proportion and count over time (from baseline to 3 weeks) were analyzed using a two‐factor repeated measurement analysis (analysis of variance [ANOVA]). The optimal cutoff points for potential peripheral blood biomarkers to evaluate any grade of irAEs were determined using receiver operating characteristic (ROC) curve analysis. Univariable and multivariable logistic regression analyses were used to assess the risk factors for any grade of irAEs. Analyzed factors included age, gender, histological subtype, Eastern Cooperative Oncology Group Performance Status (ECOG‐PS), median duration of pembrolizumab treatment, treatment history before pembrolizumab, history of autoimmune disease, and eosinophil proportion. Multivariable analysis included age, histological subtype, ECOG‐PS, median duration of pembrolizumab treatment, and eosinophil proportion as potential risk factors associated with irAE occurrence. Statistical analyses were performed using the EZR software (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [9].
3. Results
3.1. Patient Characteristics
The demographic characteristics of the study population are shown in Table 1. The sex ratio differed significantly between the non‐irAE and irAE groups. Age, primary tumor site, histological subtype, and ECOG‐PS did not differ between the two groups. Table S3 shows comorbidities and disease histories of the patients. Two patients had a history of autoimmune disease (remitting seronegative symmetrical synovitis with pitting edema syndrome and ulcerative colitis). No patients were on continuous systemic steroid use before pembrolizumab treatment. Treatment histories prior to pembrolizumab initiation included platinum‐based therapy in 108 patients, taxane‐based therapy in 19 patients, avelumab in four patients, and other therapies in five patients (Table S4).
TABLE 1.
Clinical features of patients.
| Characteristics | All patients | Non‐irAE | irAE | p |
|---|---|---|---|---|
| Total | 124 (100) | 84 (68.0) | 40 (32.0) | |
| Age, n (%) | 0.58 | |||
| < 65 years | 22 (17.7) | 17 (20.2) | 5 (12.5) | |
| ≥ 65 years | 99 (79.8) | 65 (77.4) | 34 (85.0) | |
| Not evaluable | 3 (2.4) | 2 (2.4) | 1 (2.5) | |
| Gender, n (%) | < 0.05 | |||
| Male | 102 (82.3) | 65 (77.4) | 37 (92.5) | |
| Female | 22 (17.7) | 19 (22.6) | 3 (7.5) | |
| Primary tumor site, n (%) | 0.28 | |||
| Bladder | 72 (58.1) | 52 (61.9) | 20 (50.0) | |
| Upper urinary tract | 45 (36.3) | 26 (31.0) | 19 (47.5) | |
| Bladder + Upper urinary tract | 6 (4.8) | 5 (6.0) | 1 (2.5) | |
| Prostate | 1 (0.8) | 1 (1.2) | 0 (0.0) | |
| Histological subtype, n (%) | 0.52 | |||
| Urothelial carcinoma | 109 (87.9) | 74 (88.1) | 35 (87.5) | |
| Urothelial carcinoma + another subtype | 5 (4.0) | 3 (3.6) | 2 (5.0) | |
| Urothelial papilloma | 1 (0.8) | 0 (0.0) | 1 (2.5) | |
| Not evaluable | 9 (7.3) | 7 (8.3) | 2 (5.0) | |
| ECOG‐PS, n (%) | 0.96 | |||
| 0–1 | 86 (69.4) | 59 (70.2) | 27 (67.5) | |
| ≥ 2 | 15 (12.1) | 10 (11.9) | 5 (12.5) | |
| Not evaluable | 23 (18.5) | 15 (17.9) | 8 (20.0) | |
Abbreviation: ECOG‐PS, Eastern Cooperative Oncology Group‐Performance Status.
3.2. Frequency and Timing of irAE Occurrence
The irAE profiles are shown in Table 2. A total of 47 irAEs were observed. Grade 1 irAEs were observed in 31.9% of the patients (15 events), Grade 2 irAEs in 44.7% (21 events), and Grade ≥ 3 irAEs in 23.4% (11 events). The most common irAEs were gastrointestinal (25.5%, 12 events), followed by endocrine (19.1%, nine events), skin (19.1%, nine events), and pulmonary irAEs (14.9%, seven events); other irAEs accounted for 21.3% (10 events).
TABLE 2.
Profile of irAEs.
| Total | Grade ≤ 2 | Grade ≥ 3 | |
|---|---|---|---|
| Profile of irAEs, n (%) | 47 (100) | 36 (76.6) | 11 (23.4) |
| Gastrointestinal | 12 (25.5) | 7 | 5 |
| Endocrine | 9 (19.1) | 6 | 3 |
| Skin | 9 (19.1) | 9 | 0 |
| Pulmonary | 7 (14.9) | 7 | 0 |
| Others | 10 (21.3) | 7 | 3 |
The median duration of any‐grade irAEs after the initial pembrolizumab treatment was 66 days (range, 2–473 days). Among the 47 irAEs, 19.1% (nine events) occurred within the first 21 days of treatment, and 80.9% (38 events) occurred at least 21 days after treatment (Table 3).
TABLE 3.
Occurrence of irAEs of any grade over time.
| Time from initiation of pembrolizumab to irAEs occurrences | |||
|---|---|---|---|
| Total | 0–21 days | 21 ~days | |
| Events, n (%) | 47 (100) | 9 (19.1) | 38 (80.9) |
3.3. Association Between Variations in Eosinophil Proportion and irAE Occurrence
To elucidate whether eosinophil proportion was associated with irAE occurrence, we assessed the variations in eosinophil proportion and count from baseline to 3 weeks after pembrolizumab treatment. This variation was significantly higher in the irAE group compared with that in the non‐irAE group (irAE: 2.2%–4.3%; non‐irAE: 2.4%–3.1%; time, p < 0.05; interaction, p < 0.05; Figure 1A).
FIGURE 1.
Analysis of eosinophil proportion from baseline to 3 weeks after treatment. (A) Graph showing the variation of the eosinophil proportion from baseline to 3 weeks in patients with metastatic urothelial carcinoma (baseline samples: non‐irAE, n = 84, and irAE, n = 40; 3‐week samples: non‐irAE, n = 57, and irAE, n = 29). (B and C) Graph showing the variation of eosinophil proportion from baseline to 3 weeks in patients with (B) upper urinary tract urothelial cancer (baseline samples: non‐irAE, n = 26, and irAE, n = 19; 3‐week samples: non‐irAE, n = 19, and irAE, n = 12) or (C) bladder cancer (baseline samples: non‐irAE, n = 52, and irAE, n = 20; 3‐week samples: non‐irAE, n = 36, and irAE, n = 16). (D) Graph showing the variation of eosinophil proportion from baseline to 3 weeks in the irAE group with bladder cancer (baseline samples: Grade ≤ 2, n = 14, and Grade ≥ 3, n = 6; 3‐week samples: Grade ≤ 2, n = 11, and Grade ≥ 3, n = 5). Results are presented as the mean ± SD. (A–D) Two‐factor repeated measurement analysis (analysis of variance [ANOVA]). irAE, immune‐related adverse event.
Next, we examined whether the variation in eosinophil proportion differed according to the primary tumor site, such as upper urinary tract urothelial or bladder cancer. The variation in eosinophil proportion from baseline to 3 weeks in patients with upper urinary tract urothelial cancer was not significantly different between the non‐irAE and irAE groups (non‐irAE: 2.2%–3.7%; irAE: 2.8%–4.4%; time, p < 0.05; interaction, p = 0.97; Figure 1B). Notably, the variation in eosinophil proportion between the baseline and 3‐week samples from patients with bladder cancer was significantly higher in the irAE group than in the non‐irAE group (irAE: 1.9%–4.1%; non‐irAE: 2.5%–2.8%; time: p < 0.05; interaction: p < 0.05; Figure 1C). In contrast, the variation in eosinophil count from baseline to 3 weeks did not significantly differ between the non‐irAE and irAE groups (Figure S2A–C). Given the association between eosinophil proportion and irAE occurrence, subsequent analyses were performed using eosinophil proportion as the primary exploratory biomarker. Further, we performed an analysis excluding nine patients who experienced irAEs by 3 weeks after pembrolizumab treatment. Consistent with the results presented in Figure 1B,C, while eosinophil proportion in patients with upper urinary tract urothelial cancer was not significantly different between the groups (non‐irAE: 2.2%–3.8%; irAE: 2.8%–3.5%; time, p = 0.09; interaction, p = 0.52; Figure S1A), eosinophil proportion in patients with bladder cancer was significantly higher in the irAE group than in the non‐irAE group (non‐irAE: 2.5%–2.8%; irAE: 2.0%–4.1%; time, p < 0.05; interaction, p < 0.05; Figure S1B). Furthermore, two patients with pre‐existing autoimmune diseases (remitting seronegative symmetrical synovitis with pitting edema syndrome and ulcerative colitis) were excluded. Nonetheless, the variation in eosinophil proportion remained significantly higher in the irAE group than in the non‐irAE group (Figure S3A–C). These results indicate that the eosinophil proportion in bladder cancer 3 weeks after treatment was upregulated by any grade of irAEs.
The variation in eosinophil proportion was not associated with any irAE category in an exploratory analysis of irAE categories (Figure S4A–E). We further examined whether the variation in eosinophil proportion differed according to irAE severity. There was no significant association between the variation in eosinophil proportion and the severity of irAEs in patients with bladder cancer (Grade ≤ 2, 1.7%–3.8%; Grade ≥ 3, 2.2%–4.6%; time: p < 0.05; interaction: p = 0.82; Figure 1D). These results indicate that the severity of irAEs did not change in relation to the variation in eosinophil proportion from baseline to 3 weeks.
3.4. Eosinophils 3 Weeks After Treatment Are Associated With irAE Occurrence in Bladder Cancer
The optimal cutoff value for eosinophil proportion in 3‐week samples to differentiate the occurrence of any grade of irAEs in patients with bladder cancer was 5.0%. ROC curve analysis revealed an area under the curve of 0.655, a 95% confidence interval (CI) of 0.48–0.82, a sensitivity of 0.43, and a specificity of 0.88 (Figure 2). Univariable and multivariable logistic regression analyses showed that an eosinophil proportion of ≥ 5.0% in 3‐week samples was an independent risk factor for the development of irAEs of any grade in patients with bladder cancer (univariable logistic regression analysis: odds ratio (OR) = 6.22, 95% CI = 1.48–26.1; multivariable logistic regression analysis: OR = 3.61, 95% CI = 1.00–13.0; p < 0.05; Table 4).
FIGURE 2.
Optimal cutoff value of eosinophil proportion. Receiver operating characteristic curve analysis of eosinophil proportion in a 3‐week sample for the occurrence of any‐grade irAEs in patients with bladder cancer.
TABLE 4.
Univariate and multivariate logistic regression analysis of risk factors for the occurrence of any grade irAEs in patients with bladder cancer.
| Univariate | Multivariate | |||||
|---|---|---|---|---|---|---|
| OR | 95% CI | p | OR | 95% CI | p | |
| Age: ≥ 65 years | 2.00 | 0.37–10.7 | 0.42 | 1.19 | 0.29–4.88 | 0.81 |
| Gender: male | 2.20 × 107 | 0‐inf | 0.99 | |||
| Histological subtype: urothelial carcinoma | 0.88 | 0.07–10.5 | 0.92 | 0.34 | 0.04–2.55 | 0.29 |
| ECOG‐PS: ≥ 2 | 0.67 | 0.24–1.85 | 0.44 | 2.56 | 0.56–11.8 | 0.23 |
| Duration of pembrolizumab treatment: ≥ 125 day | 2.08 | 0.62–6.96 | 0.23 | 1.52 | 0.51–4.57 | 0.45 |
| Prior treatment: platinum‐based chemotherapy | 0.50 | 0.03–8.56 | 0.63 | |||
| Prior treatment: taxane‐based chemotherapy | 1.73 | 0.39–7.63 | 0.47 | |||
| Prior treatment: avelumab | 1.16 × 10−7 | 0‐inf | 0.99 | |||
| Comorbidities: autoimmune disease | 3.80 × 10−7 | 0‐inf | 0.99 | |||
| Proportion of eosinophils: ≥ 5.0% | 6.22 | 1.48–26.1 | < 0.05 | 3.61 | 1.00–13.0 | < 0.05 |
Abbreviations: CI, confidence interval; ECOG‐PS, Eastern Cooperative Oncology Group‐Performance Status; OR, odds ratio.
4. Discussion
We found that an elevated eosinophil proportion 3 weeks after pembrolizumab treatment was associated with the occurrence of irAEs of any grade in patients with bladder cancer.
The KEYNOTE‐045 study showed that the median time to the onset of the first irAE induced by pembrolizumab was 107 days (range, 1–647 days). Consistent with the findings of this large‐scale clinical trial, the current study showed that the median time of any‐grade irAE occurrence was 66 days (range, 2–473 days). In contrast, the median time to the onset of the first AEs induced by chemotherapy, such as paclitaxel and docetaxel, in the KEYNOTE‐045 study was 21 days (range: 1–70 days) [3]. These data demonstrate that the time to the onset of the first irAE induced by pembrolizumab was longer than that induced by chemotherapy, making it difficult to predict the occurrence of irAEs. Interestingly, our data showed that increased eosinophil levels 3 weeks after treatment reflected the onset of irAEs of any grade in patients with bladder cancer. As 80.9% of patients experienced irAEs of any grade at least 21 days after treatment (Table 3), our approach utilizing eosinophils may be an effective criterion for diagnosing irAEs. Furthermore, evaluation of eosinophils at Week 3 is a viable and readily applicable approach in clinical practice, as the pembrolizumab administration is scheduled at 3‐week intervals.
Eosinophils are innate immune cells involved in the elimination of infections and autoimmune disorders [10]. Recently, several studies have reported an association between eosinophils and the efficacy of immune checkpoint inhibitors, including the reports on mUC [6, 7, 11, 12, 13, 14, 15, 16, 17]. For example, in patients with mUC treated with pembrolizumab, a higher eosinophil count was associated with better survival [13]. After the administration of immune checkpoint inhibitors, the production of eosinophils in the bone marrow is stimulated by interleukin (IL)‐5 secreted by CD4 + T cells, and eosinophils accumulate in the peripheral blood [18]. Peripheral blood eosinophils infiltrate tumors via IL‐33 secreted by CD4+ T cells [18]. Tumor‐infiltrating eosinophils interact with antitumor immune molecules, such as CD8+ T cells, and exhibit antitumor effects [18]. Although the mechanisms by which patients with mUC receiving pembrolizumab exhibit increased eosinophil levels are unclear, this study indicates that the interaction between CD4+ and CD8+ T cells could be a potential mechanism of eosinophil accumulation in patients treated with immune checkpoint inhibitors.
Several studies have reported eosinophils as biomarkers for predicting irAEs [6, 7, 19, 20, 21, 22]. However, Furubayashi et al. reported that eosinophils had no significant relationship with the incidence of irAEs in patients with mUC treated with pembrolizumab [14]. Contrarily, we demonstrated an association between eosinophils and irAE occurrence. These conflicting results may reflect differences in patient characteristics and the analytical methods used. However, previous studies examined the association between eosinophils and irAE occurrence without subdividing the primary tumor sites. We found that fluctuations in eosinophil levels differed between patients with upper urinary tract cancer and those with bladder cancer when the primary tumor site was subdivided. To our knowledge, this is the first study to report that eosinophil elevation is associated with any‐grade irAEs, especially in patients with bladder cancer.
An increased level of eosinophils is also associated with better immune checkpoint inhibitor efficacy and prognosis [6, 7, 11, 12, 13, 14, 15, 16, 17]. For example, Furubayashi et al. showed that a high eosinophil count is associated with better prognosis in patients with mUC undergoing pembrolizumab treatment [13]. Considering this, we believe that physicians should not discontinue immune checkpoint inhibitor treatment due to concerns about increased irAE incidence in cases of increased eosinophils. The current study emphasizes the importance of monitoring changes in eosinophil proportion, as confirming an elevated eosinophil proportion may enable early diagnosis of irAEs and intervention when symptoms appear.
mUC occurs throughout the lumen of the urinary tract, including the upper urinary tract (renal pelvis and ureter) and bladder. As the histological subtype of upper urinary tract cancer is UC, approximately 90% or more of patients are treated for upper urinary tract cancer according to the treatment of bladder cancer [23]. However, recent comprehensive studies have shown differential effects of immune checkpoint inhibitors in bladder and upper urinary tract cancers [3, 24, 25, 26]. The KEYNOTE‐045 study showed that the hazard ratio (HR) for OS was 0.77 for bladder cancer and 0.53 for upper urinary tract cancer [3]. For patients with resected UC treated with nivolumab, the HR for progression‐free survival was 0.62 for bladder cancer, 1.23 for renal pelvis cancer, and 1.56 for ureter cancer [25]. The HR for OS in patients with bladder cancer who were treated with avelumab after platinum‐based chemotherapy was 0.62, and that in patients with upper urinary tract cancer was 0.89 [26]. Additionally, several studies have reported that UC has genetic alterations associated with the effects of immune checkpoint inhibitors [27, 28]. DNA sequencing data have shown that mismatch repair‐related genes are altered more in upper urinary tract cancer than in bladder cancer. In contrast, other DNA damage repair‐related genes are altered more in bladder cancer [29]. Although it is unclear whether these differences in genetic alterations and clinical outcomes reflect the differences in eosinophil variations between the upper urinary tract and bladder cancers, several studies have explained the association between them.
This study had some limitations. First, a small number of participants were included in this study. Second, this study was designed based on previous reports demonstrating the clinical relevance of eosinophil proportion in patients treated with immune checkpoint inhibitors [6, 7, 8, 13, 14, 15]. Therefore, eosinophil proportion was used as the primary analytical variable. However, variations in the total white blood cell count can influence eosinophil proportion. Third, the number of irAE events within individual organ‐specific categories was limited, and overlapping irAEs were observed in some patients. Therefore, the findings regarding organ‐specific associations should be interpreted with caution. Forth, there were a few participants with upper urinary tract urothelial cancer, which may have influenced the results. Fifth, owing to the retrospective nature of this study, we could not account for bias in patient selection. Sixth, the sensitivity of the eosinophil cutoff value obtained from the ROC analysis was low. Finally, irAEs that were not documented in the medical records might have been overlooked. Large‐scale prospective interventional studies are required to confirm the findings of this study.
5. Conclusion
An increased eosinophil proportion may help clinicians predict the occurrence of irAEs in patients with bladder cancer treated with pembrolizumab.
Author Contributions
Yusuke Noda: investigation, writing – review and editing. Taku Naiki: conceptualization, investigation, formal analysis, writing – original draft, writing – review and editing, methodology. Rika Banno: investigation, writing – review and editing. Yoshihiko Tasaki: conceptualization, investigation, formal analysis, funding acquisition, writing – original draft, writing – review and editing, methodology. Kunihiro Odagiri: conceptualization, investigation, formal analysis, writing – original draft, writing – review and editing, methodology. Toshiharu Morikawa: investigation, writing – review and editing. Keitaro Iida: investigation, writing – review and editing. Yoshihisa Mimura: investigation, writing – review and editing. Maria Aoki: investigation, writing – review and editing. Nobuhiko Shimizu: investigation, writing – review and editing. Yosuke Sugiyama: investigation, formal analysis, writing – review and editing, methodology. Toshiki Etani: investigation, writing – review and editing. Ryosuke Ando: investigation, writing – review and editing. Noriyasu Kawai: investigation, writing – review and editing. Takashi Nagai: investigation, writing – review and editing. Yoko Furukawa‐Hibi: supervision, writing – review and editing, investigation. Takahiro Yasui: supervision, writing – review and editing, investigation. Teruki Isobe: investigation, writing – review and editing. Shuzo Hamamoto: investigation, writing – review and editing.
Funding
This work was supported by the Japan Society for the Promotion of Science (25K18657).
Ethics Statement
This study was approved by the Ethics Committee of Nagoya City University Hospital (approval number 60‐18‐0060) and was conducted in accordance with the tenets of the Declaration of Helsinki (as updated in 2024).
Consent
As this was a retrospective study, patient consent was not required. Patients could also choose to opt out of the study by using the authors' institutional websites.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figure S1: Analysis of eosinophil proportion excluded patients who experienced irAE by 3 weeks after pembrolizumab (A and B) Graph showing the variation of eosinophil proportion from baseline to 3 weeks in patients with (A) upper urinary tract urothelial cancer (baseline samples: non‐irAE, n = 26, and irAE, n = 14; 3‐week samples: non‐irAE, n = 19, and irAE, n = 8) or (B) bladder cancer (baseline samples: non‐irAE, n = 52, and irAE, n = 16; 3‐week samples: non‐irAE, n = 36, and irAE, n = 12). Results are presented as the mean ± SD. (A and B) Two‐factor repeated measurement analysis (analysis of variance [ANOVA]). irAE, immune‐related adverse event.
Figure S2: Analysis of eosinophil count from baseline to 3 weeks after treatment. (A) Graph showing the variation of the eosinophil count from baseline to 3 weeks in patients with metastatic urothelial carcinoma (baseline samples: non‐irAE, n = 70, and irAE, n = 34; 3‐week samples: non‐irAE, n = 52, and irAE, n = 28). (B and C) Graph showing the variation of eosinophil count from baseline to 3 weeks in patients with (B) upper urinary tract urothelial cancer (baseline samples: non‐irAE, n = 20, and irAE, n = 15; 3‐week samples: non‐irAE, n = 19, and irAE, n = 11) or (C) bladder cancer (baseline samples: non‐irAE, n = 45, and irAE, n = 18; 3‐week samples: non‐irAE, n = 32, and irAE, n = 16). Results are presented as the mean ± SD. (A–C) Two‐factor repeated measurement analysis (analysis of variance [ANOVA]). irAE, immune‐related adverse event.
Figure S3: Analysis of eosinophil proportion from baseline to 3 weeks after treatment excluding the patients with a history of autoimmune disease. (A) Graph showing the variation of the eosinophil proportion from baseline to 3 weeks in patients with metastatic urothelial carcinoma (baseline samples: non‐irAE, n = 83, and irAE, n = 39; 3‐week samples: non‐irAE, n = 56, and irAE, n = 28). (B and C) Graph showing the variation of eosinophil proportion from baseline to 3 weeks in patients with (B) upper urinary tract urothelial cancer (baseline samples: non‐irAE, n = 26, and irAE, n = 18; 3‐week samples: non‐irAE, n = 19, and irAE, n = 11) or (C) bladder cancer (baseline samples: non‐irAE, n = 51, and irAE, n = 20; 3‐week samples: non‐irAE, n = 35, and irAE, n = 16). Results are presented as the mean ± SD. (A–C) Two‐factor repeated measurement analysis (analysis of variance [ANOVA]). irAE, immune‐related adverse event.
Figure S4: Analysis of eosinophil proportion from baseline to 3 weeks after treatment, by specific categories of irAE. Graph showing the variation of the eosinophil proportion from baseline to 3 weeks in patients with bladder cancer, by specific categories of irAE. (A–E) Non‐irAE group versus (A) Endocrine irAE (baseline samples: non‐irAE, n = 52, and Endocrine irAE, n = 5; 3‐week samples: non‐irAE, n = 32, and Endocrine irAE, n = 3) or (B) Gastrointestinal irAE (baseline samples: non‐irAE, n = 52, and Gastrointestinal irAE, n = 7; 3‐week samples: non‐irAE, n = 32, and Gastrointestinal irAE, n = 6) or (C) Pulmonary irAE (baseline samples: non‐irAE, n = 52, and Pulmonary irAE, n = 3; 3‐week samples: non‐irAE, n = 32, and Pulmonary irAE, n = 3) or (D) Skin irAE (baseline samples: non‐irAE, n = 52, and Skin irAE, n = 5; 3‐week samples: non‐irAE, n = 32, and Skin irAE, n = 3) or (E) Other irAE (baseline samples: non‐irAE, n = 52, and Other irAE, n = 3; 3‐week samples: non‐irAE, n = 32, and Other irAE, n = 2). Results are presented as the mean ± SD. (A–E) Two‐factor repeated measurement analysis (analysis of variance [ANOVA]). irAE, immune‐related adverse event.
Table S1: Number of patients who were analyzed for variation in eosinophil proportion.
Table S2: Number of patients who were analyzed for variation in eosinophil proportion by each primary tumor site.
Table S3: Comorbidities and past histories of diseases.
Table S4: Prior treatment history with pembrolizumab.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Siegel R. L., Giaquinto A. N., and Jemal A., “Cancer Statistics, 2024,” CA: a Cancer Journal for Clinicians 74, no. 1 (2024): 12–49, 10.3322/caac.21820. [DOI] [PubMed] [Google Scholar]
- 2. von der Maase H., Hansen S. W., Roberts J. T., et al., “Gemcitabine and Cisplatin Versus Methotrexate, Vinblastine, Doxorubicin, and Cisplatin in Advanced or Metastatic Bladder Cancer: Results of a Large, Randomized, Multinational, Multicenter, Phase III Study,” Journal of Clinical Oncology 18, no. 17 (2000): 3068–3077, 10.1200/jco.2000.18.17.3068. [DOI] [PubMed] [Google Scholar]
- 3. Bellmunt J., de Wit R., Vaughn D. J., et al., “Pembrolizumab as Second‐Line Therapy for Advanced Urothelial Carcinoma,” New England Journal of Medicine 376, no. 11 (2017): 1015–1026, 10.1056/nejmoa1613683. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Powles T., Valderrama B. P., Gupta S., et al., “Enfortumab Vedotin and Pembrolizumab in Untreated Advanced Urothelial Cancer,” New England Journal of Medicine 390, no. 10 (2024): 875–888, 10.1056/nejmoa2312117. [DOI] [PubMed] [Google Scholar]
- 5. Buder‐Bakhaya K. and Hassel J. C., “Biomarkers for Clinical Benefit of Immune Checkpoint Inhibitor Treatment‐A Review From the Melanoma Perspective and Beyond,” Frontiers in Immunology 9 (2018): 1474, 10.3389/fimmu.2018.01474. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Tasaki Y., Hamamoto S., Sugiyama Y., et al., “Elevated Eosinophils Proportion as Predictor of Immune‐Related Adverse Events After Ipilimumab and Nivolumab Treatment of Advanced and Metastatic Renal Cell Carcinoma,” International Journal of Urology 30, no. 10 (2023): 866–874, 10.1111/iju.15220. [DOI] [PubMed] [Google Scholar]
- 7. Tasaki Y., Sugiyama Y., Hamamoto S., et al., “Eosinophil May Be a Predictor of Immune‐Related Adverse Events Induced by Different Immune Checkpoint Inhibitor Types: A Retrospective Multidisciplinary Study,” Cancer Medicine 12, no. 24 (2023): 21666–21679, 10.1002/cam4.6724. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Tasaki Y., Hamamoto S., Yamashita S., et al., “Eosinophil Is a Predictor of Severe Immune‐Related Adverse Events Induced by Ipilimumab Plus Nivolumab Therapy in Patients With Renal Cell Carcinoma: A Retrospective Multicenter Cohort Study,” Frontiers in Immunology 15 (2024): 1483956, 10.3389/fimmu.2024.1483956. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Kanda Y., “Investigation of the Freely Available Easy‐To‐Use Software ‘EZR’ for Medical Statistics,” Bone Marrow Transplantation 48 (2013): 452–458, 10.1038/bmt.2012.244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Varricchi G., Galdiero M. R., Loffredo S., et al., “Eosinophils: The Unsung Heroes in Cancer?,” Oncoimmunology 7, no. 2 (2018): e1393134, 10.1080/2162402x.2017.1393134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Minohara K., Matoba T., Kawakita D., et al., “Novel Prognostic Score for Recurrent or Metastatic Head and Neck Cancer Patients Treated With Nivolumab,” Scientific Reports 11 (2021): 16992, 10.1038/s41598-021-96538-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Nishikawa D., Suzuki H., Beppu S., et al., “Eosinophil Prognostic Scores for Patients With Head and Neck Squamous Cell Carcinoma Treated With Nivolumab,” Cancer Science 112, no. 1 (2021): 339–346, 10.1111/cas.14706. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Furubayashi N., Minato A., Negishi T., et al., “The Association of Clinical Outcomes With Posttreatment Changes in the Relative Eosinophil Counts and Neutrophil‐to‐Eosinophil Ratio in Patients With Advanced Urothelial Carcinoma Treated With Pembrolizumab,” Cancer Management and Research 13 (2021): 8049–8056, 10.2147/cmar.s333823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Furubayashi N., Minato A., Negishi T., et al., “Association Between Immune‐Related Adverse Events and Efficacy and Changes in the Relative Eosinophil Count Among Patients With Advanced Urothelial Carcinoma Treated by Pembrolizumab,” Cancer Management and Research 14 (2022): 1641–1651, 10.2147/cmar.s360473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Furubayashi N., Minato A., Negishi T., et al., “The Eosinophil Changes, Efficacy and Safety of Pembrolizumab in Advanced Urothelial Carcinoma Patients With an Older Age and a Poor Performance Status,” Oncotargets and Therapy 15 (2022): 1321–1330, 10.2147/ott.s389138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Furubayashi N., Minato A., Tomoda T., et al., “Change in the Neutrophil‐to‐Eosinophil Ratio After Avelumab Maintenance for Advanced Urothelial Carcinoma: The UROKYU Study,” Anticancer Research 44, no. 4 (2024): 1675–1681, 10.21873/anticanres.16966. [DOI] [PubMed] [Google Scholar]
- 17. Gambale E., Maruzzo M., Messina C., et al., “Neutrophil‐to‐Eosinophil Ratio Predicts the Efficacy of Avelumab in Patients With Advanced Urothelial Carcinoma Enrolled in the MALVA Study (Meet‐URO 25),” Clinical Genitourinary Cancer 22 (2024): 102099, 10.1016/j.clgc.2024.102099. [DOI] [PubMed] [Google Scholar]
- 18. Blomberg O. S., Spagnuolo L., Garner H., et al., “IL‐5‐Producing CD4(+) T Cells and Eosinophils Cooperate to Enhance Response to Immune Checkpoint Blockade in Breast Cancer,” Cancer Cell 41 (2023): 106–123.e10, 10.1016/j.ccell.2022.11.014. [DOI] [PubMed] [Google Scholar]
- 19. Hommes J. W., Verheijden R. J., Suijkerbuijk K. P. M., and Hamann D., “Biomarkers of Checkpoint Inhibitor Induced Immune‐Related Adverse Events—A Comprehensive Review,” Frontiers in Oncology 10 (2020): 585311, 10.3389/fonc.2020.585311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Nakamura Y., Tanaka R., Maruyama H., et al., “Correlation Between Blood Cell Count and Outcome of Melanoma Patients Treated With Anti‐PD‐1 Antibodies,” Japanese Journal of Clinical Oncology 49, no. 5 (2019): 431–437, 10.1093/jjco/hyy201. [DOI] [PubMed] [Google Scholar]
- 21. Chu X., Zhao J., Zhou J., et al., “Association of Baseline Peripheral‐Blood Eosinophil Count With Immune Checkpoint Inhibitor‐Related Pneumonitis and Clinical Outcomes in Patients With Non‐Small Cell Lung Cancer Receiving Immune Checkpoint Inhibitors,” Lung Cancer 150 (2020): 76–82, 10.1016/j.lungcan.2020.08.015. [DOI] [PubMed] [Google Scholar]
- 22. Diamantopoulos P. T., Gkoufa A., Anastasopoulou A., et al., “Exploring the Dynamics of Immune Checkpoint Inhibitor‐Induced Eosinophilia in Advanced/Metastatic Melanoma: A Comprehensive Retrospective Analysis,” Cancer Medicine 14, no. 7 (2025): e70679, 10.1002/cam4.70679. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Green D. A., Rink M., Xylinas E., et al., “Urothelial Carcinoma of the Bladder and the Upper Tract: Disparate Twins,” Journal of Urology 189, no. 4 (2013): 1214–1221, 10.1016/j.juro.2012.05.079. [DOI] [PubMed] [Google Scholar]
- 24. Tomiyama E., Fujita K., Hashimoto M., et al., “Comparison of Molecular Profiles of Upper Tract Urothelial Carcinoma vs. Urinary Bladder Cancer in the Era of Targeted Therapy: A Narrative Review,” Translational Andrology and Urology 11, no. 12 (2022): 1747–1761, 10.21037/tau-22-457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Bajorin D. F., Witjes J. A., Gschwend J. E., et al., “Adjuvant Nivolumab Versus Placebo in Muscle‐Invasive Urothelial Carcinoma,” New England Journal of Medicine 384, no. 22 (2021): 2102–2114, 10.1056/nejmoa2034442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Powles T., Park S. H., Voog E., et al., “Avelumab Maintenance Therapy for Advanced or Metastatic Urothelial Carcinoma,” New England Journal of Medicine 383, no. 13 (2020): 1218–1230, 10.1056/nejmoa2002788. [DOI] [PubMed] [Google Scholar]
- 27. Burr M. L., Sparbier C. E., Chan Y. C., et al., “CMTM6 Maintains the Expression of PD‐L1 and Regulates Anti‐Tumour Immunity,” Nature 549, no. 7670 (2017): 101–105, 10.1038/nature23643. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. George S., Miao D., Demetri G. D., et al., “Loss of PTEN Is Associated With Resistance to Anti‐PD‐1 Checkpoint Blockade Therapy in Metastatic Uterine Leiomyosarcoma,” Immunity 46, no. 2 (2017): 197–204, 10.1016/j.immuni.2017.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Carlo M. I., Ravichandran V., Srinavasan P., et al., “Cancer Susceptibility Mutations in Patients With Urothelial Malignancies,” Journal of Clinical Oncology 38, no. 5 (2020): 406–414, 10.1200/jco.19.01395. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1: Analysis of eosinophil proportion excluded patients who experienced irAE by 3 weeks after pembrolizumab (A and B) Graph showing the variation of eosinophil proportion from baseline to 3 weeks in patients with (A) upper urinary tract urothelial cancer (baseline samples: non‐irAE, n = 26, and irAE, n = 14; 3‐week samples: non‐irAE, n = 19, and irAE, n = 8) or (B) bladder cancer (baseline samples: non‐irAE, n = 52, and irAE, n = 16; 3‐week samples: non‐irAE, n = 36, and irAE, n = 12). Results are presented as the mean ± SD. (A and B) Two‐factor repeated measurement analysis (analysis of variance [ANOVA]). irAE, immune‐related adverse event.
Figure S2: Analysis of eosinophil count from baseline to 3 weeks after treatment. (A) Graph showing the variation of the eosinophil count from baseline to 3 weeks in patients with metastatic urothelial carcinoma (baseline samples: non‐irAE, n = 70, and irAE, n = 34; 3‐week samples: non‐irAE, n = 52, and irAE, n = 28). (B and C) Graph showing the variation of eosinophil count from baseline to 3 weeks in patients with (B) upper urinary tract urothelial cancer (baseline samples: non‐irAE, n = 20, and irAE, n = 15; 3‐week samples: non‐irAE, n = 19, and irAE, n = 11) or (C) bladder cancer (baseline samples: non‐irAE, n = 45, and irAE, n = 18; 3‐week samples: non‐irAE, n = 32, and irAE, n = 16). Results are presented as the mean ± SD. (A–C) Two‐factor repeated measurement analysis (analysis of variance [ANOVA]). irAE, immune‐related adverse event.
Figure S3: Analysis of eosinophil proportion from baseline to 3 weeks after treatment excluding the patients with a history of autoimmune disease. (A) Graph showing the variation of the eosinophil proportion from baseline to 3 weeks in patients with metastatic urothelial carcinoma (baseline samples: non‐irAE, n = 83, and irAE, n = 39; 3‐week samples: non‐irAE, n = 56, and irAE, n = 28). (B and C) Graph showing the variation of eosinophil proportion from baseline to 3 weeks in patients with (B) upper urinary tract urothelial cancer (baseline samples: non‐irAE, n = 26, and irAE, n = 18; 3‐week samples: non‐irAE, n = 19, and irAE, n = 11) or (C) bladder cancer (baseline samples: non‐irAE, n = 51, and irAE, n = 20; 3‐week samples: non‐irAE, n = 35, and irAE, n = 16). Results are presented as the mean ± SD. (A–C) Two‐factor repeated measurement analysis (analysis of variance [ANOVA]). irAE, immune‐related adverse event.
Figure S4: Analysis of eosinophil proportion from baseline to 3 weeks after treatment, by specific categories of irAE. Graph showing the variation of the eosinophil proportion from baseline to 3 weeks in patients with bladder cancer, by specific categories of irAE. (A–E) Non‐irAE group versus (A) Endocrine irAE (baseline samples: non‐irAE, n = 52, and Endocrine irAE, n = 5; 3‐week samples: non‐irAE, n = 32, and Endocrine irAE, n = 3) or (B) Gastrointestinal irAE (baseline samples: non‐irAE, n = 52, and Gastrointestinal irAE, n = 7; 3‐week samples: non‐irAE, n = 32, and Gastrointestinal irAE, n = 6) or (C) Pulmonary irAE (baseline samples: non‐irAE, n = 52, and Pulmonary irAE, n = 3; 3‐week samples: non‐irAE, n = 32, and Pulmonary irAE, n = 3) or (D) Skin irAE (baseline samples: non‐irAE, n = 52, and Skin irAE, n = 5; 3‐week samples: non‐irAE, n = 32, and Skin irAE, n = 3) or (E) Other irAE (baseline samples: non‐irAE, n = 52, and Other irAE, n = 3; 3‐week samples: non‐irAE, n = 32, and Other irAE, n = 2). Results are presented as the mean ± SD. (A–E) Two‐factor repeated measurement analysis (analysis of variance [ANOVA]). irAE, immune‐related adverse event.
Table S1: Number of patients who were analyzed for variation in eosinophil proportion.
Table S2: Number of patients who were analyzed for variation in eosinophil proportion by each primary tumor site.
Table S3: Comorbidities and past histories of diseases.
Table S4: Prior treatment history with pembrolizumab.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.