Abstract
Background/Aim: Immune checkpoint inhibitors (ICIs) have revolutionized the treatment of metastatic urothelial carcinoma (mUC). However, they could be associated with immune-related adverse events (irAEs), which may be clinically significant. Identifying clinical characteristics that may be associated with a higher risk of irAEs is of great importance. Patients and Methods: We retrospectively collected data from all patients who received anti-PD1 or anti-PD-L1 for metastatic UC at our Institution from January 2017 to December 2022. Patients were dichotomized according to baseline neutrophil-to-lymphocyte ratio (NLR), systemic immune-inflammation index (SII), and platelet-to-lymphocyte ratio (PLR) values. We performed univariate and multivariate logistic regression to determine the association between baseline characteristics and the development of irAEs. Results: A total of 119 patients were identified. At a median follow-up of 29.6 months, 96 patients progressed and 82 died. Forty-five patients developed irAEs of any grade, 8 patients developed grade 3 toxicities. In the univariate analysis PS of 0 (p<0.01), baseline NLR <3.52, baseline PLR <194 (p=0.04) and baseline SII <906 (p=0.01) were significantly associated with a higher risk of developing irAEs, whereas in the multivariate analysis only PS=0 (p<0.01) and NLR <3.52 (p=0.03) maintained their correlation. Median progression-free survival (mPFS) and overall survival (mOS) were significantly longer in patients with NLR <3 (mPFS 3.8 vs. 2.6 months, p=0.01; mOS 15.3 vs. 5.6 months, p=0.002) and PS=0 (mPFS 4.8 vs. 2.1 months, p<0.001; mOS 15.3 vs. 3.8 months, p<0.001). Conclusion: Low baseline NLR, PLR, and SII and good PS are associated with a higher risk of developing irAEs in patients treated with ICIs for mUC.
In recent years, immune checkpoint inhibitors (ICIs) have revolutionized the treatment of metastatic urothelial carcinoma (mUC), becoming the treatment of choice after first-line platinum-based chemotherapy and showing interesting results in patients ineligible for first-line platinum-based treatment (1-3).
However, despite the good tolerability of these treatments, patients treated with antibodies targeting programmed death 1 (PD1) or programmed death-ligand 1 (PD-L1) could develop immune-related adverse events (irAEs), sometimes leading to clinically significant toxicities that could require the initiation of immunosuppressive therapy or lead to the withdrawal or permanent cessation of the treatment.
Many authors have attempted to identify clinical characteristics that may be associated with a higher risk of developing irAEs (4, 5). Some of systemic inflammation markers calculated from peripheral blood count values (such as neutrophil-to-lymphocyte ratio, NLR, platelet-to-lymphocyte ratio, PLR, C-reactive protein and the combination of low hemoglobin levels with PLR), which have already been shown to be prognostic factors in patients with mUC (6-12), have also been shown to be associated with a higher risk of immune toxicities in patients with other solid malignancies (4, 5, 13).
The aim of our study was therefore to investigate the potential association between three baseline systemic inflammation indexes and the development of irAEs in a real-world cohort of patients treated with ICIs for mUC.
Patients and Methods
We retrospectively reviewed the electronic records of all patients who received anti-PD1 or anti-PD-L1 treatment for metastatic UC at our Institution from January 2017 to December 2022. Inclusion criteria included a histological diagnosis of UC, the presence of metastases, at least one course of treatment with ICIs, and the availability of all necessary data.
For each patient, we collected baseline demographic data, tumor characteristics, details of the course of immunotherapy (line of treatment, start and stop dates, date of disease progression), treatment toxicity (type, grade, date of onset, need to discontinue treatment or use corticosteroids), date of death or last follow-up, and baseline blood count values. For each patient, baseline NLR, PLR and systemic immune-inflammation index (SII) (neutrophil×platelet/lymphocyte) were calculated. Adverse events were graded according to the Common Terminology Criteria for Adverse Events (CTCAE) v5.0 and radiological response was defined according to Response Evaluation Criteria in Solid Tumors (RECIST) v1.1.
Progression-free survival (PFS) was calculated using the Kaplan–Meier method from the date of treatment initiation to the date of disease progression or death (whichever occurred first); PFS was censored at the last patient follow-up visit without progression. Overall survival (OS) was calculated from the date of drug initiation to the date of death from any cause or censored at the last date known to be alive.
Key metrics were summarized using descriptive statistics. Categorical variables were compared using the Chi-square test or Fisher’s exact test, depending on the samples’ numerosity. Significant differences in numerical variables were tested using the Mann–Whitney test.
Patients were dichotomized according to baseline NLR, SII, and PLR values, using both prespecified cut-offs based on literature reference (≥3 vs. <3 for NLR, ≥180 vs. <180 for PLR, ≥1,375 vs. <1,375 for SII) (7, 13-26) and cut-off values calculated for our cohort from receiver operating characteristic (ROC) curves, selecting for each variable the values associated with the highest Youden’s index.
We performed univariate and multivariate logistic regression to determine the association between baseline characteristics and the development of irAEs of any grade; the covariates that showed an association with the oncological outcome with a p-value lower than 0.05 in the univariate analysis were included in the multivariate analysis.
Results were considered statistically significant if their p-values were <0.05. All statistical analyses were performed using “R” software v4.0.5. At the time of their first visit to our institution, all patients gave their written consent for their clinical data to be used for scientific purposes. The study was conducted in accordance with the Declaration of Helsinki. Data collection was approved by the local Ethics Committee on April 20th, 2022 (approval number 0008904/22).
Results
Patients’ characteristics. From January 2017 to December 2022, a total of 119 patients were treated with anti-PD1 or anti-PDL1 ICIs for mUC at our Institution. At a median follow-up time of 29.6 months, 96 patients (80.7%) progressed and 82 (68.9%) died. Regarding the type of ICI used, 22 patients (18.5%) were treated with atezolizumab, 29 (24.4%) with durvalumab and 68 (57.1%) with pembrolizumab. Patients’ characteristics are shown in Table I.
Patients’ characteristics (N=119).
Immune-related adverse events. In the total population, 45 patients (37.8%) developed irAEs of any grade. The most common adverse events reported were skin toxicities (pruritus and rash 19 patients), endocrine and metabolic disorders (hypothyroidism 13 patients, hyperglycemia 4 patients) and diarrhea (13 patients). A total of 8 patients (6.7%) developed grade 3 toxicities (one case of nephritis, one case of liver enzyme elevation, one case of hyperglycemia, one case of rash, one case of encephalitis, three cases of diarrhea). All irAEs are listed in Table II. Median time from the start of immunotherapy to the onset of an irAE was 75 days (range=20-576 days).
Immune-related adverse events reported in our cohort.
Comparison of baseline characteristics of patients experiencing irAEs. We analyzed the baseline clinical characteristics of patients who experienced the development of immune-related toxicities and compared them with those who did not; the results are shown in Table III.
Comparison of patients who did or did not develop immune-related adverse events (irAEs).
There was no significant difference in terms of age, line or type of treatment between the two groups; patients who did not develop irAEs had a significantly worse PS compared to those who experienced such events (PS 1-2: 59.5% vs. 26.7%, p<0.001).
Patients who did not develop an irAE had higher median NLR (3.67 vs. 2.53), SII (999 vs. 635) and PLR (181 vs. 143); the difference was statistically significant for NLR and SII (p=0.009 and p=0.02 respectively), but not for PLR (p=0.15). The results are graphically represented in Figure 1.
Boxplots representing baseline inflammatory indexes according to the development of immune-related adverse events. From left to right: neutrophil-to-lymphocyte ratio (NLR), platelets-to-lymphocyte ratio (PLR), systemic immune inflammation index (SII). Values on the y-axis are on a logarithmic scale.
Determination of baseline blood count cut-off for prediction of toxicities. In order to define optimal cut-off values for the biomarkers, we performed ROC analysis of baseline PLR, NLR, and SII as predictors of toxicity of any grade.
Cut-off values based on the development of irAE were 3.52 (sensitivity 52.3; specificity 77.8) for NLR, 906 (sensitivity 56.7; specificity 66.7) for SII and 194 (sensitivity 45.9; specificity 73.3) for PLR; results are shown in Figure 2. Areas under the curve (AUC) were 0.64, 0.62, and 0.58 respectively.
Receiver operating characteristic (ROC) curves for development of adverse events of any grade for neutrophil-to-lymphocyte ratio (left chart), platelet-to-lymphocyte ratio (middle chart), and systemic immune inflammation index (right chart).
Univariate and multivariate analysis for the appearance of irAE. We investigated the association between baseline clinical characteristics (including blood count indexes) and the risk of developing irAEs; in the univariate analysis, for each biomarker we used both the cut-off values calculated for our cohort (3.52 for NLR, 194 for PLR, 906 for SII) and the cut-off values more commonly used in clinical practice (3 for NLR, 180 for PLR, 1,375 for SII). The results are shown in Table IV.
Univariable and multivariable analysis for the appearance of immune-related adverse events (irAEs).
Age, sex, type of ICI (anti-PD1/anti-PD-L1) and drug used (pembrolizumab, atezolizumab or durvalumab) did not show an association with the risk of developing irAEs (p=0.41, 0.45, 0.57 and 0.46 respectively).
Characteristics associated to a significantly higher risk of developing irAEs of any grade were a PS of 0 (odds ratio, OR=4.03, 95%CI=1.69-9.93, p<0.01), baseline NLR <3 (OR=3.22, 95%CI=1.39-7.52, p<0.01) or <3.52 (OR=3.90, 95%CI=1.73-9.38), baseline PLR <194 (OR=2.34, 95%CI=1.06-5.36, p=0.04), baseline SII <906 (OR=2.62, 95%CI=1.22-5.79, p=0.01).
In the multivariate analysis PS=0 (OR=3.48, 95%CI=1.48-8.58, p<0.01) and NLR <3.52 (OR=3.89, 95%CI=1.14-14.38, p=0.03) were significantly associated with a higher risk of developing irAEs (Table IV).
Association of baseline NLR and PS with treatment outcomes and competing risk analysis. PFS and OS were significantly longer in patients with NLR <3 (median PFS 3.8 vs. 2.6 months, p=0.01; median OS 15.3 vs. 5.6 months, p=0.002); therefore, patients with low NLR were treated with a higher number of cycles (median duration of treatment: 5 vs. 3 cycles, p<0.001). PFS and OS were significantly longer in patients with PS=0 (median PFS 4.8 vs. 2.1 months, p<0.001; median OS 15.3 vs. 3.8 months, p<0.001).
In order to exclude the possible bias of a higher incidence of AEs due only to the longer treatment duration and longer follow-up time in patients with a low NLR or PS=0, we performed a competing risk analysis confronting the cumulative incidence of irAEs in the two groups, with death as the competing event.
Low NLR maintained an association with higher risk of developing adverse events for both cut-off values (p<0.05 for both NLR <3 and NLR <3.52); PS equal to 0 also maintained an association with higher risk of developing irAEs (p<0.01).
Association between development of irAEs and treatment outcomes. Patients who experienced the development of irAEs had a significantly longer PFS (median PFS: 9.9 months, 95%CI=5.2-16.2 vs. 2.5 months, 95%CI=2.1-2.8; p<0.001) and a significantly longer OS (median OS: 18.9 months, 95%CI=11.8- not reached vs. 3.8 months, 95%CI=3.1-7.1; p<0.001).
Risk factors for G3-G4 adverse events. In our cohort, eight patients (6.7%) experienced grade 3-4 adverse events (Table II). Severe adverse events were less frequent in patients with NLR <3 (3.8% vs. 9.1%, OR=0.39, 95%CI=0.04-2.33, p=0.39), PLR <180 (6.3% vs. 7.1%, OR 0.88, 95%CI=0.16-4.99, p=1.00), SII<1375 (6.0% vs. 8.3%, OR=0.71, 95%CI=0.13-4.82, p=0.69), PS <1 (6.3% vs. 7.1%, OR=0.88, 95%CI=0.16-4.99, p=1.00). All differences were not statistically significant.
Discussion
Immunotherapy has dramatically improved the treatment options and prognosis of patients with mUC. Pembrolizumab is the current treatment of choice for patients who have progressed after first-line chemotherapy for metastatic disease (1, 2), while avelumab has been shown to prolong OS when used as a maintenance treatment for patients who have completed first-line platinum-based CT without subsequent progression (1, 16).
However, treatment with ICIs has also fewer positive aspects. Not all patients benefit from immunotherapy, with a significant proportion showing progressive disease after the first cycles of treatment (for example, 48.5% of patients treated with second-line pembrolizumab in the KEYNOTE-045 pivotal trial) (2). Moreover, ICIs can cause irAEs that can affect a patient’s quality of life, cause delays or interruptions in treatment, and can persist for months after treatment is stopped.
In a real-world population of patients treated with anti-PD1 or anti-PD-L1 ICIs for mUC, we found that a good PS and low values of systemic inflammation indexes (SII, PLR and NLR) could identify patients at higher risk of developing immune-related toxicities; however, only for PS and NLR the result was confirmed in the multivariate analysis and only for NLR the association was statistically significant using both the cut-off value calculated for our population and a more “standardized” cut-off (Table IV).
The prognostic role of systemic inflammation indexes or other factors related to inflammation (such as anemia or cachexia), in patients with mUC has been documented by many researchers (6-12, 27), with some authors attempting to incorporate these markers in new prognostic models, alongside other clinical parameters (performance status, metastatic sites) already used in clinical practice (17); more recently, the correlation between inflammatory markers and irAEs has been demonstrated (4, 13, 18).
The association between low NLR and higher risk of irAEs has been reported by many authors (13, 18-21) and confirmed in meta-analyses (22-23); there have been similar reports for SII (21) and PLR (13). Interestingly, some authors have reported a higher risk of severe or high grade irAEs in patients with a high NLR (24-26). In our population, grade 3 toxicities were more common in patients with a high NLR, but probably due to the low number of these events, the difference was not statistically significant.
There is currently no definitive biological mechanism that could explain the association between these biomarkers and the outcome of patients treated with immunotherapy. One possible explanation is that these indexes are an indirect evaluation of both the host’s immunocompetence and the chronic inflammation caused by the tumor.
High levels of platelets and neutrophils are in fact indirect markers of systemic inflammation, a phenomenon associated with immunosuppression (28, 29); neutrophils have an immunomodulatory role, as they have been shown to interact with other immune cells (such as T-lymphocytes) and to suppress lymphocyte activity (30-32). Composite indexes could evaluate these parameters simultaneously with immune activity, represented by lymphocytes (33).
The association we found between a better PS and a higher risk of developing irAE of any grade (Table IV), although counterintuitive, has also been shown by other authors (34, 35). Shimozaki et al., by analyzing a heterogeneous cohort of patients treated with ICIs for various malignancies, found that a PS <2 was associated with a higher risk of irAEs (OR 2.11), although the association was not statistically significant in the multivariate analysis (34). In the paper by Pavan et al., there was a higher proportion of people with PS=0 in the group of patients who developed an irAE, although the difference narrowly missed statistical significance (31.7% vs. 16.9%, p=0.053) (13).
A common problem with many studies evaluating the predictive or prognostic role of these biomarkers is that the results are not always easily comparable due to differences in the definition of the threshold value (4, 22).
We believe that identifying patients at higher risk of developing treatment-related toxicities could help physicians in their daily practice to choose the best option for each patient and to identify those who need more frequent and closer monitoring for AEs.
Many different biomarkers are being studied as predictive factors for irAEs in patients treated with immunotherapy (4); one of the strengths of systemic inflammation indexes is that they can be calculated from blood tests that are highly accessible, inexpensive, and routinely performed in clinical practice. Tumor characteristics, such as histology of the primary tumor, mutational burden and tumor burden, had been associated with differences in incidence of irAEs (5, 36, 37); unfortunately, in our cohort the analysis of fibroblast growth factor receptors 2-3 alterations (FGFR2/3) had been performed for only a minority of the patients. Our population was too small to evaluate separately patients affected from neoplasm originating from the bladder and patients with upper tract urothelial cancers, although it is well known that the latter exhibit different characteristics both from a molecular perspective and in terms of clinical behavior and response to immunotherapy (38, 39).
We found an association between the development of irAEs and longer PFS and OS, as reported in genitourinary oncology from many other authors before (40-42); however, our sample size was too small to perform statistical analyses (such as landmark-time analysis) that could properly handle immortal-time bias, thus limiting the reliability of our finding.
This study has several limitations, the first of which is its retrospective design. The study was conducted in a single institution, thus limiting the generalizability of our findings to diverse patient populations (especially for those with different demographic or clinical characteristics). The treatments administered were heterogeneous in terms of the drugs used and the treatment regimen; however, it should be noted that they all have a similar mechanism of action and that we did not find an association between any specific treatment and the development of irAEs. As highlighted previously, data regarding FGFR2-3 alterations were available only for a few patients, and therefore had not been included in the analyses. The sample size was small, which limits the validity of the results and the statistical power of our analysis. Moreover, it was not possible to validate the results in an independent validation cohort. Finally, the blood tests were carried out in different laboratories.
Conclusion
Low baseline NLR, PLR, and SII and good performance status are associated with a higher risk of developing irAEs in patients treated with ICIs for mUC. These findings warrant validation and confirmation in larger, prospectively designed, multi-centric clinical trials as they may have implications for patient selection and treatment management in daily clinical practice.
Footnotes
Authors’ Contributions
All Authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Conflicts of Interest
The Authors have no financial relationships or conflicts of interest to declare.
Funding
This research received “Ricerca Corrente” funding from the Italian Ministry of Health to cover publication costs.
- Received July 18, 2024.
- Revision received August 12, 2024.
- Accepted August 23, 2024.
- Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.