Abstract
Background:
Immune checkpoint inhibitors (ICIs) have changed the treatment landscape for urothelial carcinoma. However, the relationship between immune-related adverse events (irAEs) and patient outcomes remains poorly understood.
Objective:
We investigated the association between irAEs and oncological outcomes in patients with advanced urothelial carcinoma receiving ICIs, and whether the administration of systemic corticosteroids diminishes therapeutic impact.
Methods:
the association between irAEs occurrence and clinical progression-free survival (PFS), overall survival (OS), and cancer-specific survival (CSS) was tested by means of multivariable Cox or competing-risks regression, when appropriate. Patients experiencing irAEs were further stratified based on systemic corticosteroids administration. A sensitivity analysis was conducted by repeating all the analyses with median time to irAE as landmark point.
Results:
We relied on individual participant data from two prospective trials for advanced urothelial cancer: IMvigor210 and IMvigor211. A total of 896 patients who received Atezolizumab for locally-advanced or metastatic urothelial carcinoma were considered. Overall, irAEs were recorded in 195 patients, median time to irAEs was 64 days. On multivariable analysis, irAEs were inversely associated with the risk of disease progression [HR 0.50 95%CI (0.40–0.61) p<0.001], overall mortality [HR 0.51 95%CI (0.41–0.64) p<0.001], and cancer-specific mortality [SHR 0.55 95%CI (0.45–0.72) p<0.001]. Moreover, our results did not refute the supposition that the administration of systemic corticosteroids does not impact oncological outcomes [PFS: HR 0.92 95%CI (0.62–1.34), p= 0.629; OS: HR 0.86 95%CI (0.51–1.64), p= 0.613; CSS: SHR 0.90 95%CI (0.60–1.36), p=0.630]. The sensitivity yielded consistent findings.
Conclusions:
The development of irAEs while receiving atezolizumab treatment was associated with improved oncological outcomes, namely overall and cancer-specific mortality, and progression free survival. These findings seem to not be substantially affected by administration of systemic corticosteroids.
Keywords: Urothelial Cancer, Metastatic Disease, Autoimmune Adverse Events, Systemic Corticosteroids, Immunotherapy
INTRODUCTION
In recent years, many immune checkpoint inhibitors (ICIs) have been investigated for locally advanced or metastatic urothelial carcinoma (mUC)(2). Among them, atezolizumab demonstrated clinically relevant benefits that led to its approval as a second-line therapy in patients after platinum-based chemotherapy failure, and more recently, as a first-line therapy for cisplatin-ineligible patients (1) despite its use has been removed from the US market for urothelial carcinoma.
Presently, predictors of response to ICIs both at a patient and tumor level remain poorly understood. Several studies that explored the prognostic ability of tumor markers, such as programmed death-ligand 1 (PD-L1) expression or tumor mutational burden (TMB), yielded conflicting results (1). Recent work suggests that circulating tumor DNA (ctDNA) may be a promising prognostic biomarker, but this has yet to be prospectively confirmed in clinical trials(3). Despite a growing body of evidence on the identification of drivers of response to ICIs and on how this response can be enhanced, several clinical observations warrant further investigation. One of these is the association between the occurrence of characteristic immune-related adverse events (irAEs) and oncological outcomes. Previous studies have explored the role of irAEs in response to different ICIs, as these phenomena are thought to represent the bystander effects of activated T-cells in patients responding to ICIs. At the molecular level, irAEs may be attributed to either a state of immune system disinhibition, which may lead to both break of self-recognition mechanisms and tumor response, or to the molecular mimicry between cancer neo-antigens and self antigens expressed by the affected organ(4). Nevertheless, several questions regarding this phenomenon have yet to be answered. Notably, the association with irAEs and outcomes is not consistently observed across different malignancies and antibody targets. These studies may suffer from immortal time bias, namely a statistical distortion caused when a cohort is designed so that follow-up includes a period of time where participants cannot experience any event (such as from randomization until irAE occurrence), leading to an overestimation of an effect(5). In the context of advanced urothelial cancer, a single pooled analysis(6) of seven randomized clinical trials (RCTs) demonstrated a survival advantage for individuals experiencing irAEs without accounting for immortal time bias. In this study (6), the authors relied on a heterogeneous pool of patients who received either anti-PD-1 (Nivolumab, and Pembrolizumab) or anti-PD-L1 antibodies (Atezolizumab, Avelumab, and Durvalumab) (6). Additionally, their definition of irAEs was strictly related to the use of corticosteroids (either topical or systemic), thus whether the survival advantage is somehow affected by steroid administration remains to be addressed. Taken together, these facts might partly impair the generalizability of their results(6).
Herein, we aimed to investigate the impact of irAEs on oncological outcomes in a homogeneous population of patients treated with atezolizumab for locally advanced or metastatic urothelial carcinoma. We relied on individual participant data from two clinical trials: IMvigor211 (7) and IMvigor210 (8,9). To corroborate our findings, we also assessed whether oncological outcomes were influenced by systemic corticosteroids administration among patients experiencing irAEs.
MATERIALS AND METHODS
Study population
For the present study, we relied on individual participant data from two multicenter prospective clinical trials: IMvigor210 cohort 1 and cohort 2(8,9), and IMvigor211 immunotherapy arm(7). Of the 931 patients enrolled on IMvigor211, 467 received atezolizumab-based treatment and were thus considered in our analysis. Our overall patient population consisted of 836 individuals. Data were used according to Hoffmann–La Roche policy and made available through Vivli, Inc. (www.vivli.org) (10).
Cohort 2 of IMvigor210 and the IMvigor211 trial enrolled patients with locally advanced or metastatic urothelial cancer, who had received no more than two previous lines of therapy and who had progressed during or following a prior platinum-based regimen. Conversely, cohort 1 of IMvigor210 enrolled patients with locally advanced or metastatic urothelial carcinoma who were treatment-naive and cisplatin-ineligible. The criteria for cisplatin-ineligibility were impaired renal function (glomerular filtration rate [GFR] > 30 but < 60 mL/min), hearing loss (measured by audiometry) of 25 dB, or Grade ≥2 peripheral neuropathy.
The two trials included in our analysis shared the following key inclusion criteria: a) an Eastern Cooperative Oncology Group performance status (ECOG PS) of 0 or 1, b) measurable disease defined by RECIST v1.1(11), c) predominance of transitional cell carcinoma histology, d) adequate hematologic and end-organ function, and e) no autoimmune disease or active infections. On the other hand, key exclusion criteria included history of autoimmune disease and the need for chronic systemic immunosuppressors.
According to the protocols for both studies, patients received atezolizumab 1200 mg intravenously every 3 weeks until unacceptable toxicity, RECIST v1.1 progression, or informed consent withdrawal. Atezolizumab treatment could continue beyond radiographic progression if deemed a clinical benefit by the investigator.
PD-L1 expression was centrally reviewed and graded on immunohistochemistry as follows: IC2/3 (PD-L1 expression on ≥5% of tumor-infiltrating immune cells), IC1 (PD-L1 expression on ≥1% and <5% of tumor-infiltrating immune cells), or IC0 (PD-L1 expression on <1% of tumor-infiltrating immune cells).
The Bellmunt score (0 vs ≥1) was calculated based on the previously published literature, assigning 1 point for each of the following risk factors: hemoglobin level below 10, presence of liver metastases, and the Eastern Cooperative Oncology Group PS (ECOG PS) of 1 or higher (12).
The indication to administer systemic corticosteroids in the context of irAEs was only stated in the IMvigor211 protocol. Per the study’s guidelines, oral or parenteral corticosteroids were administered in case of Grade ≥3 or recurrent Grade 2 irAEs and reported in accordance with the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events, version 4.0 (13). The dose and duration of immunosuppressive treatment, as well as Atezolizumab discontinuation, was decided by the treating physician. Once irAEs had resolved, atezolizumab treatment was allowed to resume.
Definition of irAEs
According to the definition provided in the IMvigor211 trial, the following conditions were considered suggestive of an irAE related to atezolizumab ICIs administration: immune-related pneumonitis, hypoxia or dyspnea Grade ≥ 3, colitis, endocrinopathies (such as diabetes mellitus, thyroiditis, pancreatitis, or adrenal insufficiency), vasculitis, hepatitis, transaminitis [(AST or ALT>3×ULN and bilirubin > 2× ULN) or AST/ALT > 10 × ULN], systemic lupus erythematosus, Guillain-Barré syndrome, myasthenia gravis, or skin reactions (such as pruritus, vitiligo, or pemphigoid).
Definition of oncological outcomes
Progression-free survival (PFS) was defined as time from randomization (or start of treatment in the IMvigor210 trial) to disease progression, or death, whichever occurred first, as assessed by the independent review facility according to RECIST v1.1(11). Cancer-specific survival (CSS) was defined as the time from randomization (or start of treatment in the IMvigor210 trial) to the time of death from cancer-related causes. Overall survival (OS) was defined as the time from randomization (or start of treatment in the IMvigor210 trial) to the time of death from any cause.
Statistical Analysis
For continuous variables, descriptive statistics consisted of medians and interquartile ranges (IQR). For categorical variables, descriptive statistics consisted of absolute and relative frequencies. Kruskall-Wallis and Chi‐square tests were used to compare continuous and categorical variables, respectively. Patients were stratified according to occurrence of irAE after atezolizumab treatment.
Our analyses consisted of four main steps. First, the Kaplan-Meier method was used to graphically depict PFS and OS of patients stratified by irAE occurrence (yes vs no). Subsequently, the log-rank test was used to test the equality of the PFS and OS functions between the two groups. We relied on the competing risks method to display the cumulative incidence functions for CSS and other cause of mortality (OCM) rates for the two aforementioned groups. The Gray test was employed to test the equality of the cumulative incidence functions in the presence of competing events. Second, multivariable Cox regression was used to test the association between the occurrence of irAE after ICI administration and PFS or OS. Instead, a multivariable competing risks regression was used for CSS, after accounting for the occurrence of OCM, i.e. the competing event. A directed acyclic graph (DAG) was used to identify and explicitly represent the set of potential confounders to be adjusted in regression models(14) (supplementary figure 1). Analyses were accordingly adjusted for potential confounders: age, BMI, previous chemotherapy regimens (cisplatin only vs carboplatin only vs cisplatin+carboplatin vs none), pathologic stage (T1,a,is,x vs T2 vs T3 vs T4), histology (pure Transitional Cell Carcinoma vs mixed histology), number of metastasis, and PD-L1 expression via immunohistochemistry. All analyses were performed using R statistical software version 4.2.1 (www.r-project.org). All tests were two‐sided, with a significance level set at 0.05.
Subgroup Analyses
In an effort to corroborate our findings, the aforementioned analyses were repeated after stratifying patients for corticosteroid administration.
Sensitivity Analyses
As an additional sensitivity analysis, we performed landmark analyses with 64 days as the landmark time point, in an attempt to mitigate the potential impact of immortal-time bias on OS and CSS (i.e. all patients in the analysis are required to live long enough to develop irAEs). This landmark time point was chosen because it represents the median time from randomization to irAE occurrence. Furthermore, we conducted a time-dependent Cox regression analysis to examine the influence of irAE on the mentioned oncological outcomes. The details of this analysis have been previously documented(15).
RESULTS
Descriptive Statistics
We identified 896 patients who received atezolizumab for locally advanced or metastatic urothelial carcinoma across the two studies. These individuals were stratified into two groups, based on the presence or absence of irAEs after atezolizumab treatment. Overall, irAEs were recorded in 195 patients during treatment, while the remaining 701 patients did not report any irAEs over the study period. Median time to irAE was 64 days. Patient demographics, baseline, pathological characteristics, and previously well-established prognostic factors are listed in Table 1. Age, sex, BMI, presence of histological variants, and metastatic burden were similar between the two groups. The presence of visceral metastases [62 vs 151 patients (32 vs 22%) p=0.004] and the rate of T3–4 disease [93 vs 259 patients (48 vs 37 %) p<0.001] were higher in the group experiencing irAEs. On the other hand, the frequency of chemotherapy naïve patients [12 vs 90 patients (6 vs 13%), p=0.009] and the rate of IC2/3 of PD-L1 expression [63 vs 186 patients (32 vs 27%); p<0.001] were higher in the irAE group.
Table 1.
Baseline characteristics
| Patients experiencing immune-related AE | Patients without immune-related AE | p value | |
|---|---|---|---|
|
| |||
| N° of patients | N= 195 | N= 701 | |
|
| |||
| Median Age (IQR) | 68 (61–74) | 67 (60–74) | 0.691 |
|
| |||
| Sex | |||
| • Female | 58 (30%) | 314 (44%) | <0.001 |
| • Male | 137 (70%) | 387 (56%) | |
|
| |||
| Median BMI (IQR) | 26.5 (24.0–30.0) | 26.0 (23.1–29.4) | 0.847 |
|
| |||
| Events | |||
| • Death | 135 (69%) | 569 (81%) | 0.002 |
| • Cancer Specific Death | 120 (62%) | 525 (74%) | 0.002 |
| • Progression* | 145 (74%) | 555 (79%) | <0.001 |
|
| |||
| N° of prior systemic therapies: | |||
| • 0 | 12 (6%) | 90 (13%) | 0.009 |
| • 1 | 105 (54%) | 311 (44%) | |
| • 2+ | 78 (40%) | 300 (43%) | |
|
| |||
| Type of previous systemic therapies: | |||
| • Carboplatin only | 48 (25%) | 171 (24%) | 0.196 |
| • Cisplatin + carboplatin | 21 (11%) | 41 (6%) | |
| • Cisplatin only | 112 (57%) | 390 (56%) | |
| • Other platinum combination | 2 (1%) | 10 (1%) | |
|
| |||
| Histology: | |||
| • Transitional cell carcinoma | 178 (91%) | 641 (91%) | 0.944 |
| • Transitional cell carcinoma w/ mixed histology | 17 (9%) | 60 (9%) | |
|
| |||
| Tumor stage: | |||
| • Tx/a/is/1 | 64 (33%) | 253 (36%) | 0.035 |
| • T2 | 38 (19%) | 189 (27%) | |
| • T3 | 61 (31%) | 177 (25%) | |
| • T4 | 32 (17%) | 82 (12%) | |
|
| |||
| N° of metastasis sites: | |||
| • 0 | 18 (9%) | 51 (7%) | 0.445 |
| • 1 | 64 (33%) | 200 (28%) | |
| • 2 | 61 (31%) | 240 (34%) | |
| • 3+ | 52 (27%) | 210 (31%) | |
|
| |||
| Presence of visceral metastasis | |||
| • Yes | 133 (68%) | 550 (78%) | 0.002 |
| • No | 62 (32%) | 151 (22%) | |
|
| |||
| PD-L1 expression | |||
| • IC0/1 | 132 (68%) | 515 (73%) | 0.038 |
| • IC2/3 | 63 (32%) | 186 (27%) | |
|
| |||
| Bellmunt Risk Score: | |||
| • 0 | 67 (34%) | 275 (39%) | 0.22 |
| • 1+ | 128 (66%) | 426 (61%) | |
Only disease progression with known time to event was considered.
Survival Analyses
Overall, 704 patients died (79%) throughout the duration of the study period. Of the 704 patients who died, 135 patients experienced an irAE (69%) and 569 patients did not experience an irAE (81%; p=0.002). When stratified according to cause of death, 645 patients (71%) died due to urothelial cancer, while 59 patients (7%) died due to other causes. Moreover, of the 195 patients with documented irAEs, 120 patients (62%) died from urothelial cancer. Conversely, of the 569 patients with no documented irAEs, 525 patients (92%) died from urothelial cancer (p=0.002). With regards to PFS, 700 patients (78%) experienced disease progression over the study period. Specifically, disease progression was observed in 145 patients (74%) in the irAE group and in 555 patients (79%) in the remaining patient groups (p<0.001). The progression-free, cancer-specific, and overall survival curves are displayed in Figures 1, 2, and 3, respectively. The occurrence of irAEs was associated with longer PFS and OS (Log-rank p: <0.0001 in both cases). Median OS was 15.7 months for those patients experiencing irAEs and 7.9 months for those patients not experiencing irAEs (p< 0.0001). At 20-months follow-up, the OS rate was 43% in the irAEs group and 25% in the absence of irAEs. On multivariable Cox regression analysis (Table 2), the development of irAEs during atezolizumab treatment was inversely associated with risk of disease progression [HR 0.50 95% CI (0.40–0.61) p<0.001] and overall mortality [HR 0.51 95% CI (0.41–0.64) p<0.001]. Similarly, on competing risk regression, the development of irAEs was associated with lower risk of cancer-specific mortality [SHR 0.55 95% CI (0.45–0.72) p<0.001].
Figure 1.
Progression-free survival curves, stratified according to the occurrence of immune-related adverse events
Figure 2.
Cancer-specific survival and other cause mortality cumulative incidence curves, stratified according to the occurrence of immune-related adverse events.
Figure 3.
Overall survival curves, stratified according to the occurrence of immune-related adverse events
Table 2.
Univariate and Multivariate Cox regression analysis predicting OS and PFS. Univariate and Multivariate competing risk regression analysis predicting CSS.
| OS | CSS | PFS | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||
| Unadjusted | Adjusted | Unadjusted | Adjusted | Unadjusted | Adjusted | |||||||
|
| ||||||||||||
| HR (95% CI) | p | HR (95% CI) | p | sHR (95% CI) | p | sHR (95% CI) | p | HR (95% CI) | p | HR (95% CI) | p | |
|
| ||||||||||||
| Immune-related events | 0.59 (0.49–0.72) | <0.001 | 0.51 (0.41–0.64) | <0.001 | 0.59 (0.48–0.73) | <0.001 | 0.55 (0.45–0.72) | <0.001 | 0.55 (0.45–0.66) | <0.001 | 0.50 (0.40–0.61) | <0.001 |
|
| ||||||||||||
| Age | 1.00 (0.99–1.00) | 0.2 | 1.00 (0.98–1.01) | 0.41 | 0.99 (0.98–1.00) | 0.13 | 0.99 (0.98–1.00) | <0.001 | 0.99 (0.98–1.00) | 0.095 | 0.99 (0.98–1.00) | 0.161 |
|
| ||||||||||||
| BMI | 0.97 (0.95–0.98) | <0.001 | 0.98 (0.96–0.99) | 0.004 | 0.97 (0.95–0.98) | <0.001 | 0.97 (0.96–0.99) | <0.001 | 0.97 (0.96–0.99) | 0.001 | 0.98 (0.96–0.99) | 0.005 |
|
| ||||||||||||
| Chemotherapy regiments: | ||||||||||||
| • Cisplatin only | Reference | Reference | Reference | Reference | Reference | Reference | ||||||
| • Carboplatin only | 1.68 (1.22–2.29) | 0.001 | 0.76 (0.10–5.56) | 0.785 | 0.78 (0.64–0.95) | 0.015 | 1.16 (0.93–1.45) | 0.9 | 0.79 (0.11–5.63) | 0.811 | 0.71 (0.10–5.19) | 0.734 |
| • Cisplatin+carboplatin | 1.67 (1.16–2.40) | 0.006 | 0.82 (0.11–6.12) | 0.85 | 0.76 (0.57–1.02) | 0.068 | 1.05 (0.76–1.46) | 0.69 | 0.75 (0.10–5.44) | 0.778 | 0.70 (0.09–5.16) | 0.725 |
| • No chemotherapy | 1.10 (0.44–2.73) | 0.842 | 0.43 (0.03–5.09) | 0.504 | 1.01 (0.18–5.85) | 0.99 | 1.16 (0.18–7.37) | 0.85 | 0.98 (0.10–9.47) | 0.989 | 0.56 (0.06–5.78) | 0.63 |
|
| ||||||||||||
| Histology | ||||||||||||
| • Transitional cell carcinoma | Reference | Reference | Reference | Reference | Reference | Reference | ||||||
| • Transitional cell carcinoma w/ mixed histology | 1.14 (0.80–1.62) | 0.482 | 1.28 (0.80–2.04) | 0.31 | 1.32 (0.87–2.00) | 0.19 | 1.27 (0.75–1.42) | 0.79 | 1.13 (0.74–1.71) | 0.58 | 1.62 (1.03–2.55) | 0.037 |
|
| ||||||||||||
| Stage | ||||||||||||
| • T1/a/is/x | Reference | Reference | Reference | Reference | Reference | Reference | ||||||
| • T2 | 1.16 (1.06–1.45) | <0.001 | 1.22 (1.07–1.66) | 0.007 | 1.13 (1.05–1.31) | <0.001 | 1.09 (1.02–1.34) | 0.021 | 1.36 (1.13–1.98) | 0.047 | 0.49 (0.16–1.46) | 0.198 |
| • T3 | 1.13 (1.05–1.37) | <0.001 | 1.22 (1.07–1.66) | 0.006 | 1.15 (1.06–1.36) | <0.001 | 1.19 (1.10–1.46) | 0.003 | 1.14 (0.16–1.23) | 0.116 | 0.57 (0.19–1.71) | 0.314 |
| • T4 | 1.28 (1.10–1.80) | <0.001 | 0.66 (0.22–2.04) | 0.474 | 1.36 (1.15–1.88) | <0.001 | 1.24 (0.74–2.55) | 0.738 | 1.68 (0.24–1.94) | 0.467 | 0.93 (0.30–2.91) | 0.907 |
|
| ||||||||||||
| N° of metastasis sites: | 1.28 (1.20–1.36) | <0.001 | 1.17 (1.07–1.27) | <0.001 | 1.27 (1.17–1.38) | <0.001 | 1.16 (1.10–1.53) | 0.002 | 1.13 (1.07–1.20) | <0.001 | 1.05 (0.97–1.14) | 0.215 |
|
| ||||||||||||
| Visceral mets | ||||||||||||
| • No | Reference | Reference | Reference | Reference | Reference | Reference | ||||||
| • Yes | 1.67 (1.39–2.00) | <0.001 | 0.98 (0.76–1.27) | 0.874 | 1.72 (1.39–2.12) | <0.001 | 1.62 (1.13–2.26) | 0.033 | 1.43 (1.19–1.72) | <0.001 | 1.14 (0.90–1.45) | 0.288 |
|
| ||||||||||||
| PD-L1 expression | ||||||||||||
| • IC0/1 | Reference | Reference | Reference | Reference | Reference | Reference | ||||||
| • IC2/3 | 0.57 (0.45–0.73) | <0.001 | 0.80 (0.62–1.05) | 0.105 | 0.70 (0.54–0.91) | 0.007 | 0.74 (0.27–0.96) | <0.001 | 0.76 (0.61–0.95) | 0.018 | 0.84 (0.66–1.08) | 0.179 |
|
| ||||||||||||
| Bellmunt Risck Score | ||||||||||||
| • 0 | Reference | Reference | Reference | Reference | Reference | Reference | ||||||
| • 1+ | 2.14 (1.78–2.57) | <0.001 | 2.11 (1.72–2.58) | <0.001 | 2.13 (1.77–2.57) | <0.001 | 2.07 (1.53–2.59) | <0.001 | 1.62 (1.37–1.91) | <0.001 | 1.56 (1.30–1.87) | <0.001 |
Subgroup Analyses
The subgroup analysis of the irAEs group relied on 195 patients. These individuals were stratified in two groups, based on the administration of systemic corticosteroids or lack thereof. Specifically, the group of patients receiving systemic corticosteroids consisted of 99 individuals (51%) while the second group consisted of 96 individuals (49%). We repeated the aforementioned survival analyses to assess the potential confounding impact of systemic corticosteroids administration. The survival curves for PFS, CSS and OS are displayed in Supplementary Figures 2, 3, and 4, respectively. At the Log-rank test, there was no statistical difference between the two groups for PFS and OS (p= 0.55 and p =0.09, respectively). Median OS was 18.3 months for those patients receiving systemic corticosteroids and 12.1 months for the non-corticosteroid group. The 20-months OS rates for those receiving systemic corticosteroids and for those who did not were 44.9% and 41.1%, respectively. On multivariable Cox regression analysis (Supplementary Table 1), the administration of corticosteroids was not associated with worse oncological outcomes [PFS HR 0.92 95%CI (0.62–1.34); p= 0.629; CSS SHR 0.90 95% CI (0.60–1.36); p=0.630; OS HR 0.86 95% CI (0.51–1.64); p= 0.613].
Sensitivity Analyses
Overall, 771 patients were alive and not censored 64 days after randomization or start of treatment. Of these, 182 experienced irAEs. Kaplan Meier curves and Multivariable Cox regression are provided in Supplementary Figure 5, 6, and 7 and Supplementary Table 2 respectively, confirming that irAE development was associated with improved survival outcomes [PFS: HR 0.53 95% CI (0.41–0.67); CSS: SHR 0.53, 95% CI (0.41–0.67); OS: HR 0.54, 95% CI (0.43–0.72); all p< 0.001]. Similarly, Time-dependent Multivariable Cox regression analsysis are shown in Supplementary Table 3 [PFS: HR 0.73 95% CI (0.61–0.83); CSS: SHR 0.75, 95% CI (0.62–0.87); OS: HR 0.76, 95% CI (0.63–0.88); all p< 0.001]
DISCUSSION
The implementation of immune checkpoint inhibitors (ICIs) into clinical practice has mutated the treatment paradigm of several advanced malignancies. However, the proportion of individuals who respond to therapy remains less than optimal (4), and predictors of response at the patient and tumor level remain poorly understood. With that in mind, we aimed to assess whether irAEs were associated with improved prognosis in patients. IrAEs are thought to represent the bystander effects of activated T-cells in patients responding to ICIs. As such, they should be considered as a systemic manifestation of immune cells activation against self antigens, and possibly anti-tumoral T-cell mediated activity.
We gained access to data from a homogeneous population in order to test our hypothesis. In the patient population of IMvigor210 cohort 1 and cohort 2 and of IMvigor211 trials, we observed improved oncological outcomes, namely PFS, CSS and OS, if patients were experiencing irAEs while on atezolizumab treatment. Furthermore, this association did not seem to be affected by systemic corticosteroids administration. Our results deserve further considerations.
First, the group of patients who experienced irAEs and the group who did not shared similar baseline characteristics. In fact, no statistical difference was observed concerning age, sex, BMI, presence of mixed histologies, and metastatic burden (all p>0.05). Yet, the two groups differed in PD-L1 expression on immunohistochemistry, as this was significantly higher in patients experiencing irAEs. To the best of our knowledge, no translational research studies demonstrated any biological mechanism through which increased tumor PD-L1 expression may trigger irAEs. Thus, despite this observation being hypothesis generating, no causal association can be established. Notably, PD-L1 expression and response to ICIs have been inconsistent across several trials, as a significant proportion of PD-L1 negative patients respond to immune checkpoint blockade (16,17).
Second, we observed that the development of irAEs during atezolizumab treatment was inversely associated with the risk of disease progression [HR 0.50 95% CI (0.40–0.61) p<0.001], overall mortality [HR 0.51 95% CI (0.41–0.64) p<0.001], and cancer-specific mortality [SHR 0.55 95% CI (0.45–0.72) p<0.001] on multivariable Cox and competing risks regression analyses. The association between immune-mediated toxicities and survival outcomes has been described in other malignancies treated with anti-PD-1 or anti-PD-L1 antibodies (18–21).The first piece of evidence showing the relationship between the development of irAEs and treatment response comes from retrospective analyses with a composite population receiving ipilimumab for either metastatic melanoma or renal cell carcinoma (22). Similarly, individuals experiencing irAEs demonstrated an increased response rate when compared to those who did not experience irAEs, as shown in two multicentric large series of patients receiving nivolumab for metastatic melanoma (23,24). However, these findings were not consistent with the ones from a retrospective series, which failed to confirm the association between irAEs and oncological outcomes in metastatic lung cancer and melanoma patients receiving an ICI, after correcting for immortal time bias (15). Differently from our data, a subset of these patients received ICI combination therapy, ICI rechallenge after having developed side effects, or subsequent ICI regimens after progression(15).Conversely, these heterogeneity of treatment within the population might have introduced unaccounted variability, weakening the association between irAEs and oncological outcomes.
Third, the observed effect of irAEs on outcomes was not affected by systemic corticosteroid administration, [PFS HR 0.92 95%CI (0.62–1.34), p= 0.629; OS HR 0.86 95% CI (0.51–1.64), p= 0.613; CSS SHR 0.90 95% CI (0.60–1.36), p=0.630]. The oncological safety of corticosteroids during ICI treatment has been previously investigated in several retrospective studies with conflicting results across cancer types (22,25,26). One could expect that corticosteroid administration would diminish the effect of ICIs due to their immunosuppressive properties. Yet, our findings seem to suggest that once the anti-tumoral immune response is mounted, its effect is not impaired by corticosteroid administration. Indeed, further studies are needed to confirm our clinical observation. For what concerns urothelial cancer, the findings from the pooled analysis corroborated our results, as they showed little impact of corticosteroid administration on oncological outcomes(6). However, whether the duration of steroids treatment influences oncological outcomes still needs to be fully elucidated (27).
The main strength of our findings relies on the homogeneity of our data, which was derived from two studies that tested a single ICI, atezolizumab, in the same clinical setting with a unique definition of irAEs. In literature, the association of irAEs and overall survival benefit in urothelial carcinoma patients was investigated only in a single pooled analysis of seven prospective studies (6), which resulted in a rather heterogeneous population in terms of treatment regimens, protocols, and clinical definitions. In fact, the authors did not account for immortal time bias, a well-known statistical distortion that might have inflated the impact of irAE on oncological outcomes(15). In addition to this statistical issue, the impact of anti-PD-1 and anti-PD-L1 antibodies was not investigated relative to irAEs as a whole. Given the different mechanisms of action and the biological targets of anti-PD-1 and anti-PD-L1 antibodies, the authors did not adjust their OS analysis for PD-L1 expression, nor did they perform any sub-analysis accounting for it. Moreover, the definition of irAEs in the study by Maher et al. encompassed the adverse events of special interest (AESI) of the seven trials treated with a topical or systemic corticosteroid. From this definition, several issues emerged. First, the decision to use corticosteroids might have varied among study protocols and among the anti–PD-1/L1 antibodies, resulting in potential misclassification of patients experiencing irAEs and not receiving immunosuppressive agents. Second, the strict dependency of their irAE definition to the use of corticosteroids (either topical or systemic) might have resulted in the survival advantage being affected by steroid administration. Third, some trial protocols poorly defined which clinical manifestations were considered part of the AESI, partly impairing the generalizability of their results. Finally, according to the exclusion criteria of the IMvigor210 and IMvigor211 trials, patients suffering from pre-existing autoimmune diseases were excluded, while vitiligo, psoriasis, thyroid disease, adrenal disease, or controlled diabetes were permitted in some of the studies considered in the pooled analysis by Maher et al.
Despite several strengths, our study is not devoid of limitations. First, the development of irAEs may be partially independent from the treatment. Specifically, it can be related to pre-existing sub-clinical organ-specific inflammation (caused by viral, bacterial, and autoimmune etiologies) unmasked by treatment itself. Indeed, it was not possible to distinguish between response-independent and response-dependent irAEs. Second, irAEs were investigated as a whole, because we could not identify a specific immune manifestation that better correlates with the observed improved oncological outcomes. Third, we could not demonstrate any dose response relationship between irAE severity and ICI efficacy, given the paucity of patients experiencing grade III or IV irAEs. Fourth, the occurrence of irAEs cannot be implemented as selection tool for the optimal candidates to ICIs, as they clinically appear only after treatment initiation.
CONCLUSIONS
We observed that the development of irAEs in patients receiving atezolizumab for locally advanced or metastatic urothelial carcinoma was associated with improved oncological outcomes, in terms of progression-free, cancer-specific, and overall survival. Moreover, the administration of systemic corticosteroids did not seem to interfere with the antitumoral action of the activated lymphocytes, and thus did not affect treatment response. Further translational studies are needed to corroborate the latter clinical finding.
Supplementary Material
Supplementary Figure 1. Directed acyclic graphs showing potential confounding factors. Gray circles: confounding factors; yellow circle exposure; cyan circle: outcome of interest; white circle: other variable. Arrows: biasing paths
Supplementary Figure 3. Cancer-specific survival and other cause mortality cumulative incidence curves, stratified according to corticosteroid use among patients who experienced immune-related adverse events
Supplementary Figure 2. Progression-free survival curves, stratified according to corticosteroid use among patients who experienced immune-related adverse events.
Supplementary Figure 5. Sensitivity analysis. Progression-free survival curves, stratified according to the occurrence of immune-related adverse events. Landmark time: 64 days.
Supplementary Figure 6. Sensitivity analysis. Cancer-specific survival, stratified according to the occurrence of immune-related adverse events. Landmark time: 64 days.
Supplementary Figure 4. Overall survival curves, stratified according to corticosteroid use among patients who experienced immune-related adverse events.
Supplementary Figure 7. Sensitivity analysis. Overall survival curves, stratified according to the occurrence of immune-related adverse events. Landmark time: 64 days
Acknowledgment:
this publication is based on research using data from Hoffmann–La Roche that has been made available through Vivli, Inc. Vivli has not contributed to or approved, and is not in any way responsible for, the contents of this publication.
Funding:
This study was supported in part by the Cancer Center Support Grant to MD Anderson Cancer Center (grant P30 CA016672) from the National Cancer Institute. PM was supported by the MD Anderson Khalifa Scholar Award, the Andrew Sabin Family Foundation Fellowship, a Translational Research Partnership Award (KC200096P1) by the United States Department of Defense, an Advanced Discovery Award by the Kidney Cancer Association, a Translational Research Award by the V Foundation, the MD Anderson Physician-Scientist Award, and philanthropic donations by the family of Mike and Mary Allen.
Footnotes
Conflict of interest disclosure: PM has received honoraria for service on a Scientific Advisory Board for Mirati Therapeutics, Bristol Myers Squibb, and Exelixis; consulting for Axiom Healthcare Strategies; non-branded educational programs supported by Exelixis and Pfizer; and research funding for clinical trials from Takeda, Bristol Myers Squibb, Mirati Therapeutics, Gateway for Cancer Research, and UT MD Anderson Cancer Center.
AN has served as a consultant for Merck, Astra Zeneca, Janssen, Incyte, Roche, Rainier Therapeutics, Clovis Oncology, Bayer, and Astellas/Seattle Genetics, Ferring, Immunomedics. Has received research funding from Merck, Ipsen, and Astra Zeneca. Has received Travel expenses/Honoraria from Roche, Merck, Astra Zeneca, and Janssen
AM and GF own equities of Oltre Medical Consulting, Toulouse, France
REFERENCES
- 1.EAU guidelines on Muscle-invasive and Metastatic Bladder Cancer 2022. Available from: https://uroweb.org/guidelines/muscle-invasive-and-metastatic-bladder-cancer [Google Scholar]
- 2.Niglio SA, Jia R, Ji J, Ruder S, Patel VG, Martini A, et al. Programmed Death-1 or Programmed Death Ligand-1 Blockade in Patients with Platinum-resistant Metastatic Urothelial Cancer: A Systematic Review and Meta-analysis. Eur Urol. 2019. Dec;76(6):782–9. [DOI] [PubMed] [Google Scholar]
- 3.Powles T, Assaf ZJ, Davarpanah N, Banchereau R, Szabados BE, Yuen KC, et al. ctDNA guiding adjuvant immunotherapy in urothelial carcinoma. Nature. 2021. Jul;595(7867):432–7. [DOI] [PubMed] [Google Scholar]
- 4.Das S, Johnson DB. Immune-related adverse events and anti-tumor efficacy of immune checkpoint inhibitors. J Immunother cancer. 2019. Nov;7(1):306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Yadav K, Lewis RJ. Immortal Time Bias in Observational Studies. JAMA [Internet]. 2021;325(7):686–7. Available from: 10.1001/jama.2020.9151 [DOI] [PubMed] [Google Scholar]
- 6.Maher VE, Fernandes LL, Weinstock C, Tang S, Agarwal S, Brave M, et al. Analysis of the Association Between Adverse Events and Outcome in Patients Receiving a Programmed Death Protein 1 or Programmed Death Ligand 1 Antibody. J Clin Oncol Off J Am Soc Clin Oncol. 2019. Oct;37(30):2730–7. [DOI] [PubMed] [Google Scholar]
- 7.Powles T, Durán I, van der Heijden MS, Loriot Y, Vogelzang NJ, De Giorgi U, et al. Atezolizumab versus chemotherapy in patients with platinum-treated locally advanced or metastatic urothelial carcinoma (IMvigor211): a multicentre, open-label, phase 3 randomised controlled trial. Lancet (London, England). 2018. Feb;391(10122):748–57. [DOI] [PubMed] [Google Scholar]
- 8.Rosenberg JE, Hoffman-Censits J, Powles T, van der Heijden MS, Balar AV, Necchi A, et al. Atezolizumab in patients with locally advanced and metastatic urothelial carcinoma who have progressed following treatment with platinum-based chemotherapy: a single-arm, multicentre, phase 2 trial. Lancet (London, England). 2016. May;387(10031):1909–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Balar AV, Galsky MD, Rosenberg JE, Powles T, Petrylak DP, Bellmunt J, et al. Atezolizumab as first-line treatment in cisplatin-ineligible patients with locally advanced and metastatic urothelial carcinoma: a single-arm, multicentre, phase 2 trial. Lancet (London, England). 2017. Jan;389(10064):67–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.https://vivli.org. [Google Scholar]
- 11.Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. 2009. Jan;45(2):228–47. [DOI] [PubMed] [Google Scholar]
- 12.Bellmunt J, Choueiri TK, Fougeray R, Schutz FAB, Salhi Y, Winquist E, et al. Prognostic factors in patients with advanced transitional cell carcinoma of the urothelial tract experiencing treatment failure with platinum-containing regimens. J Clin Oncol Off J Am Soc Clin Oncol. 2010. Apr;28(11):1850–5. [DOI] [PubMed] [Google Scholar]
- 13.Services H Common Terminology Criteria for Adverse Events. Definitions. 2020;2009. [Google Scholar]
- 14.Shapiro DD, Msaouel P. Causal Diagram Techniques for Urologic Oncology Research. Clin Genitourin Cancer. 2021. Jun;19(3):271.e1–271.e7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ghisoni E, Wicky A, Bouchaab H, Imbimbo M, Delyon J, Gautron Moura B, et al. Late-onset and long-lasting immune-related adverse events from immune checkpoint-inhibitors: An overlooked aspect in immunotherapy. Eur J Cancer. 2021. May;149:153–64. [DOI] [PubMed] [Google Scholar]
- 16.Powles T, Csőszi T, Özgüroğlu M, Matsubara N, Géczi L, Cheng SY-S, et al. Pembrolizumab alone or combined with chemotherapy versus chemotherapy as first-line therapy for advanced urothelial carcinoma (KEYNOTE-361): a randomised, open-label, phase 3 trial. Lancet Oncol. 2021. Jul;22(7):931–45. [DOI] [PubMed] [Google Scholar]
- 17.Powles T, van der Heijden MS, Castellano D, Galsky MD, Loriot Y, Petrylak DP, et al. Durvalumab alone and durvalumab plus tremelimumab versus chemotherapy in previously untreated patients with unresectable, locally advanced or metastatic urothelial carcinoma (DANUBE): a randomised, open-label, multicentre, phase 3 trial. Lancet Oncol. 2020. Dec;21(12):1574–88. [DOI] [PubMed] [Google Scholar]
- 18.Paderi A, Giorgione R, Giommoni E, Mela MM, Rossi V, Doni L, et al. Association between Immune Related Adverse Events and Outcome in Patients with Metastatic Renal Cell Carcinoma Treated with Immune Checkpoint Inhibitors. Cancers (Basel). 2021. Feb;13(4). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Walker KJ, Bouzubar N, Robertson J, Ellis IO, Elston CW, Blamey RW, et al. Immunocytochemical localization of estrogen receptor in human breast tissue. Cancer Res. 1988. Nov;48(22):6517–22. [PubMed] [Google Scholar]
- 20.Toi Y, Sugawara S, Kawashima Y, Aiba T, Kawana S, Saito R, et al. Association of Immune-Related Adverse Events with Clinical Benefit in Patients with Advanced Non-Small-Cell Lung Cancer Treated with Nivolumab. Oncologist. 2018. Nov;23(11):1358–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Rogado J, Sánchez-Torres JM, Romero-Laorden N, Ballesteros AI, Pacheco-Barcia V, Ramos-Leví A, et al. Immune-related adverse events predict the therapeutic efficacy of anti-PD-1 antibodies in cancer patients. Eur J Cancer. 2019. Mar;109:21–7. [DOI] [PubMed] [Google Scholar]
- 22.Beck KE, Blansfield JA, Tran KQ, Feldman AL, Hughes MS, Royal RE, et al. Enterocolitis in patients with cancer after antibody blockade of cytotoxic T-lymphocyte-associated antigen 4. J Clin Oncol Off J Am Soc Clin Oncol. 2006. May;24(15):2283–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Freeman-Keller M, Kim Y, Cronin H, Richards A, Gibney G, Weber JS. Nivolumab in Resected and Unresectable Metastatic Melanoma: Characteristics of Immune-Related Adverse Events and Association with Outcomes. Clin cancer Res an Off J Am Assoc Cancer Res. 2016. Feb;22(4):886–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Weber JS, Hodi FS, Wolchok JD, Topalian SL, Schadendorf D, Larkin J, et al. Safety Profile of Nivolumab Monotherapy: A Pooled Analysis of Patients With Advanced Melanoma. J Clin Oncol Off J Am Soc Clin Oncol. 2017. Mar;35(7):785–92. [DOI] [PubMed] [Google Scholar]
- 25.Arbour KC, Mezquita L, Long N, Rizvi H, Auclin E, Ni A, et al. Impact of Baseline Steroids on Efficacy of Programmed Cell Death-1 and Programmed Death-Ligand 1 Blockade in Patients With Non-Small-Cell Lung Cancer. J Clin Oncol Off J Am Soc Clin Oncol. 2018. Oct;36(28):2872–8. [DOI] [PubMed] [Google Scholar]
- 26.Faje AT, Lawrence D, Flaherty K, Freedman C, Fadden R, Rubin K, et al. High-dose glucocorticoids for the treatment of ipilimumab-induced hypophysitis is associated with reduced survival in patients with melanoma. Cancer. 2018. Sep;124(18):3706–14. [DOI] [PubMed] [Google Scholar]
- 27.Eggermont AMM, Kicinski M, Blank CU, Mandala M, Long GV, Atkinson V, et al. Association Between Immune-Related Adverse Events and Recurrence-Free Survival Among Patients With Stage III Melanoma Randomized to Receive Pembrolizumab or Placebo: A Secondary Analysis of a Randomized Clinical Trial. JAMA Oncol. 2020. Apr;6(4):519–27. [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
Supplementary Figure 1. Directed acyclic graphs showing potential confounding factors. Gray circles: confounding factors; yellow circle exposure; cyan circle: outcome of interest; white circle: other variable. Arrows: biasing paths
Supplementary Figure 3. Cancer-specific survival and other cause mortality cumulative incidence curves, stratified according to corticosteroid use among patients who experienced immune-related adverse events
Supplementary Figure 2. Progression-free survival curves, stratified according to corticosteroid use among patients who experienced immune-related adverse events.
Supplementary Figure 5. Sensitivity analysis. Progression-free survival curves, stratified according to the occurrence of immune-related adverse events. Landmark time: 64 days.
Supplementary Figure 6. Sensitivity analysis. Cancer-specific survival, stratified according to the occurrence of immune-related adverse events. Landmark time: 64 days.
Supplementary Figure 4. Overall survival curves, stratified according to corticosteroid use among patients who experienced immune-related adverse events.
Supplementary Figure 7. Sensitivity analysis. Overall survival curves, stratified according to the occurrence of immune-related adverse events. Landmark time: 64 days