PD‐1/L1 inhibitors can improve but not replace chemotherapy for advanced urothelial carcinoma: A systematic review and network meta‐analysis

Abstract Background Programmed cell death‐1/ligand 1 inhibitors are a new treatment strategy for advanced urothelial carcinoma. Therefore, a comparative evaluation of their efficacy and toxicity compared with chemotherapy is necessary. Methods We comprehensively searched PubMed, Web of Science, Embase, and Cochrane Library databases and performed a meta‐analysis of randomized controlled trials up to July 2021. We considered overall survival as the primary outcome, and progression‐free survival, objective response rate, and treatment‐related adverse events as secondary outcomes. Results Overall, 3584 patients from five studies were evaluated. Compared with first‐line chemotherapy, programmed cell death‐1/ligand 1 inhibitors were significantly associated with worse progression‐free survival (p < 0.001) and adverse objective response rates (p < 0.001). However, the treatments were not significantly different in terms of overall survival (p = 0.33). Compared with second‐line chemotherapy, programmed cell death‐1/ligand 1 inhibitors significantly improved overall survival (p < 0.001), and there was no statistically significant difference in progression‐free survival (p = 0.89) or objective response rate (p = 0.34). Compared with chemotherapy, programmed cell death‐1/ligand 1 inhibitors were well tolerated (first‐line chemotherapy: p < 0.001; second‐line chemotherapy: p < 0.001). Conclusions The efficacy of programmed cell death‐1/ligand 1 inhibitors in patients with advanced urothelial carcinoma is not superior to that of first‐line platinum‐based chemotherapy but is better than second‐line chemotherapy; however, programmed cell death‐1/ligand 1 inhibitors are safer than first‐ and second‐line chemotherapy and have a broader prospect for use in combination therapy.

Molecular Mechanism and Translational Medicine of Guangzhou Bureau of Science and Information Technology, Grant/Award Number: 013-163 programmed cell death-1/ligand 1 inhibitors were well tolerated (first-line chemotherapy: p < 0.001; second-line chemotherapy: p < 0.001).

Conclusions:
The efficacy of programmed cell death-1/ligand 1 inhibitors in patients with advanced urothelial carcinoma is not superior to that of first-line platinum-based chemotherapy but is better than second-line chemotherapy; however, programmed cell death-1/ligand 1 inhibitors are safer than first-and second-line chemotherapy and have a broader prospect for use in combination therapy.

| INTRODUCTION
In 2020, there were 573,278 and 212,536 new cases of, and deaths from, bladder cancer, respectively [1], of which approximately 90% were urothelial carcinoma (UC) [2].The 5-year survival rate of locally advanced or metastatic UC is only 6.4% [3].However, limited treatments are available for advanced UC, especially patients with unresectable locally advanced disease who can only receive chemotherapy.
Cisplatin has been considered the most effective cytotoxic treatment for advanced UC since 1980 [4].Platinum-based chemotherapy is now the first-line treatment for advanced UC.Although chemotherapy plays a pivotal role in UC treatment, its development has plateaued.Over the past four decades, various doses, schedules, and sequences of platinum-based regimens have been evaluated, even for platinum-free double line or newer agent combinations.However, none of the new therapies have been an improvement on standard chemotherapy regimens [5].Furthermore, the therapeutic effect of second-line chemotherapy drugs is unsatisfactory, with the median overall survival (OS) afforded by vinflunine, paclitaxel, and docetaxel being only 5-7 months, and less than 10% of patients achieving an objective response [6].Thus, to prolong survival in patients with advanced UC, new therapeutic schedules are necessary.
Immunotherapy has rapidly advanced recently and is now widely used in cancer treatment.The US Food and Drug Administration (FDA) has approved five programmed cell death-1/ligand 1 (PD-1/L1) inhibitors (atezolizumab, pembrolizumab, nivolumab, durvalumab, and avelumab) for patients with advanced UC in whom platinum-based chemotherapy has failed.Furthermore, atezolizumab and pembrolizumab have been approved as first-line treatments for patients ineligible for platinumbased therapy [7].However, the results of some recent randomized clinical trials (RCTs) on the efficacy and safety of PD-1/L1 inhibitors found contradictory results.Therefore, we conducted a systematic review and metaanalysis to clarify the efficacy and safety of PD-1/L1 inhibitors compared with those of first-and second-line chemotherapy to guide clinical practice.

| Literature search
Two authors (Longkun Mao and Meihua Yang) independently retrieved literature published by December 31, 2022 from PubMed, Cochrane Library, Embase, and Web of Science Core Collection without regional restrictions.The search was performed to identify all articles on the use of PD-1/L1 inhibitors for treating advanced UC.The following terms were included in the search (all fields): (immunotherapy OR pembrolizumab OR nivolumab OR atezolizumab OR avelumab OR durvalumab) AND (advanced UC OR metastatic UC).The related article function was also used to expand the search scope.The computer search supplemented a manual search of the reference lists of all retrieved research, review articles, and conference abstracts.

| Study eligibility
The inclusion criteria were as follows: (a) studies on advanced UC patients treated with a PD-1/L1 inhibitor; (b) RCTs; (c) studies that reported data on median OS or progression-free survival (PFS) or objective response rate (ORR) and treatment-related adverse events (TRAEs); (d) studies containing at least two arms of the chemotherapy and immunotherapy groups; and (e) publications in English.Studies that did not meet the inclusion criteria were excluded.When more than one publication reported on the same study population, studies with the most updated and/or comprehensive data were included.

| Data extraction
Data extraction and assessment were performed independently by two different investigators (Longkun Mao and Meihua Yang), and disagreements were resolved through consensus or by discussion with a third author (Xinxiang Fan).The following information was extracted from each study: (a) first author and publication year, study phase, study name, and clinical trial number; (b) whether neoadjuvant therapy was used, whether the patients were eligible to receive platinum-based therapy, name and dose of the PD-1/L1 inhibitors used, and name of the chemotherapeutic drugs used; (c) number of patients who received each treatment, and median follow-up duration, median OS, median PFS, and ORR in the intention-to-treat population; and (d) numbers of TRAEs, numbers of patients with TRAEs of different grades, and the most common TRAEs.

| Quality assessment
Two authors (Mao and Yang) independently assessed the quality of all included studies using RevMan 5.3 (Cochrane Collaboration), while a third author (Fan) addressed disagreements by consultation.The Cochrane Collaboration tool was used to assess the risk of bias, including sequence generation, incomplete outcome data, allocation concealment, selective outcome reporting, blinding of participants, personnel and outcome assessors, and other sources of bias.

| Statistical analysis
The meta-analysis endpoints were median OS, median PFS, ORR, and TRAEs.Forest plots were used for the summary variables of survival data and dichotomous outcomes, and all results were reported with 95% confidence intervals (CIs).The two-side Q statistic and I 2 test were used to assess statistical heterogeneity between different studies; p-values ≤ 0.05 were considered significant.The fixed-effects model was chosen if there was no significant heterogeneity between studies; otherwise, the random-effects model was used.

| EVIDENCE SYNTHESIS
We ultimately selected five eligible studies that included 3584 patients for analysis (Figure 1), of whom 1752 (48.9%) received PD-1/L1 inhibitors [8][9][10][11][12]; hence, we included them in the meta-analysis and divided them into three subgroups based on the PD-1/L1 inhibitors used.Examination of these studies and the references listed in the review articles did not lead to any further evaluation studies.The agreement rates of the two reviewers regarding study selection and evaluation of trial quality were both 100%.

| Characteristics of eligible studies
Characteristics of the included studies are listed in Table 1.The five meta-analyses were divided into the following groups: immunotherapy compared with firstline chemotherapy (cisplatin or carboplatin combined with gemcitabine) and immunotherapy compared with second-line chemotherapy (paclitaxel, docetaxel, or vinflunine).The first-line chemotherapy group contained three studies and included 2159 patients eligible for platinum-based chemotherapy.Only two studies were in the second-line chemotherapy group, with a total of 1473 patients who had shown progression after platinum chemotherapy or were ineligible for platinum-based chemotherapy.The outcomes of all included studies are shown in Table 2.

| Quality assessment of the included studies
The risk of bias in the analyzed RCTs was evaluated and summarized in Appendix 1. Generally, a low risk of bias was identified, which met the general requirement for meta-analyses.
For patients receiving PD-1/L1 inhibitors, the most common TRAEs were fatigue, pruritus, and anemia, whereas chemotherapy was associated with a greater risk of neutropenia and reductions in white blood cell and platelet counts.

| DISCUSSION
This study is the first systematic meta-analysis to explore the advantages and disadvantages of immunotherapy compared with first-and second-line chemotherapy for advanced UC.This analysis included 3584 patients from five well-organized RCTs and indirectly   tumors.PD-1/L1 inhibitors can relieve this inhibition and have achieved widespread and successful application since atezolizumab and pembrolizumab were approved by the FDA in 2017 for UC patients who were not suitable for cisplatin [7].Our results showed that there was no statistically significant difference in OS between immunotherapy (8.6-15.7 months) and first-line chemotherapy (7.4-14.3months).For advanced UC patients who had progressed after a platinum-containing regimen, the median OS after receiving PD-1/L1 inhibitors was 8.6-10.3months, which showed a significant survival benefit compared with second-line chemotherapy (7.4-8.0 months).
Single-agent immunotherapy has demonstrated OS benefits that were not superior to those obtained with first-line chemotherapy in advanced UC patients, but the patients tolerated the treatments well, and the mechanism is complementary to that of other treatments.Therefore, researchers have felt that combination therapy is a viable future direction, with some trials of combination therapy achieving gratifying results.
The combination of cytotoxic chemotherapy and immunotherapy has been successfully used to treat cancer.For example, Galsky et al. [13] proved that gemcitabine plus cisplatin (GC) plus ipilimumab lead to a significant expansion of peripheral blood CD4 cells in metastatic UC patients, which is associated with improved survival.Recently, the KEYNOTE-361 (NCT02853305) [10] and IMvigor 130 (NCT02807636) [9] trials reported that the addition of PD-1/L1 inhibitors (pembrolizumab and atezolizumab, respectively) to first-line chemotherapy could achieve improvements in OS (16-17 vs. 13.4-14.3months) and PFS (8.2-8.3 vs. 6.3-7.1 months) compared with chemotherapy alone in advanced UC.Moreover, the rate of complete response to combination therapy was almost double that of chemotherapy, and the safety of combination therapy was similar to that of chemotherapy alone in the IMvigor 130 trial.However, PD-1/ L1 inhibitors plus chemotherapy did not show consistent benefits compared with chemotherapy alone, arguing for the current platinum-based chemotherapy to remain the gold-standard in the first-line setting for the management of advanced UC.
In addition to cytotoxic drugs, the combination of targeted therapy with immunotherapy is favored by many researchers.Fibroblast growth factor receptor (FGFR) helps cells grow, survive, and multiply [14].FGFR inhibitors such as erdafitinib, infigratinib, and rogaratinib have been proven to be effective in previous clinical trials (median OS of patients treated with erdafitinib and infigratinib was 13.8 months and 7.8 months, respectively, while the OR of patients treated with rogaratinib has not been reported), and erdafitinib was recently approved by the FDA for use in patients with metastatic UC who progressed on platinum-based chemotherapy [15].FGFR inhibitors are undoubtedly becoming a valuable part of the treatment for advanced bladder cancer.Currently, many clinical trials are being conducted to explore how best to combine FGFR inhibitors with PD-1/L1 inhibitors to bring more survival benefits to patients [16].Another common target is nectin-4, a type 1 transmembrane protein that belongs to a family of cell adhesion molecules that are highly expressed in a variety of cancers, particularly in UC.Enfortumab vedotin (EV, also known as ASG-22CE) is a monoclonal antibody-drug conjugate (ADC) that selectively binds to nectin-4-expressing cells, results in their apoptotic death [17].The single-agent efficacy of EV in patients with metastatic UC had been confirmed in two early phase 1 clinical trials [18,19] (ORR: 40%-33%, PFS: 17-23.1 weeks, respectively).These results prompted the FDA to accelerate the approval of EV for patients with advanced UC who showed disease progression after platinum-based chemotherapy and immunotherapy in December 2019 [20].Surprisingly, a study of EV with pembrolizumab in advanced UC (EV-103) reported that the confirmed investigator-assessed ORR was 73.3% [21].Additionally, the ongoing EV-302 study is comparing the efficacy and safety of this combination with GC chemotherapy.
Cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) is another immune checkpoint that has been shown to downregulate immune responses.Combinations of PD-1/L1 and CTLA-4 inhibitors such as nivolumab plus ipilimumab have been proven to show high ORR (25.6%-38.0%) in previously treated metastatic UC [22].Additionally, an ongoing phase 3 trial [23] is expected to bring stronger evidence to support nivolumab plus ipilimumab becoming the first-line treatment for metastatic UC.The combination of durvalumab plus tremelimumab was reported to demonstrate benefits in patients with high PD-1/L1 expression in a phase 1 clinical trial [24].Although a series of subsequent trials did not achieve the primary end points, durvalumab plus tremelimumab showed higher OS and ORR than durvalumab alone or chemotherapy [12,[25][26][27][28].
Although immunotherapy has achieved much success, researchers continue to face many challenges regarding the ultimate goal of "curing cancer."For example, only approximately a quarter of patients (ORR: 23%-30%) show benefits from immunotherapy, and some of them will ultimately face tumor progression (PFS: 1.9-3.5 months).This phenomenon is known as primary or secondary immune escape, and occurs when patients who have not previously received immunotherapy fail to respond to it or have stopped responding to immunotherapy after previously showing a response [29].Primary immune escape may be associated with the suppression of T-cell responses through multiple immunosuppressive mechanisms, which are related to differences in the immune environment at baseline.Mariathasan et al. [30] demonstrated that single-agent immunotherapy performed poorly in immune-excluded UC with elevated TGFβ signaling, and that combining anti-TGFβ agents with immunotherapy could overcome primary immune escape.Additionally, for inflamed tumor mutational burden (TMB)-low tumors, antivascular endothelial growth factor (VEGF) therapy could overcome myeloid-mediated immune suppression [31].Nadal et al. [32] reported that cabozantinib (VEGF targeted therapy) plus nivolumab yielded a favorable ORR (50%) and median PFS (24.1 months) in the treatment of immunotherapy-naïve patients with chemotherapy-refractory metastatic UC; even in immunotherapy-refractory patients, the ORR was 28%.The mechanisms of secondary escape may also be diverse but are generally associated with downregulation of immunogenic antigens and growth of cancer cell clones lacking immunogenic antigens.However, a lack of relevant tumor data has made studies on the drivers of secondary escape more challenging [29].A better understanding of the different mechanisms of immune escape will facilitate the development of more targeted and effective treatments for different cancer types with distinct phenotypes at different stages.
Safety is another aspect of efficacy that is also important when evaluating a new type of treatment.Immunotherapy activates patients' own immune system, potentially affecting multiple organs throughout the body, especially the skin, gastrointestinal system, liver, and endocrine system.These adverse effects were usually grade 1/2 TRAEs, with serious adverse reactions only accounting for approximately 15% of TRAEs, and thus most patients could tolerate and continue immunotherapy, which was very different from the findings of cytotoxic chemotherapy.In other evaluations of the occurrence of serious adverse events, Weber et al. [33] observed that C-reactive protein and interleukin-6 levels above the median baseline values of these biomarkers were significantly associated with adverse events following immunotherapy.Additionally, a prospective cohort [34] study demonstrated that the predictive factors for grade 4/5 and fatal immunological toxicities were elevated neutrophil/lymphocyte ratio, performance status ≥2, and lung cancer, which could help identify patients at risk of severe adverse events.Compared with the cytotoxicity of chemotherapeutics, the effect of immunotherapy on normal cells in the patient's body was acceptable.Thus, despite the minor differences in efficacy, we were more inclined to choose safer immunotherapy as the first-line treatment rather than chemotherapy.Finally, the safety of immunotherapy also provides a basis for combinations with other treatment modalities.
The main limitation of this study was that although all the included studies were RCTs, the number of studies was relatively small, and there was only one study in each subgroup.Second, the analyses of OS, PFS, ORR, and TRAEs included all patients in the intention-to-treat analysis, and no further analysis was performed on their biomarkers.Additionally, when we compared the most common TRAEs, we did not count the specific number of patients with each adverse reaction but roughly ranked the TRAEs according to the number of studies that reported the adverse reaction.

| CONCLUSIONS
This meta-analysis found that among patients with advanced UC, PD-1/L1 inhibitors showed no superior efficacy but better safety than first-line platinum-based chemotherapy, offering broader prospects for combination therapy.

F
I G U R E 1 search flow diagram for this meta-analysis.