Current Strategies and Novel Therapeutic Approaches for Metastatic Urothelial Carcinoma

Urothelial carcinoma (UC) is a frequent cause of cancer-related deaths worldwide. Metastatic UC has been historically associated with poor prognosis, with a median overall survival of approximately 15 months and a 5-year survival rate of 18%. Although platinum-based chemotherapy remains the mainstay of medical treatment for patients with metastatic UC, chemotherapy clinical trials produced modest benefit with short-lived, disappointing responses. In recent years, the better understanding of the role of immune system in cancer control has led to the development and approval of several immunotherapeutic approaches in UC therapy, where immune checkpoint inhibitors have been revolutionizing the treatment of metastatic UC. Because of a better tumor molecular profiling, FGFR inhibitors, PARP inhibitors, anti-HER2 agents, and antibody drug conjugates targeting Nectin-4 are also emerging as new therapeutic options. Moreover, a wide number of trials is ongoing with the aim to evaluate several other alterations and pathways as new potential targets in metastatic UC. In this review, we will discuss the recent advances and highlight future directions of the medical treatment of UC, with a particular focus on recently published data and ongoing active and recruiting trials.


Introduction
Urothelial carcinoma (UC) is a common cancer worldwide, with nearly half a million of new diagnoses annually [1]. Although UC includes a group of tumors of the bladder, renal pelvis, ureter and Although cisplatin-based regimens are considered the standard first-line treatment in advanced or metastatic UC, more than 50% of patients are ineligible for cisplatin in clinical practice [23]. For this non-negligible group of patients, carboplatin plus gemcitabine has been considered the standard treatment based on the results of the EORTC 30,986 trial, with several other combinations and agents showing less favorable safety profiles and inferior outcomes compared to cisplatin-based first-line therapy [24][25][26]. Thus, the modest survival benefits observed with available treatment options highlighted the need for new effective strategies [27] and for this purpose, following small phase I trials, the role of ICIs as front-line treatment in cisplatin-ineligible patients was investigated in KEYNOTE-052 and IMvigor210 trials [28,29].
The KEYNOTE-052 [28] was a phase II trial aimed to evaluate the safety and efficacy of pembrolizumab monotherapy (200 mg flat dose every three weeks) in 370 chemo-naive, cisplatin-ineligible patients. In this setting pembrolizumab, a highly selective humanized monoclonal IgG4 isotype antibody against PD-1 protein, produced an overall response rate (ORR) of 24% with 5% of complete response (CR). Interestingly, the magnitude of ORR and survival benefit was related to programmed death ligand-1 (PD-L1) expression: in fact, in patients with PD-L1 expression combined positive score (CPS) ≥ 10%, pembrolizumab resulted in improved survival, with a median OS of 18.5 months versus 11.5 months in overall cohort. Finally, the CPS ≥ 10% population reported higher ORR (37%) compared to the CPS < 10% subgroup of patients (ORR = 18%).
Although cisplatin-based regimens are considered the standard first-line treatment in advanced or metastatic UC, more than 50% of patients are ineligible for cisplatin in clinical practice [23]. For this non-negligible group of patients, carboplatin plus gemcitabine has been considered the standard treatment based on the results of the EORTC 30,986 trial, with several other combinations and agents showing less favorable safety profiles and inferior outcomes compared to cisplatin-based first-line therapy [24][25][26]. Thus, the modest survival benefits observed with available treatment options highlighted the need for new effective strategies [27] and for this purpose, following small phase I trials, the role of ICIs as front-line treatment in cisplatin-ineligible patients was investigated in KEYNOTE-052 and IMvigor210 trials [28,29].
The KEYNOTE-052 [28] was a phase II trial aimed to evaluate the safety and efficacy of pembrolizumab monotherapy (200 mg flat dose every three weeks) in 370 chemo-naive, cisplatin-ineligible patients. In this setting pembrolizumab, a highly selective humanized monoclonal IgG4 isotype antibody against PD-1 protein, produced an overall response rate (ORR) of 24% with 5% of complete response (CR). Interestingly, the magnitude of ORR and survival benefit was related to programmed death ligand-1 (PD-L1) expression: in fact, in patients with PD-L1 expression combined positive score (CPS) ≥ 10%, pembrolizumab resulted in improved survival, with a median OS of 18.5 months versus 11.5 months in overall cohort. Finally, the CPS ≥ 10% population reported higher ORR (37%) compared to the CPS < 10% subgroup of patients (ORR = 18%).
The IMvigor210 trial [29] was a 2-cohort Phase 2 study; while cohort 2 assessed atezolizumab in a post-platinum setting, in cohort 1 the anti-PD-L1 agent was tested as first-line treatment in cisplatin-ineligible subjects. Total of 119 untreated patients were included in cohort 1 and received atezolizumab, 1200 mg flat dose every three weeks, achieving an ORR of 23% with CR and partial response (PR) of 9% and 12% respectively, regardless of PD-L1 expression. Clinical activity of Cancers 2020, 12, 1449 4 of 44 atezolizumab was higher than those observed with systemic chemotherapies traditionally used in this setting, in respect of whom the anti-PD-L1 agent showed also a more manageable safety profile.
Based on the aforementioned studies, pembrolizumab and atezolizumab were approved by the US Food and Drug Administration (FDA) and European Medicines Agency (EMA) for front-line use in cisplatin unfit patients affected by advanced or metastatic UC. However, the use of pembrolizumab and atezolizumab has been subsequently restricted, following early data from KEYNOTE-361 and IMvigor130 phase III trials which, as we shall explain later, are currently investigating combination chemo-immunotherapy in advanced or metastatic UC. In these two trials, patients with low expression of PD-L1 receiving single-agent ICI experienced worse survival compared to patients receiving standard chemotherapy [30,31].
Following platinum-based chemotherapy, large proportions of patients are either non-responders or relapsed, and therefore proceed for second-line treatment [32]. Until some years ago, taxanes or vinflunine were considered standard second-line treatments, despite disappointing ORRs and an overall modest clinical benefit [33,34]. In this scenario, recent results of Phase I to III studies with agents targeting PD-1 and PD-L1 have led to fast approval of ICIs as second-line treatments. In particular, five ICIs (two anti-PD-1 agents-pembrolizumab and nivolumab-and three anti-PD-L1 agents-atezolizumab, durvalumab, and avelumab) have been granted approval by FDA for patients with advanced or metastatic UC whose disease progressed during or following platinum-based chemotherapy [35]. Conversely, despite FDA has granted approval for the aforementioned agents, pembrolizumab is the only ICI that showed a survival benefit in a phase III randomized clinical trial and whose activity is supported by higher levels of evidence [36].
The approval of pembrolizumab in post-platinum setting was granted based on the results of the KEYNOTE-045 trial [37]. This phase III, open-label, randomized trial compared pembrolizumab (flat dose of 200 mg every three weeks) with chemotherapy by investigators' choice, including vinflunine and taxanes, in patients who recurred or progressed after a platinum-based regimen. A higher ORR was observed in patients treated with pembrolizumab (21.1% vs. 11.4% of the chemotherapy arm); moreover, an OS benefit was observed, regardless of PD-L1 expression (in the overall population 10.3 and 7.4 months, in the immunotherapy and chemotherapy arm, respectively, hazard ratio (HR) 0.73; 95% confidence interval (CI) 0.59 to 0.91; p = 0.002). Finally, pembrolizumab was associated with fewer grade 3-4 adverse events compared to vinflunine, paclitaxel, and docetaxel.
Instead, the activity of atezolizumab was tested in the phase II IMvigor210 and the phase III IMvigor211 trials [38,39]. As stated above, the IMvigor210 trial was a 2-cohort phase II study aimed to evaluate the efficacy and safety of atezolizumab (1200 mg flat dose every three weeks) in untreated, cisplatin-ineligible patients (cohort 1) as well as in patients whose disease was refractory to platinum-based chemotherapy (cohort 2) [38]. In the cohort 2, including 315 eligible subjects, an ORR of 15% was observed, with a sustained response duration and an acceptable safety profile; moreover, in patients presenting PD-L1 expression ≥5% the ORR was higher (27%) and the survival benefit longer compared to the PD-L1 ≥ 1 and < 5% cohort and the PD-L1 < 1% group. On the basis of these promising findings, the role of atezolizumab was further assessed in the confirmatory phase III, open-label, randomized IMvigor211 trial [39], which compared atezolizumab to chemotherapy by investigators' choice, including vinflunine and taxanes, in patients who recurred or progressed after a platinum-based regimen. The primary endpoint, OS in patients with PD-L1 expression ≥ 5%, did not significantly differ between the two arms, with a median OS of 11.1 and 10.6 months in atezolizumab and chemotherapy arms, respectively (HR 0.87; 95% CI 0.63-1.21; p = 0.41). Despite the negative primary endpoint, IMvigor211 provided useful data in terms of median duration of response, which was significantly higher in the ICI arm (15.9 vs. 8.3 months; HR 0.57; 95% CI 0. 26-1.26) and in terms of toxicity, with the PD-L1 inhibitor confirming a manageable safety profile. Finally, the exploratory analysis of the intention to treat population showed a survival benefit for atezolizumab (HR 0.85; 95% CI 0.73-0.99). Nivolumab is a human monoclonal IgG4 antibody that blocks the human PD-1 receptor, whose efficacy in the post-platinum setting was explored in the CheckMate 275 trial [40]; in this phase II trial, nivolumab (240 mg flat dose every two weeks) showed an ORR of 20% with 2% CR in 270 patients affected by advanced or metastatic UC. With regard to PD-L1 expression, ORR was significantly higher in the subgroup of patients with PD-L1 expression ≥5% (28.4%) compared to the PD-L1 ≥ 1% (23.8%) and the PD-L1 negative (16.1%) cohorts.
A similar level of activity was observed with post-platinum avelumab and durvalumab in the multicohort phase Ib JAVELIN trial [41] and the single-arm, phase I/II Study 1108 [42], respectively. Avelumab (10 mg/kg every two weeks), an anti-PD-L1 antibody that blocks the binding of PD-L1 to PD-1, reported an ORR of 17% with 6% CR in platinum-refractory or cisplatin unfit patients; interestingly, in PD-L1 negative subgroup ORR fell to 9% while reached the 40% in PD-L1 ≥ 5% patients.
Similarly, considering the cutoff of 25% of PD-L1 expression (assessed with immunohistochemistry (IHC) on tumor tissue via Ventana SP263 assay) in Study 1108, the subgroup of patients with PD-L1 high achieved higher response rates and survival benefit compared to PD-L1 low cohort (20 vs. 8 months) with the anti-PD-L1 human IgG1 durvalumab [42]. In Study 1108, patients were administered durvalumab intravenous infusion, 10 mg/kg every 2 weeks.
Additional data from a number of ongoing prospective clinical trials will help to confirm the activity of ICIs in previously treated and untreated patients [43]; in the era of precision, tailor-made oncology, several questions are still unanswered, including the identification of predictive biomarkers, sequential treatment strategies, and proper selection of patients in advanced or metastatic UC. A non-negligible unanswered question is how to assess PD-L1 expression. For example, in KEYNOTE-052 and IMvigor210 PD-L1 cutoff was different and it was assessed differently; in KEYNOTE-052 PD-L1 positive tumors were those presenting a CPS ≥ 10% and PD-L1 expression in formalin-fixed, paraffine-embedded tissue was determined using the PD-L1 clinical trial assay (PD-L1 IHC 22C3 pharmDx assay; Agilent Technologies, Carpinteria, CA, USA). Differently, in the IMvigor210 trial the VENTANA SP142 immunohistochemistry assay (Ventana Medical Systems, Inc.; Tucson, AZ, USA) was used to evaluate PD-L1 expression on tumor-infiltrating immune cells (IC) and a scoring criteria designated tumors as IC0, IC1, or IC2/3 (PD-L1 expression on <1%, ≥1% and <5%, or ≥5% of IC, respectively).

Target Therapies
In the recent years, genomic characterization of advanced-stage UC has given an insight on which are molecular drivers at the basis of the oncogenesis and progression of UC and that could be potentially targetable ( Figure 2) [44]. The Cancer Genome Atlas (TCGA) project for bladder cancer had the purpose to provide a comprehensive landscape of molecular alterations [45]. The first integrated analysis on 131 UC demonstrated statistically significant recurrent mutations in 32 genes. Furthermore, this analysis showed that 69% of the tumors presented potential therapeutic targets, of which 42% regarded the phosphatidylinositol-3-kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR) pathway and 44% in the receptor tyrosine kinase/MAPK pathway, and identified an in-frame activating FGFR3-TACC3 fusion in three tumors [45]. Alterations in the PI3K/AKT/mTOR pathway consisted in point mutations in PIK3CA (17%), mutation or deletion of TSC1 or TSC2 (9%), and overexpression of AKT3 (10%). Alterations in the receptor tyrosine kinase/RAS pathway included activation of FGFR3 (17%), amplification of EGFR (9%), mutations of ERBB3 (6%), and mutation or amplification of ERBB2 (9%).
The TCGA expanded cohort analysis on 412 MIBC that identified 58 significantly mutated genes and confirmed the high mutation rate of MIBC [46]. Moreover, RNA expression analysis identified five expression subtypes that may stratify response to different treatments: luminal-papillary (35%), luminal (6%), basal-squamous (35%), luminal-infiltrated (19%), and neuronal (5%) [46]. Recently, a consensus molecular classification of MIBC has been proposed on the basis of 1750 MIBC transcriptomic profiles from 18 datasets comparing six molecular classification schemes. Six molecular classes were identified: luminal papillary, luminal nonspecified, luminal unstable, stroma-rich, basal/squamous, and neuroendocrine-like [47]. This consensus classification has possible therapeutic implications. In fact, the different consensus classes are associated with different stromal components and genetic alteration that could possibly identify a subset of patients more likely to respond to immunotherapy or to target therapy. The identification of molecular alterations is of great importance since many target therapies are being studied for the management of advanced UT [48].
The TCGA expanded cohort analysis on 412 MIBC that identified 58 significantly mutated genes and confirmed the high mutation rate of MIBC [46]. Moreover, RNA expression analysis identified five expression subtypes that may stratify response to different treatments: luminal-papillary (35%), luminal (6%), basal-squamous (35%), luminal-infiltrated (19%), and neuronal (5%) [46]. Recently, a consensus molecular classification of MIBC has been proposed on the basis of 1750 MIBC transcriptomic profiles from 18 datasets comparing six molecular classification schemes. Six molecular classes were identified: luminal papillary, luminal nonspecified, luminal unstable, stroma-rich, basal/squamous, and neuroendocrine-like [47]. This consensus classification has possible therapeutic implications. In fact, the different consensus classes are associated with different stromal components and genetic alteration that could possibly identify a subset of patients more likely to respond to immunotherapy or to target therapy. The identification of molecular alterations is of great importance since many target therapies are being studied for the management of advanced UT [48].
FGFR1, FGFR2, FGFR3, FGFR4 are tyrosine kinases receptor that have been found altered in UC [49]. Activating FGFR3 mutations are most common in NMIBC, being identified in approximately two-third of these early stage tumors, while their frequency in MIBC is lower (less than 25%), including amplifications, mutations, and fusions in FGFR gene [50][51][52][53]. The activating FGFR3 mutation leads to ligand-independent receptor dimerization and constitutive downstream signal transduction [53]. The presence of activating point mutations in FGFR3 in early stage tumors is associated with favorable outcome [54]. Approximately 7% of UC present an amplification of FGFR1 [55]. FGFR1 has two splicing variants, FGFR1α and FGFR1β, that are equally expressed in normal urothelium, but the FGFR1β variant is predominant in UC and its expression correlates with tumor grade and stage [56]. The luminal-papillary subtype of the consensus classification is characterized FGFR1, FGFR2, FGFR3, FGFR4 are tyrosine kinases receptor that have been found altered in UC [49]. Activating FGFR3 mutations are most common in NMIBC, being identified in approximately two-third of these early stage tumors, while their frequency in MIBC is lower (less than 25%), including amplifications, mutations, and fusions in FGFR gene [50][51][52][53]. The activating FGFR3 mutation leads to ligand-independent receptor dimerization and constitutive downstream signal transduction [53]. The presence of activating point mutations in FGFR3 in early stage tumors is associated with favorable outcome [54]. Approximately 7% of UC present an amplification of FGFR1 [55]. FGFR1 has two splicing variants, FGFR1α and FGFR1β, that are equally expressed in normal urothelium, but the FGFR1β variant is predominant in UC and its expression correlates with tumor grade and stage [56]. The luminal-papillary subtype of the consensus classification is characterized by a high rate of FGFR3 mutations and translocations, suggesting that these tumors may respond to FGFR inhibitors [47]. Moreover, FGFR3 pathway was found to be activated in non-T-cell-inflamed tumors that are likely to present intrinsic resistance to ICIs [57]. Furthermore, immunotherapy seems to be less effective on TCGA luminal I subtype also based on an exploratory analysis of a phase 2 trial: luminal I cluster presented lower expression levels of CD8+ genes, lower PD-L1 immune cell or tumor cell expression, and lower responses to the anti-PD-L1 atezolizumab [38].
With these premises, multi-tyrosine kinase inhibitors targeting FGFR alterations have been studied in patients with metastatic UC [58]. The results of a phase 2 trial (BLC2001) testing the tyrosine kinase inhibitor of FGFR1-4 erdafitinib have been recently published [59]. In this trial, 99 patients with locally advanced or metastatic UC with FGFR3 mutation or FGFR2/3 fusion and progressed to at least one previous chemotherapy or treatment naïve if cisplatin ineligible were assigned to receive erdafitinib, 8 mg per day in a continuous regimen. The primary endpoint of the study was ORR. The treatment was found to be active with an ORR of 40% (3% with a complete response and 37% with a partial response). The median duration of progression-free survival (PFS) was 5.5 months and the median duration of OS was 13.8 months. Interestingly, the 22 patients previously treated with ICIs presented a response rate of 59%. Grade 3 or higher treatment-related adverse events were reported in nearly half the patients and the most common of any grade were hyperphosphatemia, stomatitis, diarrhea. FDA granted accelerated approval to erdafitinib for patients with FGFR3 or FGFR2 genetic alterations progressed during or following platinum-containing chemotherapy, including within 12 months of neoadjuvant or adjuvant platinum-containing chemotherapy.
Another pathway implicated in UC pathogenesis and progression is vascular endothelial growth factor receptors (VEGFR) 1 and 2 and their ligands (vascular endothelial growth factor, VEGF-A, -B, -C, and -D) [60,61]. Angiogenesis by microvessel quantification resulted to be an independent predictor of survival in patients with invasive bladder cancer and serum levels of VEGF have been correlated with tumor stage and grade, vascular invasion and presence of metastases [62][63][64].
VEGF/VEGFR inhibitors as single agents or in combination with chemotherapy have been investigated for the treatment of advanced UC. Single agent treatment with sorafenib, pazopanib, cabozantinib and sunitinib resulted to have limited activity and limited effect on clinical outcomes [65][66][67][68]. Similarly, combination therapies failed to be shown to be more active than chemotherapy alone: vandetanib combined with docetaxel or sunitinib associated with gemcitabine and cisplatin did not improve clinical activity and were more toxic [69,70].
The monoclonal antibody against VEGF Bevacizumab was evaluated in a phase II trial in association with gemcitabine and cisplatin in first line of therapy for metastatic UC: the combination treatment showed an ORR of 72% and an OS of 19.1 months [71]. Unfortunately, the subsequent phase III trial (CALGB-90601 Alliance) failed to show an advantage in OS, the primary endpoint of the study, for the combination regimen [72].
A phase III randomized trial investigated the combination of ramucirumab plus docetaxel versus placebo plus docetaxel in 530 patients with advanced or metastatic UC progressed during or after platinum-based chemotherapy [73]. The experimental arm was associated with a significantly longer PFS (4.07 months versus 2.76 months in the docetaxel-alone arm) with no OS benefit. Additional follow-up confirmed the advantage in PFS (4.1 months versus 2.8 months in the experimental arm versus the control arm, respectively; HR 0.696; p = 0.0002) and the lack of statistically significant advantage in OS for the combination treatment (9.4 months in the experimental arm versus 7.9 months in the placebo group; stratified HR 0.887; p = 0.25) [74].

Antibody-Drug Conjugates
Another interesting emerging class for the treatment for metastatic UC is antibody-drug conjugate (ADC), that consists in monoclonal antibody against a target expressed on cancer cell bounded to a cytotoxic agent with a protease-cleavable or non-cleavable linker [75]. When the monoclonal antibody binds to a tumor antigen, the drug is internalized and the active chemotherapeutic agent is released into the selected cells, leading to cell death. This mechanism of cell-killing is supposed to limit exposure and toxicity of cytotoxic agents. One of the most promising antibody-drug conjugate currently under investigation in metastatic UC is enfortumab vedotin (ASG-22ME). This ADC is composed of an anti nectin-4 (a cell adhesion molecule highly expressed in UC) monoclonal antibody liked to a micro-tubule-disrupting agent (monomethyl auristatin E). The phase 1 EV-101 trial evaluated enfortumab vedotin in patients with Nectin-4-expressing solid tumors, including 155 heavily pretreated patients with metastatic UC [76]. Single-agent enfortumab vedotin resulted to be well tolerated and showed clinically meaningful and durable responses with an ORR of 43%, a duration of response of 7.4 months, a median OS of 12.3 months, and OS rate at 1 year of 51.8%.
The phase II EV-201 single-arm study investigated enfortumab vedotin in locally advanced or metastatic UC patients previously treated with ICI and platinum-containing chemotherapy (Cohort 1) or an ICI and no prior chemotherapy (Cohort 2) [77]. The preliminary data of cohort 1 enrolling 128 patients have been presented and showed an ORR of 42% with 9% complete responses. The safety profile was manageable with fatigue (50%), alopecia (48%), and decreased appetite (41%) as most common treatment-related adverse events. Of note, one death was reported as treatment related by the investigator (interstitial lung disease). Based on these results, the FDA granted accelerated approval to enfortumab vedotin for patients with locally advanced or metastatic urothelial cancer who have previously received a PD-1/PD-L1 inhibitor and a platinum-containing chemotherapy in the neoadjuvant/adjuvant, locally advanced or metastatic setting. A phase III trial evaluating enfortumab vedotin in patients progressed to previous ICI and platinum containing chemotherapy is ongoing (NCT03474107, EV-301).
Preliminary data for the combination of enfortumab vedotin with pembrolizumab for first line treatment of cisplatinum ineligible patients with metastatic UC are encouraging. The phase Ib study EV-103 (NCT03288545) demonstrated the efficacy of this combination approach in this subset of patients with a tolerable and manageable safety profile [78,79]. At the recent 2020 American Society of Clinical Oncology (ASCO) Genitourinary Cancer Symposium, the updated results were presented by Rosenberg: at a median follow-up of 11.5 months, investigator-assessed objective response rate was confirmed to be 73.3%, with 15.6% complete responses [80]. The most common adverse events were fatigue (58%), alopecia (53%), and neuropathy (53%). A phase III trial with this combination therapy is ongoing (NCT04223856, EV-302).
Another ADC that has been evaluated in metastatic UC is Sacituzumab govitecan, a humanized anti-Trop-2 (an epithelial cell surface antigen overexpressed in UC) monoclonal antibody linked with SN-38 (the active metabolite of irinotecan). Sacituzumab govitecan has been investigated in a phase I/II basket study in 45 patients progressed after at least one prior systemic therapy [81]. The ORR was 31%, including 2 CR and 12 PR. In patients with visceral involvement the ORR was 27% and in patients previously treated with ICIs it was 23%. Median PFS and OS were 7.3 months and 18.9 months, respectively. Among grade ≥3 adverse events there were neutropenia/neutrophil count decreased (38%), anemia (11%), hypophosphatemia (11%), diarrhea (9%), fatigue (9%), and febrile neutropenia (7%). A global, single-arm, phase II trial which is ongoing (TROPHY-U-01, NCT03547973) is evaluating the antitumor activity of Sacituzumab govitecan (10 mg/kg, days 1 and 8 of 21-day cycles) in patients with advanced UC. Cohort 1 [82] assessed the activity in 35 patients progressed to platinum-based regimens and ICIs while cohort 2 [83] enrolled 18 platinum-ineligible patients who progressed after first-line ICI. The interim results of cohort 1 demonstrated an ORR of 29% with 2 confirmed CR, 5 confirmed PR, and 3 unconfirmed PR. The preliminary results of cohort 2 showed an ORR of 28% with 4 confirmed PRs, and 1 PR pending confirmation. The safety profile was consistent with prior reports in both cohorts no treatment-related deaths were reported.
The evolution of practice changing treatments, including promising therapies approved by FDA, for metastatic UC is depicted in Figure 3. Current treatment scenario in metastatic UC is reported in Figure 4.

Combination of ICIs with Cytotoxic Chemotherapy
Several trials exploring the role of ICIs plus chemotherapy in different settings and combinations are currently ongoing. As regards the first-line setting, the randomized phase III KEYNOTE-361 trial (NCT02853305) is investigating the safety and efficacy of front-line pembrolizumab with or without chemotherapy (GC in eligible patients or gemcitabine-carboplatin combination in cisplatin unfit subjects) [84]. OS and PFS are the primary endpoints of this study, with ORR and safety assessed as secondary endpoints. Similarly, the phase III IMvigor130 trial (NCT02807636) has enrolled previously untreated patients affected by advanced or metastatic UC in a 1:1:1 ratio to either atezolizumab plus platinum-gemcitabine, atezolizumab monotherapy, or platinum-gemcitabine plus placebo [85]. The primary endpoints are OS and PFS; secondary endpoints are safety, ORR and DCR. As previously stated, preliminary findings of these two trials showing a close relation between PD-L1 expression, type of treatment and clinical outcomes have had a relevant impact on current indications of pembrolizumab and atezolizumab in UC. More specifically, FDA revised previous indications for the two ICIs, which are now limited for (1) first-line treatment in cisplatin-ineligible patients whose tumors express PD-L1 (CPS ≥ 10% for pembrolizumab and PD-L1 stained tumor-infiltrating immune cells covering ≥ 5% of the tumor area in the case of atezolizumab), (2) subjects which have disease progression during or following platinum-containing chemotherapy, or (3) patients unfit for any platinum-based chemotherapy, regardless of PD-L1 expression.
The anti-PD-L1 agent atezolizumab is being also investigated in a phase II trial (NCT03093922) aimed to evaluate the safety and efficacy of two different dosing schedules of atezolizumab in combination with GC as front-line treatment for advanced or metastatic UC. Regarding less commonly used ICIs, a randomized, placebo-controlled, phase III trial (NCT03967977) has been initiated to investigate the safety and efficacy of front-line tislelizumab plus standard chemotherapy (gemcitabine plus either cisplatin or carboplatin) versus placebo plus standard chemotherapy (gemcitabine plus either cisplatin or carboplatin). Tislelizumab (BGB-A317) is a humanized monoclonal PD-1 antibody which is being evaluating in several solid tumors [86].
With regard to second-line setting, atezolizumab is currently under evaluation in cisplatin-ineligible patients in an ongoing phase II trial (NCT03737123). In this study, subjects who previously received sequential or concurrent ICI and carboplatin-based chemotherapy will be treated with atezolizumab plus docetaxel combination; conversely, patients who have already received an ICI without prior platinum-based chemotherapy will be treated with atezolizumab plus carboplatin-gemcitabine.
Another anti-PD-L1 agent, avelumab, is being investigated in phase II trial on previously untreated, cisplatin-ineligible patients (NCT03390595). In this study, patients are randomized in a 1:1 ratio to receive avelumab in combination with carboplatin-gemcitabine chemotherapy versus carboplatin-gemcitabine alone. Avelumab is also under investigation in a phase II trial comparing avelumab plus GC versus GC in cisplatin fit, treatment naïve patients (NCT03324282). PT-112, a platinum-based agent belonging to the phosphaplatin family, is under evaluation in combination with avelumab in the ongoing phase I/II PAVE-1 trial (NCT03409458).
The combination of the anti-PD-1 antibody pembrolizumab with paclitaxel is currently under investigation in a phase II trial (NCT02581982) on platinum-refractory patients. Lastly, several other combinations of anti-PD-1/PD-L1 agents with cytotoxic agents such as pemetrexed, platinum, and etoposide are being evaluated in a series of ongoing trials (NCT03744793; NCT03582475).
Ongoing phase I/I/II trials, either recruiting or active not recruiting, of ICIs in combinations with cytotoxic chemotherapy in advanced or metastatic UC are summarized in Table 1. In recent years, checkpoint-inhibition combination therapies have provided outstanding efficacy gains in several malignancies including melanoma, lung cancer and renal cell carcinoma [87,88]. The underlying rationale for these combinations lies in the synergistic effect provided by the inhibition of CTLA-4 and PD-1/PD-L1, resulting in an enhance of T-cell function through distinct pathways [89]. The results obtained in a number of cancer types have led to the recent attempt to translate these experiences in advanced or metastatic UC.
The phase 1/2 CheckMate-032 trial investigated ipilimumab plus nivolumab versus nivolumab alone in several malignancies, including platinum-refractory patients affected by advanced or metastatic UC [90,91]. In this cohort of subjects, the combination of the two immunotherapies yielded a promising response rate of 38%; moreover, subjects treated with the combination showed a median OS of 15.3 months versus 9.9 months in the nivolumab arm. The combination of an anti-PD-1 and a CTLA-4 antibody is being investigated also in the CheckMate-901 trial (NCT03036098) [92], aimed to evaluate the efficacy of nivolumab ± ipilimumab versus GC or carboplatin-gemcitabine chemotherapy. The same combination with modified schedules and additional nivolumab/ipilimumab "boost" cycles is under evaluation also in a Phase II trial (NCT03219775, TITAN-TCC) on treatment naïve and platinum-refractory patients with advanced or metastatic UC.
The anti-PD-L1 agent durvalumab, registered by FDA as monotherapy in previously treated patients affected by advanced or metastatic UC, is currently under investigation in combination with the CTLA-4 IgG2-kappa monoclonal antibody tremelimumab in the DANUBE (NCT02516241) and the NILE (NCT03682068) trials [93,94]. The DANUBE is an ongoing randomized, open-label, phase III trial aimed at ascertaining the value of front-line durvalumab ± tremelimumab versus platinum-gemcitabine chemotherapy in advanced or metastatic UC [89]. In the same setting of previously untreated patients, the NILE trial randomized subjects to three different cohorts: durvalumab plus tremelimumab plus platinum-gemcitabine; durvalumab plus platinum-gemcitabine; platinum-gemcitabine [94].
Ongoing phase II/III trials, either recruiting or active not recruiting, of ICIs in combinations with other ICIs in advanced or metastatic UC are summarized in Table 2.

Combination of ICIs with Antiangiogenic Agents
Given the importance of angiogenesis as a crucial process in the carcinogenesis and progression of UC, ICIs are under evaluation also in combination with VEGFR antibodies and TKIs, including bevacizumab, ramucirumab, lenvatinib, and several others [61,95].
As regards front-line treatment, the VEGF-A monoclonal antibody bevacizumab is being investigated in a phase II trial assessing bevacizumab plus atezolizumab in treatment-naïve, cisplatin-ineligible patients (NCT03272217).
Axitinib, a highly selective VEGFR-1, -2, and -3 inhibitor, is currently under investigation as front-line treatment in combination with avelumab in the ongoing phase II trial JAVELIN Medley VEGF (NCT03472560). Enrolled subjects are deemed ineligible for receiving cisplatin-containing first-line chemotherapy and the primary endpoint is ORR, defined as a confirmed CR or PR.
Ramucirumab is an IgG1 monoclonal antibody that binds VEGFR-2 preventing ligand binding and receptor-mediated pathway activation in endothelial cells [96]. An ongoing, phase I trial is assessing the safety of ramucirumab in combination with pembrolizumab in previously treated patients affected by a number of solid cancers, including UC (NCT02443324).
Lenvatinib is a small TKI able to inhibit VEGFR-1, FGFR1-4, stem cell factor receptor (KIT), platelet-derived growth factor receptor α (PDGFRα), and rearranged during transfection (RET) [97]. The combination of lenvatinib and pembrolizumab is being investigated as front-line treatment in the phase III LEAP-011 trial (NCT03898180) which is evaluating the combination in cisplatin-unfit subjects with PD-L1 CPS ≥10 or in patients deemed ineligible for any platinum-based regimen, regardless of PD-L1 expression.
Cabozantinib is another small TKI inhibiting a plethora of targets which play an important role in tumor growth, angiogenesis, and survival, such as VEGFR-2, MET, RET, KIT, AXL, and FLT3 [98,99]. Following the findings of a recent phase I trial where cabozantinib plus nivolumab plus ipilimumab yielded an ORR of 36% across all genitourinary cancers [100], this molecule is being evaluated in combination with pembrolizumab (NCT03534804), durvalumab (NCT03824691), atezolizumab (NCT03170960), and nivolumab plus ipilimumab (NCT03866382) in treatment-naïve and previously treated patients. Moreover, cabozantinib is also under investigation in a phase II trial (NCT04066595) which is enrolling previously treated subjects with platinum-based chemotherapy (cohort 1) and platinum-based chemotherapy plus ICIs (cohort 2).
The anti-VEGF recombinant EphB4-HSA fusion protein is currently under evaluation in combination with pembrolizumab in an ongoing phase II trial (NCT02717156). The study is enrolling treatment naïve patients affected by locally advanced or metastatic UC.
Apatinib, a small-molecule TKI which selectively inhibits VEGFR-2 resulting in a decrease in endothelial proliferation, migration, and tumor microvascular density, is under evaluation in combination with pembrolizumab in a phase I/IIa trial (NCT03407976; APPEASE). In this study, eligible subjects must have progressed during or following platinum-based chemotherapy.
Lastly, sitravatinib, a small TKI able to inhibit VEGFR, PDGFR, KIT, RET, and MET [101], is currently under investigation in combination with nivolumab in a non-randomized, Phase II trial (NCT03606174). Although all patients are planned to receive the same treatment (nivolumab 240 mg every 2 weeks or 480 mg every 4 weeks plus sitravatinib 120 mg orally once per day continuously in 28-day cycles), eligible subjects are assigned to eight different cohorts, based upon previous therapies for UC.
Ongoing phase I/II/III trials, either recruiting or active not recruiting, of ICIs in combinations with antiangiogenic agents in advanced or metastatic UC are summarized in Table 3.  Although combination therapies are displaying the ability to broaden the anticancer activity of ICIs and the majority of ongoing trials are testing ICIs in combination with other anticancer agents, some trials are evaluating the role of monotherapy in different settings.
The anti-PD-1 agent pembrolizumab is being evaluated in a randomized, double-blinded phase II trial (NCT02500121) assessing the role of maintenance pembrolizumab (200 mg flat dose every three weeks, for up to 24 months) versus placebo after front-line chemotherapy in patients affected by metastatic UC. Eligible subjects must have achieved CR, PR or stable disease (SD) after 4 to 6 cycles of first-line platinum-based chemotherapy; six-month PFS assessment, regardless of PD-L1 expression, is the primary outcome. Maintenance treatment with ICIs is also under investigation in an ongoing phase III trial (NCT02603432) comparing avelumab maintenance plus best supportive care versus best supportive care alone in patients whose disease did not progress after first-line platinum-based chemotherapy.
Atezolizumab treatment is being tested in the real-world phase III SAUL trial (NCT02928406) and preliminary results of this study assessing the role of atezolizumab in a pretreated population of 1004 UCs have been recently published [102]. Median OS and PFS were 8.7 and 2.2 months respectively, with an ORR of 13%. The trial enrolled patients who experienced progression during or after one to three prior therapies, of which 10% had ECOG-PS 2 and 98% were platinum pretreated.
Toripalimab (JS001), a recombinant, humanized PD-1 monoclonal antibody capable of preventing the binding of PD-1 with PD-L1 and PD-L2, is being evaluated as monotherapy in pretreated advanced or metastatic UC in an ongoing phase II trial (NCT03113266). The primary outcome is ORR, with duration of response, PFS, OS, and safety as secondary outcomes.
The anti-CTLA-4 agent tremelimumab is currently being evaluated as monotherapy in a phase II trial (NCT03557918) assessing ORR in patients with metastatic UC which previously received PD-1/PD-L1 blockade.
The novel anti-PD-L1 CK-301 (Cosibelimab) is being tested in a phase I trial (NCT03212404) on a number of advanced malignancies, including UC. Lastly, a phase I trial (NCT03053466) is studying the role of the anti-PD-1 agent APL-501 in patients affected by advanced solid tumors presenting at least 1% of PD-L1 expression by IHC.
Ongoing phase I/II/III trials, either recruiting or active not recruiting, of ICI monotherapy in advanced or metastatic UC are summarized in Table 4.

Novel Immunotherapy Approaches
With the aim to enhance the response to ICIs and other anticancer agents, a number of novel immunomodulatory molecules and brand-new combinations are being evaluated in UC [103,104].
A recently emerging immunotherapeutic target is represented by the indoleamine 2,3-dioxygenase-1 (IDO1), an enzyme playing a crucial role in immunosuppression, angiogenesis, and metastasis [105]; in fact, IDO1 is an immune regulatory enzyme which promotes tryptophan depletion, a mechanism necessary for T-cell survival [106]. More specifically, IDO1 enhances the activity of CD4+ T regulatory cells and myeloid-derived suppressor cells and, conversely, is able to suppress CD8+ T effector and natural killer (NK) cells [107]. Despite early promising results, the combination of pembrolizumab plus the IDO-1 inhibitor epacadostat came up short against its primary endpoints of OS and PFS; thus, the two trials assessing the role of the anti-IDO-1 ± pembrolizumab in treatment-naïve, cisplatin ineligible subjects (NCT03361865) and in platinum-refractory patients (NCT03374488) arrested recruitment. Currently, the safety of the combination of pembrolizumab plus KHK2455, a long-active selective IDO-1 inhibitor, is being evaluated in an ongoing Phase I study on platinum-refractory patients affected by metastatic UC (NCT03915405).
Another attracting target is represented by the tumor necrosis factor receptor OX40 (CD134) [108,109]; when activated by its ligand OX40L, OX40 is involved in T-cell signaling activation, promoting T-cell survival and enhancing the expression of several molecules such as Bcl-2 anti-apoptotic molecules, cytokines, cyclin A, and cytokine receptor [110]. Therefore, as OX40 may promote proliferation and survival of CD4+ and CD8+ T cells, immunostimulatory agonistic agents are currently under investigation in several solid malignancies [111]. The OX40 agonist PF-04518600 is being evaluated as monotherapy or in combination with the cytokine modulator utomilumab (PF-05082566)-a monoclonal antibody with agonist activity toward 4-1BB (CD137), a receptor expressed on NK, CD8+, and CD4+ T cells [112]. Preliminary results of this trial, which includes also a cohort of patients affected by UC, have shown an ORR of 5.4% across all cancer types; nevertheless, the promising 50% of ORR reported in the UC subgroup has led to the NCT03217747 and the Javelin Medley (NCT02554812) ongoing phase I/II trials which are evaluating the OX40 agonist PF-04518600 in combination with ICIs, radiation therapy, utomilumab, and cytotoxic chemotherapy. Finally, the hexavalent OX40 agonist INBRX-106 is currently under investigation as monotherapy or in combination with pembrolizumab for previously treated patients in a phase I trial (NCT04198766).
Another potential target is lymphocyte activation gene-3 (LAG-3, CD223), a co-inhibitory receptor able to suppress T-cell activation and cytokines secretion [115]; more specifically, LAG-3 overexpression in tumor cells is involved in the phenomenon of immune exhaustion, with suppression of T-cell function [116,117]. Thus, LAG-3 inhibitors as monotherapy or in combination with anti-PD-1 agents are currently being explored in several phase I and II trials in advanced malignancies, including pretreated UC (NCT03538028, NCT03250832). [118]. TIM-3 expression has been recently associated with poor prognosis in a number of cancer types, including UC [119][120][121]; because of the implication of TIM-3 overexpression in T-cell dysfunction and exhaustion, several TIM-3 inhibitors are currently being studied in advanced cancer. Among them, INCAGN02390 is under evaluation in a phase I trial (NCT03652077) assessing its role as monotherapy in previously treated metastatic malignancies including UC.

T-cell immunoglobulin and mucin-domain containing-3 (TIM-3) is another co-inhibitory receptor expressed on regulatory T cells, effective T cells, tumor cells, and innate immune cells (macrophage and dendritic cells)
Glucocorticoid-induced TNF receptor family related receptor (GITR) represents a co-stimulatory receptor that binds the GITR ligand (GITRL) [122]; the activation of GITR can result in signals influencing the activity of CD4+, CD8+ and regulatory T cells, playing an important role in autoimmune and inflammatory diseases as well as in anticancer immune response [123,124]. Thus, GITR seems to be a promising target for novel immunotherapy agents. A phase I/II trial analyzing the combination of nivolumab, ipilimumab, and the GITR agonist INCAGN01876 (NCT03126110) in patients with metastatic malignancies including UC is recruiting at present.
Chimeric antigen receptor (CAR)-T immunotherapy has shown impressive responses in a number of B cell malignancies and is currently being tested in several solid tumors, including advanced or metastatic UC (NCT03185468) [125]. CAR-T action is based on engineered T cells expressing a CAR; current second-generation CAR are receptors composed of (1) an extracellular, epitope-specific binding domain, (2) a transmembrane domain, (3) and an intracellular domain of the T cell receptor; this last domain consists in its turn of costimulatory molecules such as CD28, 4-1BB, and the CD3ζ chain, and is involved in a massive activation of T cells which is independent from T-cell receptor (TCR)-major histocompatibility complex (MHC) interactions [126,127].
Lastly, another promising immunotherapeutic strategy lies in tumor vaccines (TVs), which are currently under investigation in many solid tumors [128,129]. As regards UC, the majority of developing TVs concerns BCG-relapsing, non-muscle invasive disease, where neo-antigens are being studied in combination with immune-stimulating adjuvant agents, cytotoxic agents, and/or mTOR inhibitors (NCT01353222, NCT02015104, NCT01498172). Cancer vaccines are also under investigation in combination with ICIs, as in the case of the NCT03689192 and the NCT03639714 trial. In the NCT03639714 Phase 1/2 Study, two vaccines vectors (GRT-C901 and GRT-R902)-used as immune boosts-are being investigated in combination with nivolumab plus ipilimumab in patients affected by a number of solid cancers including previously treated, metastatic UC.
Ongoing phase I/II/III trials, either recruiting or active not recruiting, of novel immunotherapy approaches in advanced or metastatic UC are summarized in Table 5.

PARP Inhibitors
One of the new promising therapeutic approaches is the use of Poly(ADP-ribose) polymerase (PARP) inhibitors that target DNA repair gene mutations and have been proven active in other type of cancer like ovarian, breast, and prostate cancer [130,131].
Regarding UC, genomic alterations in DNA repair genes like ATM, ERCC2, RAD51B were found in 2-14% and in BRCA 1/2, PALB2, FANCD2, ERCC2, ATM in 3.7-12.3% of MIBC [46,132]. Moreover, patients with DNA damage response and repair (DDR) gene alterations treated with platinum based chemotherapy resulted to have better PFS and OS [133]. In fact, in multiple tumors the presence of DDR gene aberrations correlates with an enhanced sensibility to platinum compounds [134]. Based on these results, PARP inhibitors have been studied in UC as well [135][136][137][138]. At the recent ASCO Genitourinary Cancers Symposium 2020, the results of the study ATLAS (NCT03397394) were presented [134]. This phase II trial assessed the efficacy and safety of the PARP inhibitor rucaparib in 97 patients with locally advanced or metastatic UC with or without homologous recombination deficiency (HRD), progressed to one or two prior treatments. Total of 20.6% of patients were HRD-positive, 30.9% were HRD-negative, and 48.5% had unknown HRD status. Among patients with sequencing results (64 patients), deleterious alterations in BRCA1, BRCA2, RAD51C, PALB2 were infrequent (9.4%). Common alterations were found in TP53 (52.4%) and in FGF/FGFR pathway (77.6%). The results showed that there were no confirmed responses to rucaparib, 28.1% of patients achieved a stable disease as best response with no difference in efficacy between HRD-negative and HRD-positive patients. The trial was discontinued because protocol-defined continuance criteria were not meet. Two phase II trials are investigating the PARP inhibitor olaparib in monotherapy in chemotherapy naïve cisplatin ineligible patients or progressed to first line treatment selected for DDR mutations (NCT03448718) and in patients with DNA-repair defects progressed to 1 or 2 prior treatment regimens (NCT03375307).
A phase II trial is currently investigating the PARP inhibitor niraparib as maintenance therapy until disease progression or unacceptable toxicity or death in patients unselected for DDR mutational status not progressing to first line platinum-based chemotherapy (NCT03945084).
Another strategy being tested is combination therapy of PARP inhibitors with ICIs or target therapies.
Indeed, the presence of alteration in DDR genes has been associated with higher mutational load and higher response to ICIs in patients with UC [139,140]. Based on these observations, several combinations of PARP inhibitors and PD-1/PD-L1 inhibitors are currently being tested: durvalumab plus olaparib (module B, NCT02546661, active not recruiting; NCT03459846, active not recruiting), rucaparib plus nivolumab (NCT03824704, active not recruiting), niraparib plus atezolizumab (NCT03869190, recruiting).
Ongoing phase I/II/III trials, either recruiting or active not recruiting, of PARP inhibitors in advanced or metastatic UC are summarized in Table 6.

Target Therapy
As already discussed, the FGFR inhibitor erdafitinib is a promising treatment strategy in patients with FGF/FGFR alterations. In these subgroup of patients, other therapies directed at inhibiting FGFR are currently being tested: PRN1371, a FGFR 1-4 inhibitor, in a phase I trial in previously treated patients (NCT02608125); Pemigatinib, a FGFR1-3 inhibitor, in phase II trial in patients progressed to at least one prior treatment (NCT02872714, FIGHT-201); Rogaratinib (BAY1163877), a FGFR 1-4 inhibitor, in a phase II/III trial in patients progressed to at least one platinum-containing regimen (NCT03410693).
Moreover, FGFR inhibitors are being evaluated in combinations with PD-1/PD-L1 inhibitors. A study by Sweis et al. showed that FGFR3 pathways were activated in non-T-cell-inflamed UC, characterized by an absence of intratumoral T cells, thus identifying a potential targetable pathway that could help to overcome tumor-intrinsic immunotherapy resistance [57]. The updated results of the interim analysis of the phase II study FIERCE-22 (NCT03123055) evaluating the combination of the FGFR3 inhibitor vofatamab (a human IgG1 monoclonal antibody directed against FGFR3) in combination with pembrolizumab in 28 patients (20 wild-type) progressed following platinum-based chemotherapy have been presented at European Society for Medical Oncology (ESMO) Congress 2019: the combination therapy resulted to be well tolerated with encouraging ORR (29.6%) and a median PFS is 4.7 months [142].
Another interesting pathway being investigated is targeting human epidermal growth factor receptor 2 (HER2, ERBB2) considering that mutation or amplification of ERBB2 gene has been identified in 9% of MIBC [45]. Trastuzumab deruxtecan is an ADC composed of trastuzumab, a monoclonal antibody targeting HER2 conjugated to deruxtecan, a derivative of the camptothecin analog exatecan, a DNA topoisomerase 1 inhibitor. This compound is being tested in combination with nivolumab in a phase I/II trial in patients with HER2 expression of IHC 1+, 2+ or 3+, progressed to prior platinum-based therapy (NCT03523572). RC48-ADC, an anti-HER2 monoclonal antibody, in under evaluation in two phase II trial in previously treated patients, one in HER2 negative (IHC 0 or 1+, NCT04073602) and one in HER2 overexpressed tumors (IHC 2+ or 3+, NCT03809013). PRS-343, a bivalent, bispecific fusion protein composed of an anti-HER2 monoclonal antibody linked to a CD137-targeting anticalin, is being investigated in HER2 positive solid tumor malignancy, including UC, for which standard therapies are not available.
Ongoing phase I/II/III trials, either recruiting or active not recruiting, of target therapies in advanced or metastatic UC are summarized in Table 7, while miscellanea therapies are reported in Table 8.

Conclusions
In the recent years, the treatment scenario of metastatic UC has been enriched with several new therapeutic options. Immunotherapy is a very promising approach for this disease, but a high percentage of patients are still resistant to this type of treatment. In the future years, the results of the ongoing trials investigating ICIs in combination with target therapy or chemotherapy will assess if resistance to ICIs alone can be overcome. Promising treatment approaches are FGFR inhibitors and enfortumab vedotin. These two treatment strategies already showed good results in monotherapy and combination therapies with ICIs being tested. Other compounds, such as PARP inhibitors, mTOR inhibitors, anti-VEGF, tyrosine kinase inhibitors, HER2 targeting therapies, either alone or in several types of combinations, are being investigated in clinical trials.
The therapeutic approach to UC, which for many years has been dominated by platinum containing chemotherapy based on clinical and laboratory variable defining cisplatin eligibility, is now shifting toward a more personalized approach, based on the presence of molecular alteration (e.g., FGFR alterations) or PD-L1 expression.