Until recently, therapeutic options for the management of advanced urothelial cancers have been limited to traditional cytotoxic agents with modest efficacy. However, the brisk pace of recent discovery in the field of immunotherapy and molecular therapeutics has ushered in a new era for management of patients with urothelial carcinoma.
Earle F. Burgess, MD
Until recently, therapeutic options for the management of advanced urothelial cancers (UCs) have been limited to traditional cytotoxic agents with modest efficacy. However, the brisk pace of recent discovery in the field of immunotherapy and molecular therapeutics has ushered in a new era for management of patients with urothelial carcinoma. New agents demonstrate improved patient outcomes and tolerability as monotherapy, and they provide hope that treatment paradigms for this disease will continue to evolve as the prospect of novel combinations are tested. This review will discuss the recent development of immune checkpoint inhibitors in advanced UC.
The use of platinum-based multiagent chemotherapy has been the preferred choice for the management of treatment-naïve patients with advanced UCs for more than 2 decades. The combination of gemcitabine and cisplatin yields an anticipated overall survival (OS) of 14 months, a 5-year OS rate of 13%, and a progression-free survival (PFS) duration of approximately 8 months within a broad patient population.1Toxicities associated with the use of chemotherapy are well recognized and tend to be transient, although long-term effects such as renal dysfunction, neuropathy, hearing loss, or tinnitus can occur, especially in patients treated with cisplatin-containing regimens.
Patients are considered to be unfit for cisplatin use if any of the following criteria are present: ECOG performance status of 2 or greater; creatinine clearance of less than 60 mL/min; Common Terminology Criteria for Adverse Events version 4 grade 2 or higher; hearing loss; peripheral neuropathy; or New York Heart Association class III heart failure.2While alternative cisplatin dosing regimens have been tested in small studies,3cisplatin-ineligible patients are best managed by substituting carboplatin in multiagent regimens. In a randomized phase II/III study, the use of gemcitabine and carboplatin yielded an OS of 9.3 months and PFS of 5.8 months in patients unfit for cisplatin.4The inferior outcomes of the gemcitabine/carboplatin arm observed in this study, compared with the historical experience with gemcitabine and cisplatin, are most likely due to the poorer prognosis of patients deemed unfit for cisplatin and, less likely due to differences in efficacy between agents, although cisplatin use is preferred for patients when feasible.
After failure of first-line chemotherapy, use of alternate cytotoxic agents is associated with a short PFS duration and limited response rate.5,6Patient outcomes are modestly improved with the addition of anti-VEGFR2 targeted therapy7; however, cumulative disease and treatment-related (TR) adverse events (AEs) often limit therapeutic options in the post first-line setting. The limited efficacy and tolerability of traditional cytotoxic chemotherapy agents after failure of cisplatin-containing regimens or in cisplatin-ineligible patients present a large unmet medical need that can now be addressed by the availability of immunotherapeutic agents in this patient population.
Immune Checkpoint inhibition
Immune checkpoints refer to an increasing number of known molecular immune inhibitory pathways that are crucial for maintaining immune self-tolerance, although they also serve as fundamental mechanisms by which neoplastic cells evade immune surveillance.8To date, pharmacologic efforts to modulate immune checkpoints in UC have been focused on the development of humanized monoclonal antibodies against PD-1, PD-L1, and CTLA-4, as summarized below.
Atezolizumab is a humanized monoclonal antibody against PD-L1 that became the first agent approved by the FDA for advanced UC in more than 30 years when it was granted accelerated approval in May 2016 on the basis of the IMvigor 210 study results. This single-arm phase II trial included 2 patient cohorts who received atezolizumab 1200 mg intravenously (IV) every 3 weeks until progression or intolerance. The initial regulatory approval is based on the second cohort that was studied, which included patients with unresectable or metastatic disease who had failed prior platinum-based chemotherapy or who experienced disease recurrence within 12 months of perioperative platinum-based chemotherapy. The intent-to-treat (ITT) population demonstrated an objective response rate (ORR) of 15% regardless of PD-L1 status.9In this trial, PD-L1 protein expression was assessed by immunohistochemistry (IHC) using the SP142 antibody. The percentage of PD-L1 expressing tumor-infilitrating immune cells (ICs) was classified as high (≥ 5%; IC2/3), low (≥1% but <5%; IC1) or null (<1%; IC0). High PD-L1 expression on tumor-infiltrating ICs enriched but was not necessary for response. In contrast to traditional cytotoxic agents in this setting, responses appear durable, as the median duration of response (DOR) was not reached after a median follow-up of 11.7 months. Seventeen percent of patients subsequently responded beyond a first progression event, suggesting that continued treatment beyond progression is appropriate in carefully selected patients. Fatigue is fairly common, although the observed incidence of serious immune-related AEs (irAEs) was 5%.
Results of the confirmatory phase III IMvigor 211 trial were recently reported and failed to meet the study’s primary endpoint of improved OS relative to chemotherapy. The study was designed to test OS in a hierarchical fashion beginning with patients with high PD-L1expressing tumors (IC2/3) before inclusion of the entire ITT population.10Atezolizumab treatment resulted in similar OS in the IC2/3 population (11.1 months for atezolizumab vs 10.6 months for chemotherapy; P = .41) and showed modest OS improvement in the ITT population (8.6 months for atezolizumab vs 8.0 months for chemotherapy; P = .038). A subset of patients treated with atezolizumab experienced durable responses, and the safety profile was favorable compared with chemotherapy.
Cohort 1 in the IMvigor 210 trial included treatment-naïve patients with advanced or metastatic disease who were ineligible for cisplatin-containing chemotherapy.11In this cohort, 23% of the ITT population demonstrated an objective response. Response rates were again enriched by PD-L1 expression (28% vs 21% in high vs low PD-L1expressing groups, respectively) and appear durable, as the median DOR was not reached after 17.2 months of follow-up. The FDA announced on June 20 that the use of atezolizumab in cisplatin-ineligible patients should depend upon PD-L1 expression level based upon observed results in ongoing clinical trials.
Both durvalumab and avelumab are humanized antiPD-L1 antibodies that received accelerated approval by the FDA in May 2017 for patients with advanced UC who have failed first-line treatment or recurred within 12 months of perioperative platinum-based chemotherapy, similar to the conditions of atezolizumab’s approval. Durvalumab (10 mg/kg IV every 2 weeks) was evaluated in a single-arm phase I/II study (NCT01693562). The most recent data cutoff reported analysis of 191 patients, showing an ORR of 17.8% within the entire patient cohort.12In this study, PD-L1 expression was assessed using the SP263 IHC assay. PD-L1 expression was considered positive if ≥25% of tumor cells or infiltrating ICs demonstrated expression. More than 50% of patients’ tumors demonstrated a high PD-L1 expression score as defined by the assay used in the study, which also predicted for an increased likelihood of response.
ORRs for patients with high versus low PD-L1 expression were 27.6% versus 5.1%, respectively. Median DOR has not been met for the entire cohort; however, the median follow-up time was less than 6 months as of the reported data cutoff. Grade 3/4 treatment-related adverse events (TRAEs) have occurred in 6.8% of patients, including 2.1% immune-related adverse events (irAEs).
Avelumab (10 mg/kg IV every 2 weeks) was evaluated in a single-arm study (NCT01772004). The majority of UC patients had progressed on prior platinum-based chemotherapy. The most recent data cutoff included 249 patients, including 23% with upper urinary tract primary tumor sites. In the 161 patients with at least 6 months of follow-up, the confirmed ORR was 17.4% for the entire cohort and 25.4% versus 13.2%, respectively, in patients with PD-L1high and –low tumor scores.13The 24-week durable response rate was 92.3%, confirming response durability as seen with other agents. Almost 23% of patients experienced grade 1 or 2 infusion reactions, and thus premedication with an antihistamine and acetaminophen is indicated. Grade ≥3 TRAEs occurred in 8.4% of patients, including irAEs in 2.4% and 1 treatment-related (TR) death due to pneumonitis.
Therapeutic disruption of the PD-1/PD-L1 signaling axis by targeting the PD-1 receptor has also been effective in UC. Nivolumab is an antibody that is directed against PD-1 that received accelerated approval for the same indication in patients with advanced UC as the above-mentioned agents in the postplatinum setting in February 2017 based on results of a single-arm phase II trial including 270 patients, in which previously treated patients with advanced UC received nivolumab (3 mg/kg IV every 2 weeks) until progression or intolerance. Results were similar to the experience with the anti PD-L1 antibodies, as the ORR for the entire cohort was 19.6%.14Response rates for patients with the highest and lowest PD-L1 expression scores were 28.4% and 16.1%, respectively. Toxicity patterns were similar to those of other agents, including 1.1% TR mortality.
Remarkably, pembrolizumab, a humanized anti PD-1 antibody, is the fifth immune checkpoint inhibitor to show activity in patients with advanced UC. The FDA granted full approval for pembrolizumab in the postplatinum setting and accelerated approval for cisplatin-ineligible patients in the treatment-naïve setting in May 2017. Pembrolizumab is the first agent to confirm an OS benefit relative to chemotherapy after failure of platinum-based chemotherapy. In the KEYNOTE-045 study, 542 patients were randomized to receive pembrolizumab (200 mg IV every 3 weeks) or investigator’s choice chemotherapy (paclitaxel, docetaxel, or vin-flunine).15The median OS was 10.3 months in the pembrolizumab group versus 7.4 months in the chemotherapy group (HR for death, 0.73; P = .002). Interestingly, pembrolizumab was associated with a benefit relative to chemotherapy regardless of PD-L1 expression score, as the hazard ratios for death in patients with ≥10% versus <10% PD-L1 expression by combined positive score were 0.57 (0.37-0.88) and 0.80 (0.61-1.05), respectively.
Accelerated approval of pembrolizumab in the first-line setting for cisplatin-ineligible patients is based on KEYNOTE-052, on which 370 patients received pembrolizumab until progression or intolerance.16With a median follow-up of 8 months, ORR was 29%.
Role of PD-L1 Testing and Clinical Decision Making With Advanced UC
All of the active immune checkpoint inhibitors discussed are thought to act by restoring tumor-infiltrating lymphocyte (TIL) effector function by preventing the interaction of PD-L1 with PD-1 on TILs. The methodology of PD-L1 measurement (antibody and cellular compartment) and expression thresholds showed differences across the above-mentioned studies. These differences could be explained by the dynamic nature of PD-L1 expression,17discordance between primary and metastatic lesions,18and differences in PD-L1 isotype expression,19although further investigation is warranted to optimize the predictive value of this biomarker. However, despite the differences observed with PD-L1 expression patterns in these studies, patients with high PD-L1 expression scores as defined in the respective trials consistently demonstrated higher therapeutic responses relative to patients with low PD-L1 expression scores. Nonetheless, patients with low expression scores also consistently demonstrated some responsiveness to the drugs. In May 2018, the US FDA issued an alert that patients with low tumor PD-L1 expression receiving pembrolizumab or atezolizumab monotherapy in the treatment-naïve setting were found to have inferior survival compared to patients treated with chemotherapy based on interim analysis of ongoing clinical trials.
For this reason, frontline use of these agents in cisplatin-ineligible patients should be limited to those with a high tumor PD-L1 score (Pembrolizumab: Combined Positive Score ≥10; Atezolizumab: tumor infiltrating immune cell ≥5%). In other approved clinical settings, routine testing for PD-L1 expression has not been confirmed to enhance decision making and is not typically performed by the authors.
Following publication of the KEYNOTE-045 study, pembrolizumab is the only agent in the class to confirm an OS advantage to date, relative to chemotherapy, after failure of platinum-based chemotherapy; therefore, it is currently the authors’ preferred agent. The toxicity profile appears similar across the class, so if a survival advantage is identified with other agents in the ongoing confirmatory phase III trials, the choice of which drug to select as monotherapy will likely be based on the convenience of infusion schedules and treatment cost in the absence of direct comparison of these agents in future studies. Table 1 presents a summary of the current approved agents in advanced UC.
Future Directions of Immune Checkpoint Inhibitor Use in UC
Results from monotherapy trials in advanced UC have generated much excitement for expanding the role of immunotherapy in this disease. Many questions remain on how best to optimize patient selection and sequencing of immunotherapy with other agents, which is being studied in multiple ongoing clinical trials. Synergy may also exist when combining immune checkpoint inhibitors with chemotherapy, presumably through enhanced presentation of tumor-derived antigens following chemotherapy-induced tumor cell death. For example, in patients with treatment-naïve advanced nonsmall-cell lung cancer, the combination of cytotoxic chemotherapy and immune checkpoint inhibition with pembrolizumab results in a higher response rate relative to chemotherapy alone.20This hypothesis is being tested in patients with advanced UC (NCT02853305) and may reposition the use of immunotherapy in combination with cytotoxic chemotherapy as a standard first-line strategy if synergy is validated in this patient population. Table 2 presents a list of selected ongoing trials studying combinations of chemotherapy and immune checkpoint inhibitors.
In addition to testing the role of immune checkpoint blockade in earlier stages of disease, perhaps the most exciting prospect on the horizon is the potential for improved patient outcomes from targeting the PD-1/PD-L1 axis in combination with other immune checkpoint pathways. In patients with malignant melanoma, combined therapeutic targeting of PD-1 and CTLA-4 with nivolumab and ipilimumab is superior to immune checkpoint monotherapy.21,22This combination has been studied in an open label phase I/II study assessing the combination of nivolumab and ipilimumab versus nivolumab monotherapy in patients with advanced, platinum-refractory UC. In the treatment arm receiving nivolumab (1 mg/kg) and ipilimumab (3 mg/kg), the ORR was 38.5%23compared with 24.4% in the nivolumab monotherapy arm.24Serious AEs occurred in more than 30% of patients in the combination arms, underscoring that any future confirmed survival benefit resulting from combination immunotherapy will need to be balanced against potential immune-related toxicities. Durvalumab is also being studied as monotherapy and in combination with the antiCTLA-4 antibody tremelimumab versus chemotherapy in the first-line setting (NCT02516241). Results from this international phase III trial are anticipated and will likely determine the future role of combined targeting of the PD-L1 and CTLA-4 axes in advanced UC.
Multiple novel immune checkpoint pathways that may also be amenable to therapeutic targeting have been identified. For example, indoleamine 2,3-dioxygenase 1 (IDO1) is a tryptophan-catabolizing enzyme that impairs T-cell immune surveillance,25and tumor cell overexpression of IDO1 has been associated with adverse patient outcomes.26Epacadostat is an oral, selective IDO1 inhibitor that has been studied in combination with pembrolizumab in the phase I/ II setting. Preliminary results from 40 patients with advanced UC in the phase II portion demonstrated an ORR of 35% in all patients and 57% in the subgroup with 0-1 prior lines of therapy for advanced disease. Fatigue and rash are the most common TRAEs thus far, as the combination appears to be well tolerated.27Because these early results compare favorably with pembrolizumab monotherapy observed in the KEYNOTE-045 study, results from randomized studies with this combination are anticipated.
Recent advancements in the field of tumor immunology have led to the development of multiple novel therapeutic strategies in patients with advanced UCs, which represents the dawn of a new era in the management of patients with this disease. Although many patients fail to benefit from current immune checkpoint inhibitors, the pace of discovery is brisk and great optimism exists that we are in just the first stages of the immunotherapy revolution. Much work remains to be done in the years ahead to optimize the use of immunotherapy in patients with advanced UC, and current and planned clinical trials let us hope for and anticipate further progress with improved patient outcomes.