Read the Full Video Transcript

Christopher Wallis: Hello, and thank you for joining us for this UroToday Journal Club. Today, we are discussing the recently published KEYNOTE-057 trial, pembrolizumab monotherapy for the treatment of high-risk non-muscle-invasive bladder cancer unresponsive to BCG: an open-label, single-arm, multicenter, phase 2 study. I'm Chris Wallis, a Fellow in Urologic Oncology at Vanderbilt, and with me today is Zach Klaassen, an Assistant Professor in The Division of Urology at the Medical College of Georgia. This is the citation for the long-awaited KEYNOTE-057 study published recently in The Lancet Oncology.

By way of background, non-muscle-invasive bladder cancer varies in terms of both recurrence risk, and progression rates as a result of disease state and underlying histologic characteristics. Patients with high-risk disease have relatively high rates of both recurrence as well as progression, and on that basis, BCG is the recommended adjuvant treatment. However, patients with CIS may be particularly intrinsically resistant to BCG, and as a result, while the vast majority of patients have an initial complete response to BCG, a failure to maintain this response is relatively common and a large proportion of patients develop BCG-refractory or BCG-unresponsive disease.

And so work from the International Bladder Cancer Group set up benchmarks for us to consider when we're looking at novel treatment approaches in the treatment of patients with BCG-unresponsive CIS. They defined an initial complete response rate of 50% at 6 months with a durable response rate of 30% at 12 months, and 25% at 18 months as being clinically meaningful. And so, this forms the basis of how we assess the various treatment options in this disease space. A radical cystectomy remains the most definitive treatment approach. However, it is associated with significant perioperative morbidity and not insignificant perioperative mortality. As a result, many patients are unwilling or unable to undergo cystectomy, and so a variety of other treatment approaches have been utilized. This highlights a systematic review, looking at a whole variety of intravesical treatment options, including BCG with interferon, mitomycin-C, valrubicin, and others.

More recently, we've seen the approval of nadofaragene firadenovec, which is an intravesical adenovirus, which showed relatively promising results with a 3-month complete response rate, 53% in the CIS cohort, and 73% in the papillary disease cohort.

And so, when we look at more advanced stages of bladder cancer, we can see that immunotherapy forms a key component of our treatment regime in both maintenance therapy for cisplatin-eligible patients and upfront therapy for those who are cisplatin-ineligible in the first-line space, and particularly in the second-line space, pembrolizumab is the preferred treatment approach. And so, in the KEYNOTE-057 study, the authors sought to assess whether pembrolizumab could induce the clinical complete response in patients with BCG-unresponsive high-risk non-muscle-invasive bladder cancer.

To do this, they performed an open-label, single-arm, multicenter phase 2 study. They enrolled patients who are 18 years or older with histologically confirmed BCG-unresponsive non-muscle-invasive bladder cancer. They had to have at least 50% urothelial histology. In cohort A, they looked at CIS with or without papillary tumors. For those who had papillary tumors, a complete TURBT was required and muscle was required to exclude muscle-invasive disease.

There are some important exclusion criteria, including intravesical chemotherapy or immunotherapy following the TURBT, the presence of muscle-invasive or extra-vesicle disease, systemic chemotherapy, targeted small molecule therapy, radiotherapy, prior immune checkpoint inhibition, or active infection with HIV, hepatitis B or C, or pneumonitis. Pembrolizumab was administered 200 milligrams IV every 3 weeks. This was continued for up to 2 years but was stopped if there was evidence of centrally confirmed disease persistence, recurrence, or progression, whether there was unacceptable toxicity or if patients withdrew their consent. Disease assessment was performed on trial with cystoscopy, urine cytology, and biopsy or imaging as required.

This was performed initially every 3 months for 2 years, and every 6 months out to 5 years. In the presence of cystoscopic abnormalities, a directed biopsy was mandated, whereas patients who had positive cytology were required to undergo random biopsies. A CT urogram was performed every 6 months for the first 2 years, and then annually thereafter out to 5 years.

This highlights the overview of the study schema. After enrollment, we see the patients receive pembrolizumab and then have their first disease assessment at 12 weeks. For patients who do not have a complete response at 12 weeks, they discontinue their therapy at that point. For those patients who have a complete response, they continue to their further disease assessments. Patients who have no evidence of recurrence or progression or any high-grade disease, continue pembrolizumab for up to 2 years. And if high-risk disease is present at any assessment, they discontinued treatment and enter survival follow-up.

It's important to note that persistent disease is defined as CIS at the first assessment here at 12 weeks, and recurrent disease is defined as papillary tumors. And notably, low-grade Ta recurrences were not considered treatment failures, and these patients were allowed to continue with their pembrolizumab therapy after resection of these papillary lesions.

The primary outcome of KEYNOTE-057 is a complete response rate of patients with high-risk disease. A complete response rate was defined as the absence of high-risk disease or of progressive disease and this was assessed by central review. There were a variety of secondary endpoints, including, firstly, the safety of this regime, as well as the complete response rate of any disease, which then further included no high-risk disease or no low-grade Ta recurrences, as well as the duration of response for high-risk disease, duration of response overall, progression-free survival to worsening grade, stage, or death, progression-free survival to muscle-invasive disease, and overall survival. Finally, they also looked at an exploratory endpoint looking at the patient-reported quality of life.

In terms of statistical analyses, accrual of 130 patients was planned to allow for 97% power to reject the null hypothesis with a lower bound of the 95% confidence interval, excluding 20%. And 20% was selected based on historical data from the registration trial of valrubicin in the BCG-unresponsive disease space. They underwent two pre-specified interim analyses, and as a result of these, the FDA revised the target sample size to 100. They estimated the point estimate for complete response and calculated the 95% confidence intervals using the exact binomial approach and used the Kaplan-Meier method for time to event data. And additionally, they performed multiple subgroup analyses to help us better understand patient and disease characteristics associated with treatment response.

At this point in time, I am now going to hand it over to Zach to walk us through the results of KEYNOTE-057.

Zachary Klaassen: Thanks, Chris. There were 334 patients that were assessed for eligibility for this trial. Ultimately, 47 patients were assigned into cohort B and 101 patients were enrolled into cohort A and assigned to pembrolizumab. 101 of these patients received at least one dose of pembrolizumab. Ultimately, 96 were included in the efficacy analysis, and 101 of those patients that had received at least one dose were included in the safety analysis.

This looks at the baseline characteristics for patients included in this trial. The median age was 73 years. 70% of patients were older than 65 years of age, with the majority of patients being male, at 84%. 73% were ECOG performance 0 and the median previous BCG installations were 12, with an IQR of 9 to 16.5. When we look at the tumor stage at the time of enrollment, carcinoma in situ with T1 disease was 12% of patients, CIS with high-grade Ta is 25% of patients, and CIS alone was 63% of patients. There were 38% of patients that had a combined positive score greater than or equal to 10 for PD-L1 status, and there were 57% that had less than 10. Looking at the reasons for not undergoing cystectomy, the vast majority, 95%, was because they declined radical cystectomy. And in terms of the BCG failure category, the persistent disease was 26% and the recurrent disease was 69%.

This looks at the best overall response at month three by central review in patients with BCG-unresponsive CIS. In this cohort A efficacy population of 96 patients, 41% of patients had a complete response at 3 months, 58% had a non-complete response, 42% had the persistent disease, and 6% of patients had recurrent disease.

This curve looks at the estimate of the duration of complete response of high-risk non-muscle-invasive bladder cancer among patients with a complete response. The median duration of complete response was 16.2 months, with a 95% confidence interval of 6.7 to 36.2. Notably, 46% of responders remained in complete response for 12 months.

This swimmer plot looks at the time to complete response and recurrence of high-risk non-muscle-invasive bladder cancer in patients with a complete response. You can see here that the 39 patients with a complete response, 28% of patients had an ongoing response at the time of the data cutoff and 51% of patients had recurrent disease after initial complete response.

This table looks at the pathological staging at the time of radical cystectomy in patients who discontinued pembrolizumab. You can see here that this includes 38 patients, the majority of which had non-muscle-invasive bladder cancer at the time of cystectomy. 6 patients with pT0, 5 patients with pTa, 18 patients with pTis, and 6 patients with pT1. Only 3 patients, 2 with T2 and 1 with T3, had muscle-invasive bladder cancer at the time of radical cystectomy.

This table looks at treatment-related adverse events, and you can see here, there were very few grade 3 and grade 4 adverse events. Any grade 3 was 11% and any grade 4 was 2%. 2% of patients had grade 3 arthralgia, as well as grade 3 hypernatremia. So, overall, a very well-tolerated treatment.

So several discussion points from the much anticipated KEYNOTE-057 publication. Notably, one of the key premises of KEYNOTE-057 was to identify an alternative treatment option for patients not undergoing radical cystectomy. Ultimately, the results of KEYNOTE-057 showed a clinically meaningful benefit with several points to be made; First, a complete response rate of 41% of the first evaluable efficacy assessment; Secondly, after greater than 2 years of follow-up, 28% of complete responders had an ongoing response; and third, 46% of complete response patients had at least a 12-month duration of response at the time of the data cutoff.

In the 40% Of patients who underwent radical cystectomy, it appeared that the window of opportunity for curative intent was preserved and that only 3 patients, or 8% of those patients, were pathologically upstaged to muscle-invasive bladder cancer. Importantly, the safety of the radical cystectomy was also preserved during this timeframe, as no new safety signals among those undergoing cystectomy were found after discontinuing pembrolizumab.

In conclusion, the results of the KEYNOTE-057 trial support the use of pembrolizumab as a clinically active non-surgical treatment option in patients with BCG-unresponsive bladder CIS who are ineligible for or decline radical cystectomy. The FDA approval of pembrolizumab fulfills a major unmet need and establishes a foundation and benchmark for future trials that could provide higher and more durable responses for patients with BCG-unresponsive CIS.

Thank you very much and we hope you enjoyed this UroToday Journal Club discussion of the KEYNOTE-057 trial recently published in The Lancet Oncology.