SILVER SPRING, Md. — With some hesitation, an FDA advisory panel supported pembrolizumab (Keytruda) as a bladder-sparing option in high-risk, non-muscle-invasive bladder cancer (NMIBC) patients who don’t respond to standard bacillus Calmette-Guerin (BCG) therapy.
By a vote of 9-4, most Oncologic Drugs Advisory Committee (ODAC) members agreed that the immune checkpoint inhibitor’s rate of complete response (CR) and the durability of these responses represented a favorable risk-benefit profile for patients with “BCG unresponsive” disease.
“I voted yes. It was 51-49 in my head,” said Piyush Agarwal, MD, of the University of Chicago. “Ultimately I was concerned a lot about some real-world problems that we are now going to face. This is already the most expensive cancer, now it’s going to get even more expensive.”
Concerns were raised among panelists on both sides over the length of follow-up time in the study supporting the proposed indication, lack of quality-of-life data, the type of cystoscopy used, and whether the response lived up to previously proposed benchmarks for what constitutes a meaningful response.
“This is a difficult decision,” said Jonah Murdock, MD, PhD, of Washington DC Veterans Affairs Medical Center. “It’s not a slam dunk. I voted yes because the drug appears to be safe, and because it may help up to 19% of patients delay for quite a while, cystectomy, or not have one at all.”
‘Letting the Disease Fester’
Data supporting pembrolizumab in this setting came from cohort A of the KEYNOTE-057 trial, a phase II study that tested the PD-1 checkpoint inhibitor as a single-agent in high-risk NMIBC patients with urothelial carcinoma in situ (CIS), with or without papillary disease, who failed on BCG.
Among the 97 patients in the trial, the CR rate was 41% (95% CI 31%-52%), and 19% of the total population had a response out to a year. Median duration of response, recorded starting from the time of CR, was 16.2 months.
CR patients that recurred had a roughly 5-month delay in their radical cystectomy, and among the 36 patients in the study that initially responded to treatment but went on to radical cystectomy following progression, only three (8%) had progressed to muscle-invasive disease.
“I am concerned we’re just letting the disease fester and get into the muscle [making] the cystectomy less curative,” said Christian Pavlovich, MD, of Johns Hopkins University School of Medicine in Baltimore, who voted against the drug in this setting. He added that while pembrolizumab has clearly shown biologic activity in muscle-invasive disease, he was unconvinced that the data showed meaningful efficacy in BCG-unresponsive NMIBC.
One sticking point among some of the urologists on the panel was the use of white-light cystoscopy to detect disease response and recurrence (used in 94% of patients), as opposed to blue-light cystoscopy, which has been associated with improved detection rates. One panelist suggested that use of white light may have resulted in more favorable CR rates.
While anti-PD-1 immunotherapy agents in general are considered mostly well-tolerated, Chair Philip Hoffman, MD, of the University of Chicago, (who voted against recommending approval) said side effects with these drugs can be “unpredictable.”
Nearly all patients in the safety cohort of the trial (97.1%) had an adverse event (AE) on treatment, and 29.4% experienced grade ≥3 toxicities. Serious AEs occurred in 25.5%, and 9.8% stopped treatment for toxicity. Immune-related AEs were recorded in 20.6%, including serious immune-related AEs in 4.9%.
Guideline-recommended treatment for BCG-unresponsive NMIBC is radical cystectomy, but surgery is associated with poor quality of life, high complication rates (45% to 70%), and many of these patients — often elderly and frail — refuse the approach and seek other options.
FDA’s guidance for developing new drugs in this setting notes that “a high [CR] rate is not meaningful if the response duration is short.” Randomized trials using placebo as a comparator are considered unethical — as such, the agency has allowed single-arm registration trials in this setting. Recommendations from the International Bladder Cancer Group (IBCG) call for a 30% CR rate at 12 months, while the American Urological Association (AUA) has called for a 30% CR rate at 18 to 24 months.
“This did not get close to those benchmarks — not close,” said Pavlovich. “That very much affected me in terms of my voting.”
New Benchmarks Needed?
Mohummad Minhaj Siddiqui, MD, of the University of Maryland School of Medicine in Baltimore, said he had trouble reconciling the data against these benchmarks but ultimately voted in favor of the agent and suggested that maybe the IBCG/AUA benchmarks are too aggressive or unrealistic. And FDA staff said they wouldn’t have convened an advisory committee if they felt beholden to those benchmarks.
Agarwal said the trial establishes a benchmark for what a response should be in a BCG-unresponsive population, and provides carefully selected patients with an alternative to surgery. “I have a lot of faith that urologists will use this and advise patients judiciously,” he said.
Christian Hinrichs, MD, of the National Cancer Institute in Bethesda, Maryland, agreed. “I have a lot of confidence in the physicians counseling these patients to present this as an option and to present this in the context of what’s known and not known.”
Ultimately, a good proportion of patients in the trial went on to surgery. Heidi Klepin, MD, MS, of Wake Forest University Health Sciences in Winston-Salem, North Carolina, who voted in support of approval, nonetheless expressed concern over any delay in surgery and suggested post-marketing data be collected to examine outcomes in patients who receive pembrolizumab and then go on to cystectomy.
The FDA has defined BCG-unresponsive disease as patients with high-grade T1 disease following initial induction BCG, those with high-grade Ta/T1 disease who relapse within 6 months from their prior BCG treatment, and those with persistent CIS with or without recurrent Ta/T1 disease within a year of completing BCG.
CIS can be random and patchy, making it difficult to fully resect with a bladder-sparing approach. Left untreated, CIS is associated with high rates of progression to muscle-invasive disease, a common precursor to metastasis.
Although the FDA is not required to follow its advisory committees’ recommendations, it typically does.
- Reviewed by