Update Sustains Frontline Pembrolizumab Benefit in Urothelial Carcinoma

Peter H. O’Donnell, MD

Updated phase II results from the KEYNOTE-052 trial showed that first-line treatment with pembrolizumab (Keytruda) produced a long-lasting response for patients with cisplatin-ineligible advanced urothelial cancer.¹

As of the December 2016 data cutoff, the overall objective response rate (ORR) was 29% with a clinical benefit rate of 47%. Peter H. O’Donnell, MD, assistant professor of medicine at the University of Chicago, presented the results at the 2017 ASCO Annual Meeting.

“The overall objective response rate represents a 5% increase over the rate that has been previously reported,” O’Donnell said. “These results confirm that first-line pembrolizumab elicits clinically meaningful and durable antitumor activity in cisplatin-ineligible patients with metastatic urothelial cancer.”

In KEYNOTE-052, 370 patients with advanced, unresectable, or metastatic disease who were ineligible for treatment with cisplatin were assigned to 200 mg of pembrolizumab every 3 weeks. Patients continued treatment for 2 years, or until progression, toxicity, or withdrawal.

The median age was 74 and 29% of the cohort was 80 or older. The primary tumor location was the lower tract for 81% of patients and 21% had liver metastases.

At median follow-up of 9.5 months, 7% of patients had a complete response, 22% had a partial response, and 18% had stable disease. More than 40% of patients in the study had progressive disease.

O’Donnell said the response was consistent across demographic groups, including patients younger than 85 (29%; 95% CI, 20-35), patients ≥85 (28%; 95% CI, 15-44), and those with an ECOG performance status of 2 (27%, 95% CI, 20-35).

Among patients who had at least 1 post-baseline scan, 58% experienced a decrease in tumor lesions.

Median time-to-response was 2 months, and O’Donnell said 82% of responses lasted at least 6 months. Median duration of response has not been reached.

Investigators created a training set of the first 100 patients enrolled to identify the combined positive score (CPS) cut point for PD-L1 expression and a validation set that included all patients except for that first 100 to confirm results. Investigators found that a CPS ≥10% was the optimal enrichment cutoff for predicting response.

In the training set, patients with a CPS <10% had an ORR of 17% with 3 complete responses and 8 partial responses. ORR was 37% for CPS-high patients with 4 complete responses and 7 partial responses.

Response was even stronger in the validation set. ORR was 51% for CPS-high patients, with 14 complete responses and 27 partial responses. ORR was 23% for CPS-low patients, with 5 complete responses and 37 partial responses.

Two-thirds of patients reported treatment-related adverse events (AEs) of any grade, with fatigue (18%), pruritus (17%), and rash (12%) being the most common.

Seventy patients (19%) reported grade ≥3 AEs, including fatigue (2%) and colitis (2%). Other grade ≥3 AEs appeared in just 1% of patients.

“Importantly, only 7% of patients discontinued treatment because of a treatment-related adverse event,” O’Donnell said. “This is a rate that would be approximately one-third of that traditionally described for gemcitabine/carboplatin.”

Investigators recorded a single treatment-related AE that resulted in death—myositis in an 83-year-old patient.

O’Donnell added that the rate of immune-mediated AEs was similar to those reported in previous anti—PD-1 studies.

These results update findings first presented at the 2016 ESMO Congress.² At the time, overall ORR was 24.0% (95% CI, 16.0-33.6) at a median follow-up of 8 months. There were 6 complete responses, 17 partial responses, and 15 patients with stable disease. Forty-eight patients experienced disease progression.

Based on data from the KEYNOTE-052 trial, the FDA granted an accelerated approval in May 2017 for the accelerated approval to frontline pembrolizumab for patients with locally advanced or metastatic urothelial carcinoma who are not eligible for cisplatin-containing chemotherapy.

References

1. O’Donnell PH, Grivas P, Balar AV, et al. Biomarker findings and mature clinical results from KEYNOTE-052: First-line pembrolizumab (pembro) in cisplatin-ineligible advanced urothelial cancer (UC). J Clin Oncol. 35, 2017 (suppl; abstr 4502).
2. Balar A, Bellmunt J, O’Donnell PH, et al. Pembrolizumab (pembro) as first-line therapy for advanced/unresectable or metastatic urothelial cancer: Preliminary results from the phase II KEYNOTE-052 study. Presented at: 2016 ESMO Congress; October 7-11, 2016; Copenhagen, Denmark. Abstract LBA32.