Patients who received pembrolizumab plus cisplatin and gemcitabine achieved a median overall survival of 12.7 months, compared with 10.9 months with cisplatin and gemcitabine alone.
Adding pembrolizumab (Keytruda) to treatment with cisplatin and gemcitabine yielded a statistically significant and clinically meaningful improvement in overall survival in patients with advanced biliary tract cancer, vs cisplatin and gemcitabine alone, according to data from the phase 3 KEYNOTE-966 trial (NCT04003636).1
Findings presented at the 2023 AACR Annual Meeting showed that at the December 15, 2022, data cutoff, patients in the pembrolizumab arm (n = 533) experienced a median OS of 12.7 months (95% CI, 11.5-13.6), compared with 10.9 months (95% CI, 9.9-11.6) for those in the placebo arm (n = 536; HR, 0.83; 95% CI, 0.72-0.95; P = .0034). The 12- and 24-month OS rates in the pembrolizumab group were 52% and 25%, respectively. In the placebo arm, those rates were 44% and 18%, respectively.
“Results from KEYNOTE-966 support pembrolizumab plus gemcitabine and cisplatin as a new treatment option for patients with previously untreated metastatic or unresectable biliary tract cancer," lead study author Robin Kate Kelley, MD, said in a presentation of the data. Kelley is a professor of medicine in Department of Medicine at the University of California, San Francisco Helen Diller Family Comprehensive Cancer Center.
The combination of cisplatin and gemcitabine has been considered the standard of care in advanced biliary tract cancer. Although single-agent immune checkpoint inhibition has not displayed high efficacy in biliary tract cancer, gemcitabine and cisplatin could help promote an immune response, leading to the rationale to add pembrolizumab to the chemotherapy backbone.
The randomized, double-blind KEYNOTE-966 trial enrolled patients with histologically confirmed unresectable, locally advanced or metastatic extrahepatic/intrahepatic cholangiocarcinoma or gallbladder cancer who had measurable disease per RECIST v1.1 criteria. Patients were not allowed to have prior systemic therapy, and they needed to have an ECOG performance status of 0 or 1, and a life expectancy of more than 3 months.
Patients were randomly assigned to receive 200 mg of pembrolizumab or placebo once every 3 weeks for up to 35 cycles. All patients received 1000 mg/m2 of gemcitabine and 25 mg/m2 of cisplatin on days 1 and 8 of each 21-day cycle.
“It’s important to note that cisplatin was discontinued after a maximum of 8 cycles in both arms, which is standard of care for cisplatin due to its risk for cumulative toxicity,” Kelley explained. “Gemcitabine could be continued, with no maximum number of cycles, until progression or intolerable toxicity, according to investigator discretion or local practice patterns.”
Stratification factors included geographic region (Asia vs not Asia), disease stage (locally advanced vs metastatic), and site of origin (extrahepatic cholangiocarcinoma vs gallbladder vs intrahepatic cholangiocarcinoma).
OS served as the trial’s primary end point. Secondary end points included progression-free survival (PFS), overall response rate (ORR), duration of response (DOR), and safety.
Kelley noted that baseline characteristics were generally balanced between the treatment arms. The median age of patients in the pembrolizumab arm was 64 years (interquartile range [IQR], 57-71), compared with 63 years (IQR, 55-70) in the placebo arm. The majority of patients were male (53% and 51% in the pembrolizumab and placebo arms, respectively), not from Asia (55% and 54%), and had an ECOG performance status of 1 (51% and 57%). Notably, 46% and 47% of patients in the experimental and control arms, respectively, were Asian.
Site of disease origin consisted of intrahepatic cholangiocarcinoma (60% and 58% in the experimental and control arms, respectively), extrahepatic cholangiocarcinoma (18% and 20%), and gallbladder cancer (22% and 22%). Eighty-nine percent and 88% of patients had metastatic disease in the pembrolizumab and placebo arms, respectively. In the experimental arm, 6% of patients had a biliary stent or drain, 9% had prior neoadjuvant or adjuvant chemotherapy, and 55% received antibiotics within 1 month of the start of study treatment. Those rates were 8%, 9%, and 51%, respectively, in the control arm.
One percent of patients in each arm were microsatellite instability high, 68% in each arm had a PD-L1 combined positive score of at least 1, and 31% in each arm had a hepatitis B infection. Four percent and 3% of patients in the experimental and control arms had a hepatitis C infection, respectively.
Additional data showed that at the first interim analysis, the median PFS was 6.5 months (95% CI, 5.7-6.9) for patients in the pembrolizumab arm, compared with 5.6 months (95% CI, 5.1-6.6) in the placebo arm (HR, 0.86; 95% CI, 0.75-1.00; P = .0225). The 6- and 12-month PFS rates were 52% and 25% in the pembrolizumab arm, respectively, vs 46% and 20% in the placebo arm.
At the final analysis, pembrolizumab plus cisplatin/gemcitabine elicited a median PFS of 6.5 months (95% CI, 5.7-6.9), vs 5.6 months (95% CI, 4.9-6.5) for placebo plus cisplatin/gemcitabine (HR, 0.87; 95% CI, 0.76-0.99). In the pembrolizumab arm, the 12- and 24-month PFS rates were 24% and 9%, respectively. Those rates were 19% and 5%, respectively, in the placebo arm.
At the first interim analysis, patients treated with pembrolizumab plus cisplatin/gemcitabine achieved an ORR of 29% (95% CI, 25%-33%), including a complete response (CR) rate of 2% and a partial response (PR) rate of 27%. Those who received placebo plus cisplatin/gemcitabine experienced an ORR of 29% (95% CI, 25%-33%), with a CR rate of 1% and a PR rate of 27%. The difference in ORR between the two arms was 0.2% (95% CI, –5.2%-5.6%) favoring pembrolizumab.
The median DOR was 9.7 months (95% CI, 6.9-12.2) in the pembrolizumab arm, compared with 6.9 months (95% CI, 5.7-8.2) in the placebo arm. For the pembrolizumab group, the 6- and 12-month DOR rates were 67% and 41%, respectively. Those rates were 56% and 28%, respectively, in the placebo group.
At the final analysis, the ORR was 29% (95% CI, 25%-33%) for the pembrolizumab arm vs 28% (95% CI, 25%-32%) for the placebo arm, translating to a difference of 1.0% (95% CI, –4.4%-6.4%) favoring pembrolizumab. CRs were observed in 3% and 2% of patients in the pembrolizumab and placebo groups, respectively. The PR rate was 27% in both arms.
Pembrolizumab plus cisplatin/gemcitabine generated a median DOR of 8.3 months (95% CI, 6.9-10.2), compared with 6.8 months (95% CI, 5.7-7.1) for placebo plus cisplatin/gemcitabine. The 12 and 24-month DOR rates were 38% and 18% in the pembrolizumab arm, respectively, compared with 27% and 6% in the placebo arm, respectively.
Regarding safety, any-grade adverse effects (AEs) occurred in 99% (n = 524/529) of patients in the pembrolizumab arm, vs less than 100% (n = 532/534) of patients in the placebo arm. The rates of grade 3/4 AEs were 79% and 75% in the experimental and control groups, respectively. Six percent of patients in the pembrolizumab arm experienced AEs leading to death, compared with 9% in the placebo arm. Seven percent of patients in each arm discontinued all treatment due to AEs.
Any-grade AEs that occurred in at least 15% of patients included decreased neutrophil count (62% and 61% in the pembrolizumab and placebo arms, respectively), anemia (61% and 59%), nausea (44% and 46%), decreased platelet count (40% and 40%), fatigue (35% and 32%), constipation (35% and 36%), decreased appetite (27% and 29%), decreased white blood cell count (27% and 24%), pyrexia (26% and 19%), vomiting (23% and 24%), diarrhea (19% and 18%), abdominal pain (17% and 23%), rash (17% and 9%), increased aspartate aminotransaminase (17% and 18%), increased alanine aminotransaminase (16% and 21%), hypomagnesemia (15% and 15%), pruritus (15% and 10%), asthenia (14% and 18%), and peripheral edema (14% and 16%).
Any-grade immune-mediated AEs and infusion reactions occurred in 22% of patients in the pembrolizumab arm, including 7% of patients who had grade 3/4 events, compared with 13% in the placebo arm, with 4% experiencing a grade 3/4 event. One patient treated with pembrolizumab plus cisplatin/gemcitabine died due to pneumonitis, and no patients in the placebo arm died due to immune-mediated AEs/infusion reactions. Nine percent of patients in the pembrolizumab required treatment with corticosteroids, compared with 5% in the placebo arm.
The most common immune-mediated AEs/infusion reactions reported in at least 3 patients in the pembrolizumab arm were hypothyroidism, pneumonitis, hyperthyroidism, severe skin reactions, colitis, hepatitis, infusion reactions, pancreatitis, adrenal insufficiency, thyroiditis, hypophysitis, nephritis, and vasculitis.
Following the presentation, Kelley was asked how findings from KEYNOTE-966 contribute the biliary tract cancer treatment landscape following prior results from the phase 3 TOPAZ-1 trial (NCT03875235), which supported the FDA approval of durvalumab (Imfinzi) in combination with gemcitabine and cisplatin in adult patients with locally advanced or metastatic biliary tract cancers in September 2022.2
“[KEYNOTE-966] reinforces and validates the findings from TOPAZ-1, that adding an immune checkpoint inhibitor to chemotherapy does improve survival and long-term outcomes [in patients with biliary tract cancer],” Kelley said. “In distinction to TOPAZ-1, KEYNOTE-966 adds several nuances to the data. One might be that a larger study certainly adds confidence in the results, but [KEYNOTE-966] also had a slightly different patient population [with] a greater proportion of patients in Western and non-Asian sites. Further, it really addresses the variability in practice patterns worldwide by allowing the continuation of gemcitabine beyond that 6-month mark.”
Disclosures: Dr Kelley reported serving as a consult or advisor, with fees to the institution, for Agios, AstraZeneca, Bristol Myers Squibb, Exelixis, Ipsen, and Merck Sharp & Dohme; as a consultant or advisor, with fees to herself, for Compass Therapeutics, Exact Sciences, Kinnate, Regeneron, and Tyra Biosciences; receiving research support to the institution from Agios, AstraZeneca, Bristol Myers Squibb, Eli Lilly, EMD Serono, Exelixis, Genentech/Roche, Loxo Oncology, Merck Sharp & Dohme, Novartis, Partner Therapeutics, QED, Relay Therapeutics, Surface Oncology, and Taiho. The KEYNOTE-966 study was funded by Merck Sharp & Dohme.