Modest Survival Gain in Gastric/GEJ Cancer With Chemo Plus Pembrolizumab

Adding pembrolizumab (Keytruda) to chemotherapy modestly improved survival in unresectable/metastatic gastric or gastroesophageal junction (GEJ) cancer, a randomized trial showed.

Median overall survival (OS) improved from 11.5 months with chemotherapy alone to 12.9 months with the addition of the immune checkpoint inhibitor. Median progression-free survival (PFS) improved from 5.6 months to 6.9 months. Landmark survival analyses at 12 and 24 months favored the pembrolizumab arm, and the benefits were generally consistent across multiple subgroups, including age, geography, and PD-L1 expression.

Pembrolizumab did not add significantly to toxicity, as the types and severity of treatment-related adverse events (TRAEs) were similar in the two treatment groups, reported Sun Young Rha, MD, of Yonsei Cancer Center in South Korea, during the European Society for Medical Oncology (ESMO) Virtual Plenary.

"These data support pembrolizumab plus chemotherapy as a new treatment option for patients with locally advanced or metastatic, HER2-negative gastric or gastroesophageal junction adenocarcinoma," she said in conclusion.

The study continued a pattern of uneven results with PD-1/L1-targeted therapies in advanced GEJ cancer, according to invited discussant Elizabeth Smyth, MD, of the University of Cambridge in England. The KEYNOTE-859 trial that Rha reported makes an even dozen since 2017, involving five different PD1/L1 inhibitors. Six of the trials showed improved survival with anti-PD1/L1 therapy, five showed no improvement, and one demonstrated noninferiority versus chemotherapy.

"We've learned that the majority of gastric cancers are not sensitive to anti-PD-1 monotherapy, with response rates on biomarker-unselected populations of around 10%," said Smyth. "When we compare anti-PD-1 monotherapy to placebo, it is superior. However, anti-PD-1 is not superior to chemotherapy in unselected or even in PD-L1 weakly positive populations. Finally, using chemotherapy plus anti-PD-1 together improves response rates and overall survival, with the most benefit in patients with PD-L1-high tumors."

The KEYNOTE-859 data showed a statistically significant 22% reduction in the survival hazard with pembrolizumab. In absolute terms, however, the improvement amounted to less than 6 weeks, which some authorities might not consider clinically meaningful, said Smyth.

The 2-year landmark survival analysis showed a 9.3% absolute improvement (28.2% vs 18.9%) in OS with pembrolizumab and a 50% relative improvement, she continued. However, the magnitude of the improvement fell short of the 10% threshold for an ESMO Magnitude of Clinical Benefit Scale score of 4, suggesting "substantial benefit" for an intervention.

"These changes are real and substantial; however, only a third of patients are alive for 2 years," said Smyth. "How do we correctly identify the patients who are deriving this benefit?"

Available evidence suggests that use of biomarkers, such as microsatellite instability (MSI) and PD-L1 expression, offer the most promise. Though currently flawed and arbitrary, cutoffs for PD-L1 expression offer the potential to strike a balance between overtreatment and undertreatment with anti-PD-1/L1 agents.

Responding to Rha's statement that pembrolizumab and chemotherapy represent a new standard treatment option for advanced/metastatic gastric and GEJ cancers, Smyth said, "Treatment decisions will be driven by regulatory approvals, funder appraisals, and clinician and patient discussion of benefit according to PD-L1 status. PD-L1 needs further exploration to optimize confidence in testing and cutoffs."

Investigators in the international phase III KEYNOTE-859 trial randomized 1,579 patients with previously untreated unresectable/metastatic gastric or GEJ cancer to pembrolizumab or placebo. Patients in both arms received either cisplatin- or oxaliplatin-containing chemotherapy. The trial had a primary endpoint of OS.

Rha said that baseline characteristics were well balanced between the treatment arms. All patients in both arms had pathologically confirmed HER2-negative disease. About 5% had MSI-high status, 78% had a PD-L1 composite positive score (CPS) ≥1%, and 35% had a baseline CPS ≥10%.

The primary analysis yielded a hazard ratio of 0.78 in favor of pembrolizumab (95% CI 0.70-0.87, P<0.0001). The 12-month OS rate was 52.7% with pembrolizumab and 46.7% without, and the magnitude of the difference increased by 28.2% versus 18.9% at 24 months.

The data showed a statistically significant 24% reduction in the hazard for disease progression or death with pembrolizumab versus chemotherapy alone (95% CI 0.67-0.85, P<0.0001). Landmark analyses showed differences of 28.9% versus 19.3% at 12 months and 17.8% versus 9.4% at 24 months.

Pembrolizumab led to a significantly higher objective response rate (51.3% vs 42.0%, P=0.00009) and a longer median duration of response (8.0 vs 5.7 months).

Grade ≥3 TRAEs occurred in 59.4% of the pembrolizumab arm and 51.1% of the placebo group. Rates of serious TRAEs were 23.4% and 18.6%, and rates of discontinuation because of TRAEs were 26.4% and 20.1%.

As expected, immune-mediated adverse events occurred more frequently with pembrolizumab (27.1% vs 9.3%), reaching grade ≥3 in 7.9% versus 1.7%.

Comment