KEYNOTE-859 Outcomes Influence NCCN Guidelines on Pembrolizumab in Upper GI Cancer


During a Case-Based Roundtable® event, David Zhen, MD, discussed how treatment of upper gastrointestinal cancer with pembrolizumab and chemotherapy is impacted by PD-L1 composite positive score, in the second article of a 2-part series.

David Zhen

David Zhen, MD

Associate Professor, Clinical Research Division, Fred Hutch

Associate Professor

Division of Hematology and Oncology

University of Washington School of Medicine

Seattle, WA

Targeted Oncology: In addition to CheckMate 649 (NCT02872116), what datashowed the efficacy of adding immune checkpoint inhibitor (ICI) to chemotherapy in patients with advanced upper gastric or gastroesophageal junction (GEJ) adenocarcinoma?

DAVID ZHEN, MD: The next study which is fairly new led to approval [of pembrolizumab (Keytruda)] in this patient population…was not the only study [of] pembrolizumab in the advanced setting. The first study was KEYNOTE-062 [NCT02494583]. That was for [patients with] GEJ adenocarcinoma, esophageal and gastric [cancer]. That was looking at the same exact design, newly diagnosed patients randomly assigned to chemotherapy vs chemotherapy plus pembrolizumab, and that was a negative study.1

After the CheckMate 649 [regimen] got approved, it was very confusing why we had one negative study and one positive study. There are a lot of factors behind that, but I think one of the concerns was that KEYNOTE-062 might have been an underpowered study, and maybe we weren’t able to detect the difference [between the study arms]. So, the more updated study is now KEYNOTE-859 [NCT03675737]. It had a very similar design, just like CheckMate-649, where chemotherapy [plus placebo] was compared with chemotherapy plus pembrolizumab, [an ICI]. The primary end point was a signal for overall survival [OS], and then the secondary end points were progression-free survival [PFS], [objective] response rate, and safety.

What were the OS outcomes for patients receiving pembrolizumab in KEYNOTE-859?

[Looking at] the OS for the intention-to-treat group, so the entire patient population, the median OS with pembrolizumab was 12.9 months and then with placebo was 11.5 months, with an HR of 0.78 [whose confidence interval] doesn’t cross one [95% CI, 0.70-0.87].2 So it is statistically significant but a fairly small benefit when you look at the median, although you do see the [Kaplan-Meier] curves do tail out and separate so there is an overall benefit to it.

When we break it down further based on [the patients’] PD-L1 composite positive score [CPS], if we look at CPS 1 or greater, [the median OS was] 13 months with pembrolizumab vs approximately 11.5 months with placebo [HR, 0.74; 95% CI, 0.65-0.84]. So, they were fairly similar if we still just focus on a low cutoff of 1. But if we focus on higher cutoff of CPS greater than or equal to 10, now we’re seeing a bit of a more separation of the curves. The median OS here is now 15.7 months for pembrolizumab vs [almost] 12 months for placebo [HR, 0.65; 95% CI, 0.53-0.79]. I think what we can say from this, very similar to CheckMate 649, is that you could say the overall population will probably get benefit but if you really want to select that patient population who is going to get the most benefit, it’s going to be [patients with PD-L1] CPS of 10 or greater.2,3 If you look at the National Comprehensive Cancer Network guidelines, that’s what they state; it’s a category 1 [recommendation] if patients have a CPS of 10 or greater to use pembrolizumab.4 Then if it’s 1 or greater, it’s a category 2b, so there is still a lot of debate about it.

What else informs the decision between using pembrolizumab or nivolumab (Opdivo) in an appropriate patient?

We also have to keep in mind that the PD-L1 assays that are approved with each is different. With nivolumab it’s the [PD-L1 immunohistochemistry 28-8 pharmDx assay] vs pembrolizumab [which uses] a 22C3 clone. I have seen differences in terms of the values that come up, because I had one patient who was looking for clinical trials…and interestingly, their CPS was lower on one test and it was higher on the other test. It’s very confusing, which is why I still have some issues about the [use of the] PD-L1 assay.

What were the outcomes for the secondary end points in KEYNOTE-859?

When we look at the PFS, there was improvement in the overall population, but by a very small margin [HR, 0.76; 95% CI, 0.67-0.85].2 If we break it down by PD-L1, again, the patients with higher PD-L1 CPS, especially if they’re 10 or greater, derive more benefit [HR, 0.62; 95% CI, 0.51-0.76].

Safety-wise, it was similar to what we saw with chemoimmunotherapy on the CheckMate-649 study.

1. Shitara K, Van Cutsem E, Bang YJ, et al. Efficacy and safety of pembrolizumab or pembrolizumab plus chemotherapy vs chemotherapy alone for patients with first-line, advanced gastric cancer: The KEYNOTE-062 phase 3 randomized clinical trial. JAMA Oncol. 2020;6(10):1571-1580. doi:10.1001/jamaoncol.2020.3370

2. Rha SY, Oh DY, Yañez P, et al. Pembrolizumab plus chemotherapy versus placebo plus chemotherapy for HER2-negative advanced gastric cancer (KEYNOTE-859): a multicentre, randomised, double-blind, phase 3 trial. Lancet Oncol. 2023;24(11):1181-1195. doi:10.1016/S1470-2045(23)00515-6

3. Janjigian YY, Ajani JA, Moehler M, et al. First-line nivolumab plus chemotherapy for advanced gastric, gastroesophageal junction, and esophageal adenocarcinoma: 3-year follow-up of the phase III CheckMate 649 trial. J Clin Oncol. 2024: JCO2301601. doi:10.1200/JCO.23.01601

4. NCCN. Clinical Practice Guidelines in Oncology. Gastric cancer, version 1.2024. Accessed April 9, 2024.

Related Videos
Related Content