ICIs in Gastric/GEJ
How I Use Checkpoint Inhibitors in Later-line Advanced Gastric/GEJ Cancers

Released: July 18, 2019

Expiration: July 16, 2020

Tanios Bekaii-Saab
Tanios Bekaii-Saab, MD, FACP

Activity

Progress
1
Course Completed

Whereas immune checkpoint inhibitors (ICIs) have transformed cancer care, these agents benefit only a subset of patients in the refractory setting for gastric/gastroesophageal junction (GEJ) cancers. How can clinicians best identify patients who are most likely to have good outcomes with available and emerging ICIs in this setting?

Clinical Trial Data on Later-line ICIs
The most extensively studied ICIs in later-line gastric/GEJ cancers are pembrolizumab and nivolumab. In 2017, the FDA granted accelerated approval to pembrolizumab for recurrent advanced/metastatic gastric/GEJ adenocarcinoma expressing PD-L1 (defined as a combined positive score CPS ≥ 1) with progression on or after at least 2 previous chemotherapy lines based on the single-arm phase II KEYNOTE‑059 trial. Among PD-L1–unselected patients with previously treated gastric/GEJ cancer, pembrolizumab was associated with an ORR of 11.6%, with responding patients tending to have a durable response (median duration: 8.4 months). Of importance, the ORR was higher in those with PD-L1–positive (CPS ≥ 1) vs PD-L1–negative tumors (15.5% vs 6.4%, respectively).

The subsequent phase III KEYNOTE-061 and KEYNOTE-181 trials showed that pembrolizumab did not improve outcomes vs standard of care in those with previously treated, advanced gastric/GEJ cancer and low to no PD-L1 expression. In KEYNOTE-061, there was no difference in median OS for those with CPS ≥ 1. Furthermore, for those with CPS < 1, pembrolizumab had a detrimental effect on median OS, which was 4.8 months with pembrolizumab vs 8.2 months with paclitaxel (HR: 1.20). The median OS was improved with pembrolizumab vs paclitaxel only in those with CPS ≥ 10 (10.4 vs 8.0 months) or with microsatellite instability (MSI)–high disease (not reached vs 8.1 months). Similar positive results for those with CPS ≥ 10 were reported in KEYNOTE-181, which compared second-line pembrolizumab vs chemotherapy in advanced esophageal cancer. Although the FDA approval is for CPS ≥ 1, these data support CPS ≥ 10 as the threshold for using ICIs.

Notable data for nivolumab come from the randomized, placebo-controlled phase III ATTRACTION-2 trial in PD-L1–unselected Asian patients with unresectable advanced or recurrent gastric/GEJ cancer that was refractory or intolerant to standard treatment. Nivolumab significantly improved both the median OS (5.26 vs 4.14 months with placebo; HR: 0.63; P < .0001) and the ORR (11.2% vs 0%; P < .0001). Based on these results, the pan-Asian–adapted European Society for Medical Oncology guidelines recommend nivolumab monotherapy for third-line and later settings in metastatic gastric cancer, although US guidelines still consider the data for nivolumab too preliminary at this time.

It is important to note that Asian populations can have different drivers for gastric cancer compared with Western (including Asian American) populations, making Asian patients more likely to respond to ICIs even when CPS < 10. Asian gastric cancers tend to be more distal, less aggressive, and detected earlier. Epstein-Barr virus (EBV) is also more endemic in Asia, increasing the likelihood that an Asian patient has an EBV‑positive tumor—which is more likely to respond to ICIs. Altogether, these differences make it difficult to apply clinical data from one part of the world to another.

Identifying Patients Likeliest to Benefit from Later-line ICIs
These trial data show that when determining which patients with gastric/GEJ cancer could benefit from ICIs, we absolutely need to know the patient’s MSI status and PD-L1 CPS. ICIs should be considered at progression, regardless of HER2 status. Given that data suggest that trastuzumab beyond progression does not offer benefit, HER2-positive patients will follow the same algorithm as HER2-negative patients when selecting second-line and third-line treatments. At progression, if the patient has MSI-high disease and/or has a CPS ≥ 10, I would suggest using second-line pembrolizumab or, to a lesser degree, nivolumab. Those with CPS ≥ 1 and < 10 should receive second-line paclitaxel plus ramucirumab, reserving ICIs for the third line. Patients with ECOG PS 2 should receive second-line ramucirumab monotherapy, and I would not aggressively treat patients with ECOG PS 3/4. Finally, those with CPS < 1 should not receive ICIs because these agents may have a detrimental effect on patient outcomes.

Future Directions
Moving forward, we can expect to see phase III data presented from ATTRACTION-3 for nivolumab in PD-L1–unselected, refractory esophageal cancer. There are also ongoing studies on ICIs in the first-line setting for gastric/GEJ cancer (eg, ATTRACTION-4); given the positive first-line results from the phase III KEYNOTE-062 trial, I would now consider moving pembrolizumab into the first line for MSI-high or even perhaps EBV-positive patients. Other efforts are examining the combination of CTLA-4 plus PD-1 inhibitors, although given the concerns about toxicity and cost, I would not yet recommend using this combination in the clinic.

When do you recommend ICIs to your patients with advanced gastric/GEJ cancers? Please share your thoughts in the comments box!

Poll

1.
Which of the following biomarkers are you testing for in your patients with gastric/GEJ cancers? Please select all that apply. 
Submit