Asco – Merck reveals its gastric cancer hand


Merck & Co recently made an audacious bid to gain early approval for Keytruda in gastric cancer, potentially outmanoeuvring immuno-oncology competitors with the first regulatory OK, and the data backing this were disclosed at Asco today. These came from a large, but nevertheless single-arm cohort from the Keynote-059 trial.

Attention has focused on Keynote-059 since Merck revealed last month that an unexpected Keytruda filing had received FDA priority review and a September 22 action date. Merck’s high-stakes strategy in this indication could wrongfoot its anti-PD-1 rivals in the gastric cancer space, mirroring the coup pulled off for the chemo-Keytruda combination in first-line non-small cell lung cancer (see tables below).

Those rivals are Bristol-Myers Squibb and Pfizer/Merck KGaA, which have several phase III studies in gastric cancer under way for Opdivo and Bavencio respectively, with readouts over the next two years. Indeed, Bristol already has a positive phase III study under its belt, albeit one run by its partner, Ono, exclusively in Asia, where the disease is most prevalent.

Bristol has not disclosed whether it can use these data to support registrations in the west, but it seems likely that it will need to show evidence of effectiveness in western patients (Keytruda’s next advance encroaches on Opdivo and Tecentriq territory, May 24, 2017).


Merck’s move, if successful, would be particularly painful for Pfizer and Merck KGaA, since gastric cancer has been selected as one of the key target indications for Bavencio. These companies had expected to be first to file for gastric in western markets on the basis of the Javelin-Gastric300 study, in third-line use, which is to render results in August.

A potential ace in Merck’s regulatory hand could be its recent – and also unprecedented – approval for Keytruda in microsatellite-high tumours, as the Keynote-059 data show a very high response rate of 57% in this subgroup, albeit based on just seven patients. The MSI-high approval was based on data from patients with this phenotype from a number of studies, although largely in colorectal cancer.

The data that support the Keytruda filing are from cohort 1 of Keynote-059, which enrolled 259 patients with recurrent or metastatic gastric or gastro-oesophageal junction (G/GEJ) adenocarcinoma who had progressed on two or more prior chemotherapy regimens. PD-L1 positivity was set at an expression level of ≥1%.

Keytruda showed a response rate of 11.6% overall, and nearly 16% in PD-L1 positives. Response was higher in third than fourth-line patients, and among the former group there were higher rates, as expected, in PD-L1 positives (23%) than negatives (9%). The response rate in fourth line or later was less than half of those in third line.

Overall response rates in Keynote-059 Cohort 1
Combined (n=259) PD-L1 positive (n=148) PD-L1 negative (n=111)
Combined (n=259) 11.6%, with 2.3% CR, 9.3% PR 15.5%, with 2.0% CR, 13.5% PR 6.4%, with 2.8% CR, 3.7% PR
Third line (n=134) 16.4%, with 3.0% CR, 13.4% PR 22.7%, with 2.0% CR, 20.0% PR 8.6%, with 3.4% CR, 5.2% PR
Fourth line or later (n=125) 6.4%, with 1.6% CR, 4.8% PR N/A N/A

Keytruda’s 12% response rate appears similar to that seen with Opdivo in the 492-patient phase III trial ONO-4538-12, which recruited pretreated G/GEJ patients, but did not stratify them based based on PD-L1 status. This randomised study showed small absolute, but statistically significant, increases in PFS versus placebo (1.61 vs 1.45 months, HR=0.60; p<0.0001) and in OS (5.32 vs 4.14 months, HR=0.63; p<0.0001) and supports a recent Japanese filing.

Bavencio has also been tested in second-line G/GEJ cancer, where it showed an 18% ORR in PD-L1-positive and 9.1% in PD-L1-negative patients.

Other cohorts

Keynote-059 has a second cohort in first-line patients who receive Keytruda in combination with cisplatin and 5-FU or Xeloda in Japan, some early data from which were published at Asco. This showed a 60% response rate, with 69% in PD-L1 positives and 38% in PD-L1 negatives. There is also a third cohort of PD-L1-positive, first-line patients who receive Keytruda as monotherapy.

Merck & Co has three phase III studies under way with Keytruda in gastric cancer, with the Keynote-061 study in second-line patients due to render results in December.

Emboldened by its recent regulatory successes, the group clearly wants to give Keytruda a head start over rivals in gastric cancer. Investors only have to wait until September to find out whether its strategy worked.

Registration studies of checkpoint inhibitors in GE/GEJ cancer
Study Design Setting Enrolment Trial ID Data
ONO-4538-38 S-1 or CapeOx +/- Opdivo Adjuvant 700 NCT03006705 Jun 2021
Checkmate-649 Opdivo + Yervoy vs chemo 1L 1,266 NCT02872116 Jul 2019
Keynote-062 Cisplatin/5FU vs Keytruda vs combo 1L 750 NCT02494583 Jan 2019
- SOX/CapeOx +/- Opdivo 1L 680 NCT02746796 Aug 2020
Javelin Gastric 100 Bavencio vs continued chemo 1L maintenance 666 NCT02625610 Mar 2019
Keynote-061 Keytruda vs paclitaxel 2L, PD-L1 +ve 720 NCT02370498 Dec 2017
Keynote-063 Keytruda vs paclitaxel 2L Asian, PD-L1 +ve 360  NCT03019588 Jan 2019
Javelin Gastric 300 Bavencio vs chemo 3L 330  NCT02625623 Aug 2017
Fraction-GC Opdivo + Yervoy vs Opdivo + BMS-986016 Advanced 910  NCT02935634 Nov 2021
Note: Ordered by setting; S1=tegafur/gimeracil/oteracil; SOX=S1 + oxaliplatin.

To contact the writer of this story email Robin Davison at Asco in Chicago at or follow @RobinDavison2 on Twitter

Share This Article