Tidying up US stomach cancer approvals
A US adcom will next month try to unravel confusion about PD-L1 status in first-line gastric cancer.
A US adcom will next month try to unravel confusion about PD-L1 status in first-line gastric cancer.
Regulatory approvals of checkpoint blockers for stomach cancer paint a confusing picture governed by histology, PD-L1 expression and HER2 status. Now the FDA has determined to make sense of at least part of the mess, convening an advisory committee meeting that will next month discuss whether existing and future approvals should be limited to PD-L1-expressing patients.
The most exposed companies in the stomach cancer settings concerned appear to be Bristol Myers Squibb, whose Opdivo carries three all-comers first-line labels, and BeiGene, which is awaiting two US Tevimbra approvals, one of which was recently delayed. Merck’s Keytruda is at partial risk, having already had its wings clipped when a survival benefit was seen only in PD-L1 expressers.
That happened in mid-2023, and was an illustration of regulatory opinions initially diverging: on the basis of the Keynote-811 trial the EMA approved Keytruda’s Herceptin plus chemo combo in PD-L1 ≥1% expressing patients, while the FDA opted for an all-comers label – before the US regulator U-turned and restricted the label.
Adcom scope
The adcom is scheduled for 26 September, and specifically concerns approvals of checkpoint inhibitors in two cancer histologies: gastric/gastroesophageal junction adenocarcinoma, and oesophageal squamous cell carcinoma.
The adcom notice makes clear that the focus is only on first-line settings. Keytruda and Opdivo are approved for second-line use, and the latter also has a label for adjuvant oesophageal/GEJ cancer; none of those fall within the scope of this adcom.
According to the limited documents available so far, the FDA wants the adcom’s opinions on whether PD-L1 expression is adequate as a predictive biomarker, on how risk-benefit differs in different PD-L1-defined subpopulations, and – most importantly – on whether there exists enough “cumulative data to restrict the approvals of immune checkpoint inhibitors based on PD-L1 expression”.
A separate point in gastric adenocarcinoma is that Keytruda’s approvals are split between two trials, specifically in HER2 expressers and non-expressers. Opdivo plus chemo, however, is approved in all-comers even though its Checkmate-649 trial excluded HER2-positive patients; this, however, isn’t to be discussed.
So far only Bristol has issued a statement regarding the adcom, and the group accepts that “cumulative data have shown that PD-L1 expression appears to be a predictive biomarker of treatment efficacy in these patient populations”. However, it adds that Opdivo-based regimens have changed the outlook for gastric cancer patients “regardless of PD-L1 status”.
For Opdivo three approvals are at risk. For Merck two settings where Keytruda plus chemo can be given irrespective of PD-L1 status look precarious: the Keynote-859 adenocarcinoma trial showed no benefit in PD-L1 negatives, and like with Keynote-811 the EMA is at odds with the FDA, having approved only in PD-L1 ≥1% expressers; in squamous histology Keynote-590 backs an all-comers US approval but only for PD-L1 ≥10% expressers in the EU.
Three other Keytruda monotherapy studies had earlier shown what a minefield gastric/GEJ adenocarcinoma is, even though they concerned PD-L1-positive patients only. Keynote-059 backed a third-line accelerated US approval that was withdrawn after a negative adcom vote; the second-line Keynote-061 trial failed; and Keynote-062, in first-line patients, yielded inconclusive data.
Tevimbra next
It is this confusing space that BeiGene is attempting to enter with its extensively delayed Tevimbra, whose chemo combo trials Rationale-306 and 305 have been filed in support of oesophageal squamous cell and HER2-negative gastric/GEJ adenocarcinoma respectively.
The FDA has already missed a June action date on the former, a setback BeiGene put down to delays in scheduling clinical site inspections. The latter has a December PDUFA date, and is notable for its primary endpoint concerning PD-L1 ≥5% expressers, who appeared to have driven the subsequently demonstrated positive result in all-comers.
Both Tevimbra studies will be discussed at the adcom.
Stomach cancer trials to be discussed to 26 Sep adcom (except Keynote-811)
Study | Regimen | Precise 1st-line setting | HR for OS in all-comers | HR for OS in PD-L1 subgroup |
---|---|---|---|---|
Gastric and gastroesophageal junction adenocarcinoma | ||||
Checkmate-649 | Opdivo + chemo | Gastric/GEJ/oesophageal adenocarcinoma^ | 0.80 | 0.71 in PD-L1 ≥5% |
Keynote-859 | Keytruda + chemo | HER2-ve gastric/GEJ adenocarcinoma^^ | 0.78 | 0.74 in PD-L1 ≥1% |
Keynote-811 | Keytruda + Herceptin + chemo | HER2+ve gastric/GEJ adenocarcinoma* | 0.84 | 1.03 in PD-L1 <1% |
Rationale-305 | Tevimbra + chemo | HER2-ve gastric/GEJ adenocarcinoma** | 0.80 | 0.74 in PD-L1 ≥5% |
Oesophageal squamous cell carcinoma | ||||
Checkmate-648 | Opdivo + Yervoy | Oesophageal squamous cell carcinoma | 0.78 | 0.64 in PD-L1 ≥1% |
Checkmate-648 | Opdivo + chemo | Oesophageal squamous cell carcinoma | 0.74 | 0.54 in PD-L1 ≥1% |
Keynote-590 | Keytruda + chemo | Oesophageal/GEJ carcinoma*** | 0.73 | 0.62 in PD-L1 ≥10% |
Rationale-306 | Tevimbra + chemo | Oesophageal squamous cell carcinoma** | 0.66 | 0.62 in PD-L1 ≥10% |
Notes: ^EU approval is restricted to PD-L1 ≥5%; ^^EU approval is restricted to PD-L1 ≥1%; *earlier US all-comers approval already restricted to PD-L1 ≥1%; **under FDA review, all others being already approved; ***EU approval is restricted to PD-L1 ≥10%. Source: OncologyPipeline & adcom documents.
1337