A failed trial is a failed trial, says China’s regulator

Hutchmed’s plan to expand fruquintinib into uses beyond colorectal cancer has hit a roadblock, with China’s NMPA refusing to endorse a filing for second-line gastric/gastroesophageal junction adenocarcinoma based on results of the phase 3 Frutiga trial.

As a result Hutchmed has pulled the application, and says it will “evaluate a new route forward”. Though Hutchmed doesn’t spell this out in today’s statement, the glaring fact here is that Frutiga was effectively a failure, and the company was relying on subgroup analyses showing that subsequent therapy had confounded an OS benefit; the NMPA has thrown out this claim.

Taking such a hard line seems like nothing more than the logical acceptance of a real-world setting. If an established subsequent therapy can rescue placebo patients who initially progress then there seems little point in approving a novel drug that brings its own added cost and, perhaps, added toxicity.

PFS benefit not enough...

The Frutiga study, combining the VEGF inhibitor fruquintinib with paclitaxel, demonstrated a statistically significant benefit against paclitaxel alone on only one of its co-primary endpoints, progression-free survival.

The more important metric, overall survival, was a bust, with curves crossing over and a numerical 1.2-month median benefit failing to translate into a positive result across the entire trial period; the study yielded a 0.96 hazard ratio (p=0.6) and a confidence interval upper bound of 1.13, according to data presented at an ASCO virtual plenary in February.

Fruquintinib is already approved in China, where Lilly has co-marketing rights, as Elunate for third-line colorectal cancer. In the US, meanwhile, it’s been greenlit as Fruzaqla for a similar use, and there Hutchmed sold rights to Takeda for $400m, citing tough market conditions and a focus on its most profitable activities.

As such expanding the drug’s use in China is an important focus for Hutchmed, which argued that Frutiga had been confounded by control patients receiving subsequent anticancer drugs, which wiped out the possibility of an OS benefit in all-comers.

...and neither are subgroup data

At the ASCO plenary a subgroup analysis was presented, showing a 28% reduction in risk of death when the Frutiga data were cut by patients not receiving subsequent therapy (median OS 6.9 versus 4.8 months, and a nominal p value of 0.04). Meanwhile, in those who did receive subsequent therapy the hazard ratio was 0.90 (nominal p=0.3).

However, discussions with the NMPA made it clear that such an analysis couldn’t serve as the basis for approval. The regulator has evidently followed the science, as well as perhaps bearing in mind the availability in China of various other gastric cancer drugs, including Opdivo, Keytruda and Lilly’s anti-VEGFR2 MAb Cyramza.

No other gastric cancer-focused phase 3 studies are ongoing with fruquintinib, according to OncologyPipeline, and the only other pivotal trial is a Tyvyt combination in kidney cancer. A fruquintinib/Tyvyt combo is also awaiting China approval for MMR-proficient/non-MSI-high endometrial cancer, based on the uncontrolled phase 1/2 Frusica-1 trial.

Frutiga isn’t Hutchmed’s first Chinese stumble: in 2018 the phase 3 Faluca study of fruquintinib in third-line NSCLC showed a benefit on PFS but failed for OS.

Fruquintiniib's phase 3 studies

Source: OncologyPipeline.