ASCO-GI – Exelixis’s son of Cabometyx disappoints

Exelixis’s big product Cabometyx is set to come off patent in 2030, and the company hopes to replace it with a next-generation project, zanzalintinib. However, doubts about the new asset emerged this week at the ASCO Gastrointestinal Cancers symposium.

For one, zanzalintinib, both with and without Roche’s Tecentriq, worked much better in late-line colorectal cancer patients without liver metastases, versus those whose cancer had spread to the liver, the Stellar-001 study found. And on a cross-trial basis zanzalintinib plus Tecentriq doesn’t look better than Cabometyx plus Tecentriq – on efficacy or safety.

As liver involvement is a common problem in colorectal cancer, this could limit the market for zanzalintinib. This is a live issue given that Exelixis has said that the project could make $5bn by 2033, with nearly half of this coming from gastrointestinal tumours.

Exelixis’s stock fell 10% on Wednesday. Big catalysts await, with data due this year from several pivotal trials of zanzalintinib, including Stellar-303, testing a Tecentriq combo in relapsed colorectal cancer. That trial will primarily evaluate overall survival in patients without liver mets; the secondary endpoint looks at OS in all-comers.

Stellar-001

Cabometyx and zanzalintinib are both multikinase inhibitors; Exelixis hopes that the latter could have similar efficacy, but with a better safety profile, helped by a shorter half-life. That’s still an open question, the latest data suggest.

Stellar-001 is a basket trial that includes a colorectal cancer cohort, which enrolled RAS wildtype, non-MSI-high patients refractory to standard of care chemo and/or targeted therapies.

The colorectal data, revealed in an ASCO-GI abstract, showed a better response in the roughly one third of patients who didn’t have liver mets.

Not out of this world: Stellar-001 results

Source: ASCO-GI abstract.

The intent-to-treat results look in line with those seen with Cabometyx plus Tecentriq in the colorectal cancer cohort of Cosmic-021, which found an ORR of 10%, median progression-free survival of three months, and median overall survival of 14 months.

Furthermore, those figures from Cosmic-021 included RAS mutants (a post-hoc analysis showed better results in wildtype patients), and 80% of patients had liver mets – so the cross-trial comparison could flatter Stellar-001.

Any potential safety advantage with zanzalintinib also looks far from clear. The most common treatment-related adverse events in Stellar-001 were diarrhoea (52% with zanzalintinib plus Tecentriq), nausea (54%) and decreased appetite (41%). 48% of patients receiving the combo had grade 3 or 4 TRAEs, and one patient in each dosing arm had a fatal TRAE.

Meanwhile, Cosmic-021 found diarrhoea at 52%, fatigue at 42%, and nausea at 35%. Grade 3/4 TRAEs were seen in 52% of patients, and there were no treatment-related deaths.

Notably, though, Cabometyx isn’t approved in colorectal cancer; it’s only marketed for renal, liver and thyroid cancers (the last also under a different trade name, Cometriq).

As well as the upcoming Stellar-303, zanzalintinib is in various phase 3 trials in renal and head and neck cancers, and neuroendocrine tumours. Merck also plans to test it in combination with its HIF2α inhibitor Welireg in renal cancer, under a deal signed last year.

Exelixis will need to show more emphatic results than this if its franchise-management efforts are to succeed.

Phase 3 trials of zanzalintinib 

Note: *phase 2/3; **under Merck collaboration, Merck will fund one trial, Exelixis will co-fund the other. Source: OncologyPipeline & clinicaltrials.gov.