Gritstone bio’s gamble on a novel endpoint has backfired. The cancer vaccine failed to trigger hoped-for changes in circulating tumor DNA (ctDNA), causing the phase 2 trial to miss its primary endpoint and leaving the biotech clinging to immature survival data.
Investigators randomized 104 patients with metastatic microsatellite stable colorectal cancer to take one of two front-line therapies. All patients received induction and maintenance chemotherapy. Around half of the subjects also received Gritstone’s personalized neoantigen cancer vaccine, Bristol Myers Squibb’s Yervoy and Roche’s Tecentriq during the maintenance phase.
The primary endpoint looked at changes in ctDNA. On that measure, Gritstone’s drug combination was numerically worse than chemotherapy alone, with the molecular responses in the vaccine and control arms coming in at 30% and 41.7%, respectively. Gritstone attributed the result to its misunderstanding of how ctDNA would change after treatment.
“With regard to defining molecular response, we simply got it wrong,” Gritstone CEO Andrew Allen, M.D., Ph.D., said in a statement. “CtDNA levels in both arms decreased on chemotherapy for longer than we anticipated, generating similar short-term molecular response rates across arms and rendering our protocol measure of ctDNA change uninformative.”
While the study missed its primary endpoint, Allen latched on to progression-free survival (PFS) data to contend that the results are “highly encouraging.” The PFS rate was higher in the vaccine group than in the control arm after six months and nine months. However, the lines crossed around the 12-month mark, creating a short period in which PFS probability was higher in the control arm.
The hazard ratio favored the vaccine regimen, clocking in at 0.82 in the overall population, but the wide confidence intervals make it impossible to draw firm conclusions. Gritstone reported a hazard ratio of 0.52 in a subgroup of high-risk patients. PFS data are more mature in the subgroup, leading Allen to call the result “a striking signal,” but, again, the confidence intervals are wide enough that the vaccine may be less effective than the control. Allen had discussed what would be a good PFS result at an event in March.
“The separation [of the curves], of course, is likely and important, but the lifting of the tail is what we really care about,” Allen said at the time. The CEO cited Bayer’s Stivarga as an example of a drug that “shifted PFS by about two months [and is] not widely used because survival is basically no different.”
Gritstone picked ctDNA, rather than the widely used and accepted PFS, as its primary endpoint because of concerns about “pseudo progression.” The term describes people whose tumors appear to grow after treatment but then shrink. That happens when T cells enter tumors and proliferate. Initially, this causes lesions to grow, but they then collapse as the immune cells wipe out the tumor.
Evidence of pseudo progression made Gritstone “a little bit leery” about PFS, Allen said, and led it to make ctDNA the primary endpoint. The hope was that ctDNA would provide a clear signal that the therapy is working and make the case for pushing ahead to an overall survival (OS) readout that will provide the truest test of efficacy.
Instead, the study missed the ctDNA endpoint, leaving Gritstone looking to immature PFS results that, at best, show trends favoring the vaccine to make the case for its therapy. The biotech expects to have mature PFS data in the third quarter of 2024, with OS data set to follow in the first half of next year.
The timing of the readouts is important, because Gritstone’s cash is running low. The biotech ended last year with $79.2 million, a sum it told investors would fund operations into the third quarter of 2024. Gritstone proposed a public offering in the immediate aftermath of the phase 2 data Monday and later priced a $32.5 million sale at $1.65 a share.
Gritstone’s share price has bounced between a low of $1.14 and high of $3.33 over the past year. The stock fell almost 33% in the wake of the updates Monday, falling to $1.58 in after-hours trading.