Skip to main content
Premium Trial:

Request an Annual Quote

Gritstone Bio Eyes Pivotal Trial for Personalized Cancer Vaccine as Early ctDNA Endpoint Disappoints


NEW YORK – Gritstone Bio is gearing up for talks with regulators about moving its personalized neoantigen cancer vaccine, Granite, into Phase III trials even after early Phase II data showed the treatment didn't improve colorectal cancer patients' molecular responses versus chemotherapy.

On Monday afternoon, Emeryville, California-based Gritstone shared preliminary Phase II data from its ongoing Phase II/III clinical trial of the personalized Granite vaccine plus immune checkpoint inhibitors and chemotherapy versus a standard-of-care regimen involving chemo and Genentech's VEGF drug Avastin (bevacizumab) as a frontline regimen for metastatic microsatellite stable colorectal cancer patients.

Gritstone shared data from an evaluable group of 67 patients, 39 of whom were randomized to receive the Granite-containing regimen and 28 of whom received the control regimen.

Patients on the Granite arm first received induction chemotherapy while their personalized Granite vaccines were manufactured. That manufacturing process involves taking patients' tumors, using Gritstone's artificial intelligence-based Edge platform and HLA peptide sequencing to identify a series of tumor-specific neoantigens, and then engineering the vaccine specifically to target these neoantigens so that the treatment attracts CD8-positive T cells to the tumor, jumpstarting a response from the patient's immune system. On the trial, after that manufacturing process is complete, patients receive their personalized vaccine alongside Bristol Myers Squibb's CTLA4 checkpoint inhibitor Yervoy (ipilimumab), Genentech's PD-1 checkpoint inhibitor Tecentriq (atezolizumab), Avastin, and chemotherapy.

Patients on the standard-of-care control arm, meanwhile, receive induction chemo followed by Avastin and chemotherapy.

As of a clinical data cutoff of March 8 and a circulating tumor DNA cutoff of March 12, Gritstone reported that the trial had not met the endpoint it had selected as a primary outcome measure for the Phase II portion of the trial: short-term molecular response according to circulating tumor DNA.

The company had decided that relative to baseline measurements, patients whose tumors showed a 30 percent reduction in ctDNA at any time point after treatment would be considered to have a molecular response. Gritstone chose molecular response by ctDNA as its primary endpoint to get around a phenomenon that it believes would have complicated measurements of tumor shrinkage: Infiltrating immune cells may cause a tumor to appear bigger, even if the treatment was working.

"We need to open our minds to the fact that with T-cell infiltration into lesions, shrinkage may not be what we see," Gritstone CEO, President, and Cofounder Andrew Allen said back in 2021. "Molecular response may be particularly important here."

But as of Monday's preliminary readout, the molecular responses as defined by the trial's protocol weren't any better in the Granite arm than the control arm.

In fact, as of the March cutoff, the molecular response rate was slightly higher in the control arm than in the Granite arm, suggesting that the personalized neoantigen vaccine didn't improve ctDNA levels. Of the patients who received the vaccine and immunotherapy combo, 30 percent had a molecular response by ctDNA. In the control arm, meanwhile, 41.7 percent of evaluable patients experienced this molecular response.

"Short-term molecular response … is uninformative due to unanticipated continuation of ctDNA drop beyond induction chemotherapy," the company said in a press release issued alongside the data on Monday.

"Pioneering new spaces carries inherent risks, and with regard to defining molecular response, we simply got it wrong," Allen 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."

The firm noted that there seems to be a trend toward a ctDNA benefit for Granite-treated patients after a longer follow-up, with the molecular responses with the vaccine catching up to and overtaking the rates in the control arm. "Fortunately, long-term analysis demonstrates the expected correlation of ctDNA with clinical benefit and favors Granite patients," Allen said. But the observation isn't evidence of benefit and needs further confirmation.

As the ctDNA response fell flat, Gritstone chose to emphasize what it called an "early trend in progression-free survival benefit" for the Granite-treated patients.

After nine months, 78 percent of patients on the Granite-treated arm were alive without tumor progression, whereas the same was true for 62 percent of patients in the control arm. The relative progression-free survival benefit seemed to drop after about a year, though. Although the data weren't mature at the time of the data cutoff, Gritstone reported on Monday that the median progression-free survival for patients on the Granite arm was 13.01 months and the median progression-free survival for patients on the control arm was 15.28 months.

The firm also zeroed in on what it believes to be a particular benefit in a high-risk group of patients, which Gritstone defined as having baseline ctDNA levels over the median value, a variant allele frequency above 2 percent. The progression-free survival benefit in this high-risk group was more pronounced as of the recent data cutoff. High-risk patients treated with the Granite regimen lived for a median 11.57 months without their cancers progressing, whereas patients treated on the control arm lived for a median 7.06 months without their cancers progressing. "The exploratory PFS hazard ratio of 0.52 in the high-risk group, a more mature dataset, is a striking signal," Allen said.

Gritstone is anticipating more mature progression-free survival data during the third quarter of this year and plans to present overall survival data during the first half of 2025. In a Monday presentation focused on the data readout, Gritstone said the firm plans to talk with the US Food and Drug Administration about an appropriate Phase III endpoint once the Phase II progression-free survival data mature later this year, then launch a global Phase III pivotal trial in first-line microsatellite stable colorectal cancer in 2025. The firm also said it plans to discuss expanding the Granite program with prospective partners.

In the presentation, Gritstone also doubled down on what it sees as a differentiator for its personalized neoantigen vaccine approach versus its main competitors in the space, including Moderna, which has partnered with Merck on its mRNA-based bespoke cancer vaccine. While Moderna and Merck are testing their personalized cancer vaccine in patients with melanoma — a cancer type that is considered "hot" because it's generally responsive to immunotherapy — Gritstone is choosing to focus on colorectal cancer, a tumor type considered "cold." The firm is trying to turn the cold tumors hot by coupling its vaccine with immunotherapy.

Even though Gritstone emphasized that it was encouraged by its preliminary results and talked up its future regulatory discussions and Phase III strategies, investors weren't satisfied with the Phase II trial's failure to meet the ctDNA endpoint. With the preliminary readout, the company's stock price dropped more than 50 percent — from $3.28 per share on Monday afternoon to $1.29 per share on Tuesday morning.

At the same time, Gritstone announced a $32.5 million underwritten public offering Monday, which is expected to close Thursday subject to customary closing conditions.