ALX flunks its gastric test

Investor expectations had been low ahead of full readout of the Aspen-06 trial of ALX Oncology’s CD47 inhibitor evorpacept in gastric cancer – but it seems they weren’t low enough.

Waning response rates versus last October’s interim analysis mean that the trial was a bust, although the company didn’t disclose this in its press release yesterday. Instead, ALX saved a key p value for its investor presentation, and only admitted when pressed that this hadn’t hit the target.

The group also didn’t present any hoped-for data on progression-free survival, saying this and an overall survival analysis were immature.

The news is yet another nail in the coffin for the troubled CD47 field; until now, evorpacept had been one of its few success stories.

ALX talked up a subgroup analysis of Aspen-06, but investors were unconvinced, with the company's stock opening down around 30% this morning. During an investor call to discuss the data the company was cautious about committing to the phase 3 portion of the phase 2/3 Aspen-06, but there’s little else in its pipeline to pivot to.

Aspen rocky

The phase 2 part of Aspen-06 is evaluating evorpacept plus Herceptin (trastuzumab), Cyramza (ramucirumab) and paclitaxel (a combo known as TRP) versus TRP alone in second and third-line HER2-positive gastric cancer patients who had received prior HER2-targeted therapy, including AstraZeneca and Daiichi’s Enhertu.

The primary endpoint is overall response rate, and ALX’s stated aim was to show at least a 10% difference between the evorpacept and control arms.

The study achieved this, with an ORR of 40.3% with evorpacept plus TRP versus 26.6% with TRP among the full intent-to-treat population of 127 patients. However, this figure has worsened since an interim analysis last October, involving 54 patients.

The evolving Aspen-06 dataset

Note: TRP=trastuzumab, Cyramza (ramucirumab) & paclitaxel. Source: company releases.

Furthermore, the latest data cut didn’t hit statistical significance. In its investor presentation, ALX disclosed that the p value for the comparison was 0.027. When asked by analysts about the bar for statistical significance, the company’s chief executive officer, Jason Lettmann, admitted that this was 0.025.

ALX had a couple of potential explanations for the dip in response rates since October. For one, patients enrolled after the interim analysis had more aggressive disease. Secondly, fewer patients enrolled after the interim analysis had a “fresh” biopsy – defined as a biopsy after prior anti-HER2 treatment.

ALX believes that the latter is important because, it says, HER2 expression is highly variable in gastric cancer. This could mean that patients with fresh biopsies are more likely to actually have HER2-positive disease than those whose biopsies were taken some time ago.

The company highlighted a subgroup analysis of 48 patients with fresh biopsies, where it showed an ORR of 54.8% in the evorpacept plus TRP arm, versus 23.1% in the TRP arm.

When asked if the company planned to focus on this population in phase 3, ALX’s chief medical officer, Sophia Randolph, replied that it hoped at least to enrich for these patients. This raises an important practical point, however, of requiring patients to have another biopsy at enrolment – something that could slow the pace of recruitment in phase 3, she admitted.

However, when questioned whether ALX was still committed to phase 3, Lettmann replied that the group was “still digesting” the latest data. He added that the next step was talking to the FDA, and “then we’ll have a clear sense of the path forward”. This portion of the study had been slated to start by the end of 2024.

ALX is all-in on evorpacept, with its only other disclosed pipeline project being a preclinical SIRPα-directed TLR9 agonist ADC. The next big readout for evorpacept could come from the ASPEN-03 and 04 Keytruda combo trials in first-line head and neck cancer, due towards the end of this year or early 2025.