ALX Oncology has published randomized Phase 2 ASPEN-06 data in Nature Medicine showing that adding evorpacept to trastuzumab, ramucirumab and paclitaxel increased objective response rates in previously treated HER2-positive advanced gastric or gastroesophageal junction cancer. The response rate reached 40.3% with the evorpacept-containing regimen compared with 26.6% for the control regimen, while a post hoc biomarker-defined subgroup with retained HER2 expression and high CD47 expression showed a substantially larger separation. The findings strengthen the biological case for selecting patients using both HER2 persistence and CD47 expression, although the study did not satisfy every prespecified statistical criterion and ALX Oncology is not advancing a company-funded U.S. Phase 3 gastric cancer program.
ASPEN-06 enrolled 127 patients with metastatic second- or third-line HER2-overexpressing gastric or gastroesophageal junction adenocarcinoma whose disease had progressed following earlier HER2-directed therapy and fluoropyrimidine- or platinum-containing chemotherapy. Sixty-three patients received evorpacept plus trastuzumab, ramucirumab and paclitaxel, while 64 were randomized to the three-drug regimen without evorpacept.
Randomized ASPEN-06 data show a higher response rate with evorpacept added to HER2-directed therapy
The protocol-defined primary analysis showed an investigator-assessed objective response rate of 40.3% with evorpacept plus trastuzumab, ramucirumab and paclitaxel compared with 26.6% in the control arm. The 13.7-percentage-point difference exceeded the study’s prespecified threshold for demonstrating a contribution from evorpacept in the intent-to-treat population.
The statistical interpretation is more complicated than the headline response rates suggest. ASPEN-06 had two primary objectives: showing sufficient activity relative to a historical control and demonstrating a minimum response-rate improvement over the randomized internal control arm. Although the between-arm difference exceeded the prespecified threshold, the 40.3% response rate in the overall population did not meet the study’s prespecified one-sided statistical criterion against the 30% historical benchmark.
A similar pattern was observed in the prospectively defined subgroup with a fresh HER2-positive biopsy. Objective response reached 54.8% with evorpacept plus the three-drug regimen compared with 23.1% in the control arm, but the statistical test against the historical benchmark narrowly missed the predefined significance threshold.
That makes ASPEN-06 an encouraging randomized signal rather than a conventionally positive registrational trial. The study supports a contribution from evorpacept, but the strongest evidence emerged after investigators looked more deeply at which tumors retained the biological features required for the combination to work.
Retained HER2 and high CD47 expression identify the strongest evorpacept response signal
HER2 status can change after exposure to HER2-directed treatment, meaning a tumor classified as HER2-positive earlier in the disease course may no longer retain the same level of HER2 dependence when later therapy is considered. Investigators therefore evaluated whether current HER2 status, assessed through fresh tissue or circulating tumor DNA, was associated with response.
Among 95 patients with retained HER2-positive disease identified through either fresh biopsy or ERBB2 amplification in circulating tumor DNA, objective response was 48.9% with evorpacept plus trastuzumab, ramucirumab and paclitaxel compared with 25.0% in the control arm.
The signal strengthened further when CD47 expression was added to patient selection. In tumors with retained HER2 positivity and elevated CD47 expression, defined as at least 5% of tumor cells showing the highest-intensity CD47 staining, objective response reached 63.6% with the evorpacept combination versus 23.1% with control. Median progression-free survival was 19.5 months compared with 7.0 months, while median duration of response reached 25.5 months versus 8.4 months.
Those results are biologically consistent with evorpacept’s proposed mechanism. The investigational therapy blocks the CD47-SIRPα interaction, a signal tumors can use to suppress macrophage-mediated phagocytosis. By removing that inhibitory signal while trastuzumab binds HER2 on tumor cells, ALX Oncology aims to strengthen antibody-dependent cellular phagocytosis and increase immune-mediated tumor clearance.
The biomarker findings remain exploratory because the CD47 analyses were conducted post hoc rather than being prespecified as the basis for primary trial success. The subgroup sizes were also much smaller than the full randomized population, so the magnitude of benefit will require confirmation in prospectively designed studies.
Safety remained manageable but hematologic toxicity increased with the four-drug regimen
Treatment-emergent adverse events occurred in all patients in both treatment groups, reflecting the intensive chemotherapy and antibody-based backbone used in the study. Hematologic toxicities were more frequent with the addition of evorpacept. Neutropenia or decreased neutrophil count occurred in 69.8% of patients receiving evorpacept plus the three-drug regimen compared with 55.6% in the control group. Anemia occurred in 58.7% versus 38.1%, while diarrhea was reported in 41.3% versus 36.5%.
Grade 3 or higher treatment-emergent adverse events occurred in 90.5% of patients in the evorpacept arm and 79.4% of control patients. Severe neutropenia was reported in 55.6% versus 39.7%, severe anemia in 23.8% versus 17.5%, and severe leukopenia in 17.5% versus 9.5%. Despite those differences, investigators characterized the overall safety profile as manageable and generally similar between groups. Treatment-related deaths occurred in both arms, but the two fatal events in the evorpacept group were not considered related to evorpacept itself.
That safety profile is important for the broader CD47 field. Earlier approaches to CD47 blockade have sometimes been complicated by hematologic toxicity because CD47 is also expressed on normal blood cells. Evorpacept was engineered with an inactive Fc domain intended to reduce unwanted destruction of normal cells while preserving its ability to block the CD47 checkpoint.
ALX Oncology will not fund a U.S. Phase 3 gastric cancer program despite the biomarker signal
The clinical data create an unusual development situation. ALX Oncology said it has decided not to pursue a U.S. registrational Phase 3 gastric cancer program with its own resources, citing disciplined focus and resource allocation priorities. The company will instead consider development partnerships that could allow the gastric cancer program to continue.
That decision means the Nature Medicine publication does not immediately lead into a company-sponsored registration study despite the favorable biomarker-defined findings. ALX Oncology is prioritizing other evorpacept opportunities where it believes the biomarker strategy can be tested with a more focused development program.
The most important of those programs is ASPEN-09-Breast, an ongoing Phase 2 study evaluating evorpacept with trastuzumab and chemotherapy in HER2-positive metastatic breast cancer. ALX Oncology expects topline data from 80 patients in mid-2027. The breast cancer study is particularly relevant because the company is carrying forward the lesson that evorpacept activity may depend on both maintained HER2 expression and sufficient CD47 expression. If the biomarker-defined benefit observed in gastric cancer can be reproduced prospectively in breast cancer, it would provide stronger validation of ALX Oncology’s patient-selection strategy.
Cash runway gives ALX Oncology time to test the biomarker strategy in breast cancer and ALX2004
ALX Oncology ended the second quarter with $153.4 million in cash, cash equivalents and investments and expects those resources to support planned operations through the first half of 2028. The company also refinanced existing debt and obtained access to additional borrowing capacity, improving financial flexibility while its lead programs progress.
Beyond evorpacept, ALX Oncology is developing ALX2004, an EGFR-targeted antibody-drug conjugate currently in a Phase 1 dose-escalation study for EGFR-expressing solid tumors. Initial safety data from that program are expected during the second half of 2026.
ALX Oncology shares closed the previous trading session at $1.31 after declining 3.3%, extending a broader September pullback that has taken the stock down from above $2 earlier in the month. The depressed valuation increases the importance of upcoming clinical catalysts because the company remains dependent on experimental oncology programs without an approved commercial product.
The ASPEN-06 publication therefore offers both encouragement and restraint. Randomized data suggest that evorpacept contributes additional activity when added to a HER2-directed regimen, and the biomarker analysis identifies a subgroup with much stronger response durability. At the same time, the strongest CD47 findings are retrospective, the original trial did not meet every prespecified statistical objective, and ALX Oncology has chosen not to finance a Phase 3 gastric cancer program itself.
The next major test will be whether the same biomarker logic can predict benefit prospectively in HER2-positive metastatic breast cancer. If ASPEN-09-Breast reproduces the pattern seen in ASPEN-06, evorpacept could emerge as a more precisely targeted CD47 therapy rather than a broadly applied immuno-oncology agent.
