Episode 8: Neo-adjuvant immunotherapy and pCR as a biomarker in G/GEJ cancer: insights from the KEYNOTE-585 + MATTERHORN study

Prof Eric Van Cutsen from University Hospitals Leuven is invited by Prof Marc Van den Eynde from Cliniques Universitaires Saint Luc to shed light on the KEYNOTE-585 and the MATTERHORN study. These two important studies in gastric cancer investigate the addition of immunotherapy in the operative and pre-operative stages. 

In KEYNOTE-585, patients with locally advanced, resectable gastric or gastroesophageal junction (G/GEJ) cancer were randomly assigned to receive either standard treatment or pembrolizumab before and after surgery. Some patients received FLOT, while others received only the chemo doublet, following the initial treatment protocol. In contrast, all patients in the MATTERHORN study received FLOT before and after surgery, with either durvalumab or placebo added. 

Patients receiving pembrolizumab in addition to chemotherapy, whether doublet or triplet, exhibited increased pathological regression, partial responses, and notably, more pCR at surgery. Despite roughly 10% more patients experiencing a pCR with pembrolizumab, this did not lead to a statistically significant difference in EFS, which is disappointing.

The MATTERHORN study revealed promising results regarding pCR, showing a significant 12% difference in favouring patients treated with FLOT plus durvalumab over FLOT alone. While this outcome is hopeful, final data on EFS, DFS, and OS are awaited for a comprehensive assessment.

Due to the less reliable nature of PD-L1 testing on biopsies during study design, patients were included regardless of their PD-L1 status. Subgroup analysis based on PD-L1 status may offer insights but current data lack the detail necessary for such assessments. Despite their significance, neither study affects clinical practice. Formally, KEYNOTE-585 was considered a negative study, and we await the results of the MATTERHORN study. 

While pCR could be a surrogate endpoint for efficacy and survival benefit, a 10% difference may not suffice to yield long-term outcomes. Interpreting KEYNOTE-585 proves challenging as most patients received doublet chemotherapy, with some receiving FLOT, necessitating a combined analysis. Conversely, MATTERHORN offers a more straightforward approach by exclusively employing FLOT. Prof Van Custem’s personal opinion suggests pCR could be an adequate surrogate endpoint if the difference is substantial, though the exact threshold remains unclear. The forthcoming period will reveal whether the 12% disparity observed in MATTERHORN translates into benefits in EFS or DFS.

Although PD-L1 expression is not a primary endpoint in these studies, it is the best option available. In metastatic patients, a consistent trend emerges: higher PD-L1 expression in tumours correlates with greater benefits from checkpoint inhibition. Directly extrapolating this from metastatic to adjuvant or neo-adjuvant therapy is not possible but the experts agree that future developments may make this feasible, particularly concerning subgroups.