ESMO Congress 2026

The G-HOPE-001 trial discussed at ESMO26 evaluates a CLDN18.2 ADC (rendered in Japanese katakana in the on-site post as "アルコタツグタバテカン"; standard spelling to be confirmed) in CLDN18.2-positive, third-line and beyond (3L+) gastric/gastroesophageal junction (G/GOJ) cancer. As for class evidence, none of the background references provided with this material contains any published pivotal trial of a comparable CLDN18.2 ADC or in gastric cancer, so established confirmatory evidence for this class can only be marked as undisclosed/pending confirmation within these materials. One naming coincidence needs clarification: the "HOPE-001" in the background literature is a multicenter retrospective cohort study across 11 centers in Japan, enrolling 213 patients receiving first-line pembrolizumab for advanced NSCLC with PD-L1 TPS ≥50% ^[1]; it is an entirely different trial from G-HOPE-001, with no overlap in population or asset, and cannot corroborate it.

The on-site readout comes from two posts pointing to the same trial. The post by TapiaJC1 gives: a randomized study run in collaboration between China and Japan, 408 patients, in the 3L+, CLDN18.2-positive G/GOJ setting, with 81% gastric primary and 44% previously resected primary; on efficacy, ORR 37% vs 8% (IA2), PFS at IA1 of 5 vs 2 mo, HR 0.47, with toxicity seemingly similar to alternative tx. The post by m0370 adds the hazard-ratio magnitude: PFS in the 0.4 range and OS in the 0.5 range, with the poster calling the results far beyond expectation; this should be read alongside the main readout—the relatively new information it adds is mainly the OS hazard-ratio range, not a separate trial.

Relative to existing knowledge, since the materials contain no comparable randomized data for the same ADC class, the step forward lies in the trial itself: a multinational, randomized, 408-patient controlled readout whose efficacy signals point in the same direction across two independent posts (the ORR and PFS benefit, plus an OS HR in the 0.5 range). Under the classic framework for analyzing time-to-event endpoints, HRs and median survival figures must be interpreted together with the randomized comparison and event information ^[6], and the on-site posts provide no event counts, confidence intervals, or p values, so each magnitude can only be transcribed as posted without further inference.

The limitations are equally clear: all figures come from conference social-media posts; the specific comparator regimen is undisclosed; ORR (IA2) and PFS (IA1) come from different interim analyses with the timing difference unexplained; OS is expressed only as a "0.5 range" figure; and safety is covered by a single phrase—"similar to alternative tx"—without specific incidence rates. Before the formal abstract and full publication are released, all of the above should be treated as preliminary readouts pending confirmation.