Conference · for clinicians

ESMO 2026 Presidential Symposium II: KRAS G12C, EGFR/MET lung, CLDN18.2 gastric and biliary

10/26/2026 · 7 min read · updated 1 Nov 2026

Presidential Symposium II, ESMO 2026, Madrid — Sunday 25 October. Four phase 3 late-breakers (LBA5–8). All showed an overall survival benefit. Data as presented; Krascendo 1 and SAFFRON were published simultaneously in NEJM. Fact-checked against the papers and sponsor releases on 1 Nov 2026.

Bottom line: every trial beat an active control, not placebo. Two are lung (KRAS G12C; EGFR-mutant with MET after osimertinib), two are GI (CLDN18.2 gastro-oesophageal; first-line biliary).

LBA	Trial	Comparison	PFS	OS
5	Krascendo 1	Divarasib vs sotorasib/adagrasib, pretreated KRAS G12C NSCLC	9.8 vs 5.7 mo (HR 0.58)	17.7 vs 12.0 mo (HR 0.56)
6	SAFFRON	Osimertinib + savolitinib vs platinum–pemetrexed, post-osimertinib, MET-altered	7.6 vs 5.4 mo (HR 0.55)	17.8 vs 12.9 mo (HR 0.68)
7	CLARITY-Gastric 01	Sonesitatug vedotin vs investigator’s choice, CLDN18.2+ GEC, 2L+	ITT 4.2 vs 3.8 mo, not significant (HR 0.77)	3L+: 10.0 vs 5.0 mo (HR 0.47)
8	HARMONi-GI1	Ivonescimab + GemCis vs durvalumab + GemCis, 1L biliary	8.4 vs 5.8 mo (HR 0.61)	14.9 vs 12.1 mo (HR 0.75)

Median progression-free survival, experimental vs control (months)

Krascendo 1: divarasib9.8 mo

Krascendo 1: control5.7 mo

SAFFRON: osimertinib + savolitinib7.6 mo

SAFFRON: platinum–pemetrexed5.4 mo

CLARITY ITT: sonesitatug vedotin4.2 mo

CLARITY ITT: investigator's choice3.8 mo

HARMONi-GI1: ivonescimab + chemo8.4 mo

HARMONi-GI1: durvalumab + chemo5.8 mo

Same sources as above. CLARITY-Gastric01 ITT PFS was a dual primary endpoint that was not met (HR 0.77, P = .053).

Median overall survival, experimental vs control (months)

Krascendo 1: divarasib17.7 mo

Krascendo 1: sotorasib/adagrasib12.0 mo

SAFFRON: osimertinib + savolitinib17.8 mo

SAFFRON: platinum–pemetrexed12.9 mo

CLARITY 3L+: sonesitatug vedotin10.0 mo

CLARITY 3L+: investigator's choice5.0 mo

HARMONi-GI1: ivonescimab + chemo14.9 mo

HARMONi-GI1: durvalumab + chemo12.1 mo

LBA5–8, ESMO 2026; Krascendo 1 and SAFFRON per NEJM 2026, CLARITY-Gastric01 per AstraZeneca, HARMONi-GI1 per Akeso.

LBA5 — Krascendo 1 (divarasib)

• Design: open-label; n = 338 (167 vs 171). Divarasib 400 mg OD vs sotorasib 960 mg OD (n = 156) or adagrasib 600 mg BD (n = 15). 1–3 prior lines incl. platinum + PD-(L)1. No crossover. Median follow-up 14.3 months.
• PFS (BICR): 9.8 vs 5.7 mo; HR 0.58 (95% CI 0.44–0.76; P < .0001); 12-month PFS 43.9% vs 24.0%.
• OS (first interim): 17.7 vs 12.0 mo; HR 0.56 (95% CI 0.41–0.78; P = .0005); 12-month OS 67.1% vs 49.2%. Upper CI of divarasib median not yet estimable. Benefit seen regardless of PD-L1 or KEAP1 status (Roche).
• ORR (response-evaluable, n = 328): 58.4% vs 32.3%. DoR 11.9 vs 6.9 months.
• Toxicity: treatment-related grade 3–4 AEs lower (24.6% vs 34.5%) despite longer exposure (median 11.3 vs 5.0 months on treatment); serious TRAEs 7.8% vs 15.2%; discontinuation for TRAEs 4.8% vs 11.7%; no grade 5 TRAEs. More GI upset (diarrhoea 66%, nausea 53%, mostly grade 1); less transaminitis.

LBA6 — SAFFRON (osimertinib + savolitinib)

• Eligibility: EGFR-mutant NSCLC progressing on osimertinib; MET IHC 3+ in ≥ 90% cells and/or FISH ≥ 10 copies on a post-progression biopsy. Only 35% of those tested qualified; 1,746 screened, 338 randomised (171 vs 167). Osimertinib 80 mg OD + savolitinib 300 mg BD.
• PFS (BICR): 7.6 vs 5.4 mo; HR 0.55 (95% CI 0.42–0.72; P < .001); 12-month PFS 36% vs 10%. MET-overexpressed subset HR 0.51 (0.37–0.71); MET-amplified subset HR 0.49 (0.35–0.66), not formally tested.
• OS (final analysis, 53% maturity): 17.8 vs 12.9 mo; HR 0.68 (95% CI 0.50–0.91; P = .010); 12-month OS 70% vs 53%.
• ORR: 51% vs 27%.
• Toxicity: grade ≥ 3 AEs 64% vs 63%; peripheral oedema 49%, anaemia 26%, diarrhoea 26%; fatal AEs 12 (7%) vs 0 — 1 treatment related, 9 with progression.

LBA7 — CLARITY-Gastric 01 (sonesitatug vedotin, CLDN18.2 ADC)

• Eligibility: CLDN18.2 membranous staining ≥ 25% of tumour cells; ≥ 1 prior line. 2.2 mg/kg q3w (n = 267) vs investigator’s choice (n = 265). Control: 2L ramucirumab + paclitaxel (or taxane alone); 3L+ irinotecan, trifluridine/tipiracil (outside China) or apatinib (China). Median follow-up 8.6 months (cut-off 31 May 2026).
• OS, 3L+ (dual primary, n = 355): 10.0 vs 5.0 months; HR 0.47 (95% CI 0.36–0.62; P < .000001); 12-month OS 38.1% vs 18.1%.
• OS, ITT (key secondary): 10.6 vs 6.4 months; HR 0.58 (95% CI 0.47–0.73; P = .000004); 12-month OS 42.6% vs 27.8%.
• PFS, ITT (dual primary, BICR): 4.2 vs 3.8 months; HR 0.77 (95% CI 0.62–0.96; P = .053) — not met. 3L+ PFS 4.1 vs 2.7 months (HR 0.56), not formally tested.
• Caveats: 2L subgroup OS HR 0.88 (0.60–1.30) and GEJ/oesophageal primary HR 1.08 (0.68–1.71) — the benefit is driven by 3L+ gastric primaries.
• Toxicity: grade ≥ 3 TRAEs 36.4% vs 44.8%; more nausea/vomiting; less grade ≥ 3 neutropenia (10.3% vs 25.4%).
• Benefit seen at both CLDN18.2 ≥ 75% (ITT OS HR 0.59) and 25–74% (HR 0.63) expression.

LBA8 — HARMONi-GI1 (ivonescimab, PD-1 × VEGF bispecific)

• Design: double-blind; n = 682, all in China. Ivonescimab 20 mg/kg vs durvalumab 1500 mg, both with gemcitabine–cisplatin, then maintenance.
• OS (interim; median follow-up 14.3 mo): 14.9 vs 12.1 mo; HR 0.75 (95% CI 0.61–0.92; one-sided P = .0031). 12-month OS 59.7% vs 50.7%; 18-month OS 44.0% vs 32.5%. Subsequent therapy less often in the ivonescimab arm (31.8% vs 45.9%).
• PFS (investigator): 8.4 vs 5.8 mo; HR 0.61 (95% CI 0.51–0.72).
• ORR: 53.8% vs 32.5%; DCR 91.5% vs 77.8%. Age ≥ 65: OS HR 0.67; ECOG 1: OS HR 0.71.
• Toxicity: grade 3–4 TEAEs 82.0% vs 72.3% (longer exposure, 7.8 vs 5.7 mo); grade 3–4 neutropenia 54.6% vs 41.3%; serious AEs 43.4% vs 43.7%; grade ≥ 3 immune-related AEs 4.1% in both; VEGF effects — proteinuria ~32% vs 12%, hypertension 15.0% vs 2.4%, bleeding 9.7% vs 6.5% (grade ≥ 3 1.5% vs 1.8%).
• For context, TOPAZ-1 control-to-durvalumab gain was 11.5 to 12.8 months (HR 0.80).
• First phase 3 to beat the TOPAZ-1 regimen (durvalumab + GemCis).

What is not yet known

• Full papers for CLARITY-Gastric01 and HARMONi-GI1, final OS, and whether SAFFRON and HARMONi-GI1 results generalise beyond mainly Asian/Chinese populations.
• Regulatory status — divarasib, savolitinib, sonesitatug vedotin and ivonescimab are not approved for these indications in India.

Sources

1. Skoulidis F, et al. Divarasib as Compared with Sotorasib or Adagrasib in KRAS G12C–Mutated NSCLC (Krascendo 1). N Engl J Med, published online 25 Oct 2026
2. Roche press release — Krascendo 1 results, ESMO 2026 LBA5 (25 Oct 2026)
3. Osimertinib–Savolitinib in Advanced NSCLC after Progression during EGFR-TKI Therapy (SAFFRON). N Engl J Med, 2026
4. HUTCHMED announcement (SEC exhibit) — SAFFRON results incl. MET-overexpressed/amplified subsets (Oct 2026)
5. AstraZeneca press release — CLARITY-Gastric01 results, ESMO 2026 LBA7 (25 Oct 2026)
6. OncLive — CLARITY-Gastric01 shows OS benefit with sonesitatug vedotin (subgroups, PFS detail)
7. Akeso press release — HARMONi-GI1 results, ESMO 2026 LBA8 (25 Oct 2026)
8. Summit Therapeutics press release — HARMONi-GI1 OS data table (25 Oct 2026)
9. Targeted Oncology — ESMO 2026 Presidential Symposium II showcases OS-improving LBAs in lung and GI cancers (25 Oct 2026)
10. OncLive — Divarasib improves survival in pretreated KRAS G12C NSCLC (25 Oct 2026)
11. ESMO Congress 2026 — official programme
12. ClinicalTrials.gov — Krascendo 1 (NCT06497556)
13. ClinicalTrials.gov — CLARITY-Gastric 01 (NCT06346392)
14. ClinicalTrials.gov — SAFFRON (NCT05261399)
15. ClinicalTrials.gov — HARMONi-GI1 / AK112-309 (NCT06591520)

Educational summary, not medical advice. Opinions are labelled as such and are the author's own. Approvals and availability in India (CDSCO) may differ from the US/EU.

← Back to blog