PREVIEW ASH 2026 — New Orleans, Dec 12–15: the abstracts hematologists are watching, top voices and trial buzz See the ASH 2026 preview →

9 min read

From ASCO24 Hype to FDA CRL Reality: Why Did Summit's First Ivonescimab Submission Fail?

From ASCO24 Hype to FDA CRL Reality: Why Did Summit's First Ivonescimab Submission Fail?

On November 15, 2026, Summit Therapeutics (SMMT) announced that the FDA issued a Complete Response Letter for ivonescimab plus platinum chemotherapy in EGFR-mutant, locally advanced or metastatic non-squamous NSCLC after EGFR TKI therapy. The application was based on the global Phase 3 HARMONi trial. The PDUFA date was November 14, 2026.

Per Summit, the FDA stated that “substantial evidence of effectiveness is not established in the absence of a statistically significant overall survival (OS) improvement.” HARMONi met PFS. OS was a dual primary endpoint and did not reach statistical significance (Summit release).

This post follows ivonescimab in EGFR-mutant NSCLC after EGFR TKI from the HARMONi-A data at ASCO 2024, through the global HARMONi topline in 2025, the WCLC 2025 presentation and Suresh Ramalingam’s discussion, to the CRL, using the slides and physician posts from each point.

ASCO 2024: HARMONi-A

Li Zhang, MD, presented HARMONi-A at ASCO 2024, with simultaneous publication in JAMA on May 31, 2024. HARMONi-A (NCT05184712) is Akeso’s China-only Phase 3 of ivonescimab plus pemetrexed-carboplatin vs placebo plus pemetrexed-carboplatin in non-squamous EGFR-mutant NSCLC after EGFR TKI. Its ClinicalTrials.gov record lists the same protocol number, AK112-301, as the global HARMONi trial (NCT06396065). The primary endpoint was PFS by independent review; OS was a secondary endpoint.

HARMONi-A study design slide, ASCO 2024: ivonescimab or placebo plus pemetrexed and carboplatin; primary endpoint PFS by IRRC; secondary endpoints include overall survival
HARMONi-A study design. Li Zhang, ASCO 2024 (photo shared by Stephen Liu, MD).

PFS: median 7.1 vs 4.8 months, HR 0.46 (95% CI 0.34–0.62), P<0.001, in 322 patients.

HARMONi-A PFS Kaplan-Meier slide, ASCO 2024: median PFS 7.1 vs 4.8 months, HR 0.46 (0.34-0.62), p<0.001
HARMONi-A met its primary endpoint of PFS. Li Zhang, ASCO 2024 (photo shared by Stephen Liu, MD).

OS at 30% data maturity (96 events; data cut-off June 25, 2023): HR 0.72 (95% CI 0.48–1.09). The slide notes that the Chinese regulatory authority requested two OS analyses, at 30% and 52% data maturity. Per Summit, the 52% analysis (data cut-off December 2023) gave HR 0.80 (95% CI 0.59–1.08). Ivonescimab was approved in China for this setting in May 2024 on the basis of HARMONi-A, with PFS as the primary endpoint.

HARMONi-A overall survival slide, ASCO 2024: HR 0.72 (0.48-1.09) after 96 events at 30% data maturity
HARMONi-A OS at 30% data maturity. Li Zhang, ASCO 2024 (photo shared by Stephen Liu, MD).

The day before the presentation, Summit posted the HARMONi-2 topline from China: ivonescimab monotherapy vs pembrolizumab monotherapy in first-line PD-L1-positive NSCLC.

Jack West, MD, then Summit’s Vice President of Clinical Development, quoted that post:

On May 25, 2024, the week before the presentation, physicians asked where the PFS benefit came from.

May 30, 2025: the global HARMONi topline

On May 30, 2025, in time for ASCO 2025, Summit announced topline results from the global HARMONi trial. Physician posts that day cite PFS HR 0.52 and an OS HR of 0.79 (P=0.057) that was not statistically significant. HARMONi itself was not presented at ASCO 2025. Physicians reacted to the PFS result.

September 7, 2025: full data at WCLC 2025

Jonathan Goldman presented HARMONi at the WCLC 2025 Presidential Symposium in Barcelona. The trial randomized 438 patients 1:1 to ivonescimab or placebo plus carboplatin-pemetrexed after progression on a third-generation EGFR TKI. OS and PFS by IRRC were both primary endpoints. The design slide notes that the single-region HARMONi-A study in Asia was positive with PFS as its primary endpoint.

HARMONi Phase 3 study design slide, WCLC 2025: ivonescimab or placebo plus chemotherapy, N=438, primary endpoints OS and PFS by IRRC
HARMONi study design. Goldman, WCLC 2025 (photo shared by Stephen Liu, MD).

PFS by IRRC: median 6.8 vs 4.4 months, HR 0.52 (95% CI 0.41–0.66), P<0.0001. OS at the primary OS analysis (labeled the final OS analysis on the slides): median 16.8 vs 14.0 months, HR 0.79 (95% CI 0.62–1.01), P=0.0570. Median follow-up was 29.7 months overall and 9.2 months for North American and European patients.

HARMONi overall survival Kaplan-Meier slide, WCLC 2025: HR 0.79 (95% CI 0.62-1.01), P=0.0570; NA and EU follow-up not yet mature
Primary endpoint: overall survival. Goldman, WCLC 2025 (photo shared by Chul Kim, MD).

A separate slide updated OS for the full population with longer follow-up of Western patients: HR 0.78 (95% CI 0.62–0.98), nominal P=0.0332. Data cut-offs differed by region: September 2025 for North America and Europe and April 2025 for Asia. Median follow-up was 13.7 months in North America and Europe and 32.7 months in Asia.

HARMONi OS longer-term Western follow-up slide, WCLC 2025: HR 0.78, nominal p=0.0332, regional follow-up and data cut-offs
OS with longer-term Western follow-up, nominal P value. Goldman, WCLC 2025 (photo shared by Stephen Liu, MD).

The discussant: “A New Symphony or a Familiar Tune?”

Suresh Ramalingam, MD, of the Winship Cancer Institute of Emory University, discussed HARMONi.

Suresh Ramalingam presenting his WCLC 2025 discussion titled HARMONi: A New Symphony or a Familiar Tune?
Ramalingam, WCLC 2025 discussion (photo shared by Hidehito Horinouchi, MD).

His key-results slide noted that the OS P value of 0.057 had to be 0.0448 or lower for significance. Grade ≥3 adverse events were 50% vs 42%.

Ramalingam slide HARMONi key results: mPFS 6.8 vs 4.4 months, mOS 16.8 vs 14.0 months, HR 0.79, P=0.057 had to be 0.0448 or lower for significance
HARMONi key results. Ramalingam, WCLC 2025 (photo shared by Estela Rodriguez, MD).

He placed HARMONi next to two other post-TKI trials. MARIPOSA-2 (amivantamab + chemo): PFS 6.3 vs 4.7 months (HR 0.48), OS 17.7 vs 15.3 months (HR 0.73, NS). HERTHENA-Lung02 (patritumab deruxtecan): PFS 5.8 vs 5.4 months (HR 0.77), OS 16 vs 15.9 months (HR 0.98, NS). HARMONi: PFS 6.8 vs 4.4 months (HR 0.52), OS 16.8 vs 14 months (HR 0.79, NS).

Ramalingam slide comparing MARIPOSA-2, HERTHENA-Lung02 and HARMONi PFS and OS in the post-TKI setting
Post-TKI trial comparison. Ramalingam, WCLC 2025 (photo shared by Estela Rodriguez, MD).

His summary slide: the PFS benefit in HARMONi is comparable to studies of VEGF inhibitors in NSCLC; there is no clear evidence of PD-1 blockade as a contributor to efficacy; adverse events associated with both VEGF and PD-1 blockade were observed. Its conclusion: “While Ivonescimab may be effective in other types of NSCLC, in EGFRMT NSCLC its MoA (dual blockade) does not appear to offer an advantage.”

Ramalingam Taken Together slide: PFS benefit comparable to VEGF inhibitors, no clear evidence of PD-1 contribution, dual-blockade adverse events
“Taken together.” Ramalingam, WCLC 2025 (photo shared by Estela Rodriguez, MD).

His final slide asked four questions. Could this represent ‘a’ standard of care? “May be.” Are there any biomarkers to select therapy? “No.” Are there toxicities of concern? “Yes- but manageable.” Is this a new symphony? “Hardly…”

Ramalingam slide Ivonescimab in EGFR-mutant NSCLC: standard of care may be, no biomarkers, manageable toxicities, new symphony hardly
Ramalingam, WCLC 2025 (photo shared by Hidehito Horinouchi, MD).

The primary results were published in The Lancet Oncology in 2026, with the same OS result: HR 0.79, P=0.057.

November 15, 2026: the CRL

Summit’s release says the BLA “cannot be approved in its present form.” Per Summit, the FDA said the PFS benefit without a statistically significant OS benefit does not support a positive benefit-risk profile for approval. The release cites a post hoc, longer-follow-up OS hazard ratio of 0.76 in the full population and in Asian and Western patients. It does not mention resubmitting this application; it says Summit will continue developing ivonescimab in HARMONi-3 and HARMONi-7 (first-line NSCLC), HARMONi-GI3 (colorectal cancer), HARMONi-GU1 (urothelial cancer) and SMT112-216 (renal cell carcinoma). The FDA publishes complete response letters in its openFDA CRL database; this letter was not posted there as of November 15, 2026. Ivonescimab is approved in China and is investigational in the US (Summit release).

Aakash Desai infographic on the FDA CRL for ivonescimab plus chemotherapy: HARMONi design, PFS and OS results, safety
Infographic by Aakash Desai, MD, MPH, posted November 15, 2026.

Physicians on the decision:

The record

  • HARMONi-A had PFS as its only primary endpoint, with OS secondary. The global HARMONi trial had PFS and OS as dual primary endpoints.
  • OS was a dual primary endpoint. At the primary OS analysis, HR was 0.79 (95% CI 0.62–1.01), P=0.057; the significance threshold was P≤0.0448.
  • The later OS updates (HR 0.78, nominal P=0.0332 at WCLC 2025; post hoc HR 0.76 per Summit) added longer follow-up of Western patients, who had less follow-up and a later data cut than the Asian cohort. Neither was a formally tested result.
  • Per Summit, the FDA’s stated reason was the absence of a statistically significant OS improvement.
  • HARMONi-A, Akeso’s China-only Phase 3 under the same protocol number, reported a statistically significant final OS HR of 0.74 (95% CI 0.58–0.95) (Summit, November 2025). The global HARMONi analysis, with 38% of patients from Western countries, did not reach statistical significance on OS.
  • At WCLC 2025, Ramalingam’s discussion found the PFS benefit comparable to prior VEGF-inhibitor studies and no clear evidence of a PD-1 contribution in EGFR-mutant NSCLC.

Full HARMONi KOL coverage, slides and the FDA status: HARMONi trial profile on KOL Pulse.