iza-bren PANKU-Breast01: the EGFR×HER3 bispecific ADC meets its PFS primary endpoint in HR+/HER2- breast cancer, but the trial was run only in China

Data as of 2026-10-25 (Taipei time). [Fact] statements are sourced; [Inference] statements are this article's judgment.

1. One-line conclusion

[Fact] At PANKU-Breast01's interim analysis, iza-bren versus chemotherapy gave BICR mPFS 8.2 vs 4.0 months (HR 0.41; 95% CI 0.33–0.53; p<0.0001) (BMS). [Inference] Meeting the PFS primary endpoint supports validating the EGFR×HER3 dual-target design in global trials; but the trial was run by Sichuan Biokin only in mainland China, with a chemotherapy control, immature OS and an undisclosed grade ≥3 toxicity rate, so for the FDA it counts only as supportive evidence.

2. What happened

[Fact] BMS and SystImmune announced the prespecified interim analysis on 2026-10-23 (BMS). PANKU-Breast01 (NCT06343948) is a randomized Phase 3 trial comparing iza-bren (n=190) with physician's choice of chemotherapy (n=193) in unresectable locally advanced, recurrent or metastatic HR+/HER2- breast cancer after failure of at least one line of chemotherapy; the primary endpoint was PFS by BICR, with OS as the key secondary endpoint. ClinicalTrials.gov lists sites only in China, with actual enrollment of 383 (ClinicalTrials.gov), consistent with 190+193.

[Fact] Median PFS follow-up was 15.2 months; mPFS 8.2 vs 4.0 months, HR 0.41 (0.33–0.53; p<0.0001); cORR 40.5% vs 20.2%, mDoR 6.9 vs 4.2 months; OS immature. Grade ≥3 TRAEs were mainly hematologic, with the rate not disclosed; discontinuation for TRAEs 1.6%; adjudicated ILD in 2 patients (1.1%), all grade 1–2. According to the BMS press release, the results were presented as an ESMO 2026 late-breaking oral and published simultaneously in The Lancet (the paper's DOI was not checked for this article). Biokin sponsors the trial in mainland China, and SystImmune and BMS co-develop outside China (BMS).

For sponsors: iza-bren's Chinese Phase 3 data are out, and BMS's global trial design will shape control arms and enrollment competition for later-line HR+/HER2- ADC trials; sponsors planning trials in the same population should assess overlap early.

3. Why PFS was met, and what is missing

[Inference] The dual-target design is the sponsor's mechanistic explanation. BMS says iza-bren binds both EGFR and HER3, releases a topo-1 payload after internalization and blocks signaling through both (BMS); this article saw no data linking EGFR/HER3 expression to efficacy, so the explanation cannot yet be verified.

[Inference] The control-arm baseline must be considered too. The chemotherapy arm had mPFS 4.0 months and cORR 20.2% (BMS), a low later-line chemotherapy baseline; this round did not check chemotherapy-arm figures from other ADC trials, so this article does not judge whether HR 0.41 is favorable among similar trials.

[Inference] Safety is the biggest gap. ILD in 2 patients, 1.1%, all low grade, with discontinuation of only 1.6% (BMS) looks good on the surface; but the grade ≥3 TRAE rate was not disclosed, only that it was mainly hematologic (BMS), and the population was solely Chinese, so more data are needed before extrapolating to Western populations.

4. Regulatory and commercial implications

[Fact] iza-bren has FDA Breakthrough Therapy designation (previously treated EGFR-mutant NSCLC) (BMS). [Inference] PANKU-Breast01 enrolled only in China (ClinicalTrials.gov) and used a chemotherapy control, so a US application may need a trial including US patients; Biokin's position for filing in China is more complete.

[Fact] BMS says iza-bren has shown clinical benefit in Phase 3 in five indications, including this trial and TNBC (BMS). [Inference] Later-line HR+/HER2- already has several ADCs; to win a global position, iza-bren needs to prove its value in ADC-pretreated patients.

5. Recommendations for the next trial design and OncoLattice

[Inference] First, design of the global bridging trial. PANKU-Breast01 used a chemotherapy control (BMS); BMS's international trial should instead enroll ADC-pretreated patients or use an ADC as control, or the control will be outdated at launch.

[Inference] Second, biomarkers. BMS emphasizes the EGFR and HER3 dual target (BMS); global trials should prospectively collect expression of both to decide whether patient selection is needed.

[Inference] Third, the role of Asian sites. Multiregional sites in Taiwan, Japan and Korea can accelerate international multiregional trials; OncoLattice can promote Taiwanese sites as "comparable to the Chinese data, yet able to join multiregional trials".

6. Open questions and what to watch

The grade ≥3 TRAE rate, hematologic toxicity details and the proportions of prior CDK4/6 inhibitor and ADC use should appear in The Lancet paper that BMS says was published simultaneously (BMS); this article has not yet obtained the DOI. OS results will read out in later analyses of NCT06343948 (ClinicalTrials.gov). BMS's global breast cancer trial plans have not been announced.

The next thing to watch is the Lancet paper's grade ≥3 toxicity and prior-treatment makeup: they determine whether this result can be extrapolated to Western patients who widely use CDK4/6 inhibitors and ADCs.

Evidence limitations

The sources are the BMS press release (BMS) and the ClinicalTrials.gov record, which lists only Chinese sites (ClinicalTrials.gov).

Advice for readers: sponsors using iza-bren's mPFS of 8.2 months to evaluate their own HR+/HER2- trials should first read the Lancet paper to confirm the prior-treatment makeup; the control arm's mPFS of 4.0 months is chemotherapy data from a Chinese population (BMS) and should not be used directly as the control assumption for a global trial.

Note: differences from the original relay

The relayed figures match. Additions: n=383 (190 vs 193), CIs and p-values, cORR and mDoR; ILD 1.1% means 2 adjudicated grade 1–2 cases.