A Phase II Study of Trastuzumab Deruxtecan (T-DXd) Plus Pyrotinib in HER2-Positive Advanced Breast Cancer After Progression on T-DXd

HER2-positive breast cancer accounts for approximately 15-20% of all breast cancers and is characterized by high aggressiveness and a propensity for relapse and metastasis. The CLEOPATRA study established taxane plus trastuzumab and pertuzumab as the international first-line standard. For patients failing first-line therapy, trastuzumab deruxtecan (T-DXd, DS-8201) has become the globally accepted second-line standard. T-DXd is a next-generation HER2-directed antibody-drug conjugate composed of a humanized anti-HER2 monoclonal antibody, a cleavable tetrapeptide linker, and a topoisomerase I inhibitor payload (DXd), with a drug-to-antibody ratio of approximately 8. Its bystander effect enables activity against neighboring tumor cells with low or absent HER2 expression. In DESTINY-Breast03, T-DXd prolonged median progression-free survival from 6.8 months (T-DM1) to 28.8 months, reducing the risk of progression or death by 67%. Both the Chinese CSCO guidelines and major international guidelines recommend T-DXd as the highest-level second-line option.On August 12, 2026, based on the results of the DESTINY-Breast09 Phase III clinical trial, the indication of T-DXd for the first-line treatment of adult patients with unresectable or metastatic HER2-positive breast cancer was approved in China.Nevertheless, resistance to T-DXd is inevitable, and the population progressing after T-DXd continues to grow. No approved standard therapy and no high-level evidence exist for this setting. Across regimens including other ADCs, chemotherapy, HER2 monoclonal antibodies or TKIs, the ORR after T-DXd progression is approximately 14.5% with a median PFS of only 3-4 months. Real-world data show a median rwPFS of 4.7 months with tucatinib plus trastuzumab plus capecitabine and 2.6 months with sacituzumab govitecan. New clinical evidence is therefore urgently needed.

RATIONALE Mechanisms of T-DXd resistance are multifactorial and include reduced HER2 expression or altered HER2 binding, increased drug efflux (e.g., ABCC1 overexpression) and secondary genomic alterations (ERBB2, NFE2L2, KEAP1, TOP1). We hypothesize that combining T-DXd with pyrotinib may produce synergistic antitumor activity through: (1) complementary mechanisms providing vertical dual blockade of HER2 signaling (extracellular antibody-mediated payload delivery plus intracellular kinase inhibition); (2) suppression of bypass pathway activation such as PI3K/AKT, thereby reversing or circumventing resistance; (3) enhanced ADC cytotoxicity induced by TKI-mediated alteration of intracellular signaling and cell-cycle distribution; and (4) clinical precedent from the TROPHY study, in which T-DXd plus pyrotinib showed a favorable safety profile with the recommended pyrotinib dose established at 320 mg.

STUDY DESIGN This is an investigator-initiated, open-label, single-arm, exploratory phase II study conducted at Fujian Cancer Hospital. Approximately 28 evaluable participants will be enrolled. A Simon's minimax two-stage design is used, with a null hypothesis ORR of 14.5%, an alternative hypothesis ORR of 35%, a one-sided alpha of 0.05 and power of at least 80%. In stage 1, 15 participants will be evaluated; if 2 or fewer responses are observed the study will be stopped for futility, otherwise an additional 13 participants will be enrolled to a total of 28. If more than 7 responses are observed among the 28 participants, the regimen will be considered worthy of further investigation.

INTERVENTION

• Trastuzumab deruxtecan (T-DXd): 5.4 mg/kg intravenous infusion, Day 1 of each 21-day cycle. First infusion over at least 90 minutes; if tolerated, subsequent infusions may be shortened to approximately 30 minutes. Permitted dose reductions: 4.4 mg/kg and 3.2 mg/kg.
• Pyrotinib: 320 mg orally once daily, continuously, administered with or immediately after a meal. Permitted dose reduction: 240 mg once daily.

Treatment continues until radiographic or clinical disease progression, unacceptable toxicity, pregnancy, participant request, loss to follow-up, death, or investigator decision.

ASSESSMENTS Screening (within 28 days before the first dose; laboratory tests and 12-lead ECG within 7 days) includes demographics, medical and treatment history, ECOG performance status, vital signs, physical examination, hematology, serum chemistry, urinalysis, coagulation, thyroid/cardiac evaluation (echocardiography, LVEF), 12-lead ECG, virology (HBV, HCV, HIV), pregnancy testing, tumor markers (CEA, CA153, CA125, CA199) and imaging (CT/MRI of chest, abdomen, and pelvis; bone scan; brain imaging if clinically indicated).

During treatment, hematology, serum chemistry, urinalysis, ECOG assessment, vital signs, physical examination and 12-lead ECG are performed before each cycle; echocardiography, coagulation and urinalysis/stool tests every 2 cycles. Tumor imaging is performed at Cycles 3, 5 and 7 (3-5 days before Day 1) and every 3 cycles starting from Cycle 10 Day 1, with a window of ±7 days.

End-of-treatment visit occurs within 28 days after the last dose. Safety follow-up occurs 28±7 days after the last dose. Efficacy follow-up imaging is performed every 3 months in participants who discontinue for reasons other than progression or death. Survival follow-up is performed every 12 weeks (±7 days) to collect survival status and subsequent anticancer therapy.

STATISTICAL CONSIDERATIONS Efficacy analyses will be based on the Full Analysis Set (FAS), comprising all enrolled participants who receive at least one dose of study treatment (ITT principle); the Per-Protocol Set (PPS) will be used for sensitivity analyses. Safety analyses will be based on the Safety Set (SS), comprising all treated participants. Categorical variables including ORR and DCR will be summarized by frequency and percentage; the 95% CI for ORR will be calculated using the Clopper-Pearson exact method and for other categorical variables using the Wilson method. Time-to-event endpoints (PFS, OS, DoR) will be estimated by the Kaplan-Meier method, with medians and 95% CIs (Brookmeyer-Crowley). Adverse events will be graded according to NCI-CTCAE version 5.0.