Editor's Note: The 2026 European Society for Medical Oncology (ESMO) Congress was held in Madrid, Spain, from October 23 to 27. At the meeting, several major studies in HER2-low breast cancer reported updated results. In particular, the presentation of large phase III clinical trial data for several novel HER2 ADCs sparked extensive discussion among experts. These findings not only offer new hope to patients but also prompt a closer examination of the relationship between different drug design strategies and clinical efficacy.Against this background, Oncology Frontier invited Professor Jin Yang of the First Affiliated Hospital of Xi’an Jiaotong University for an in-depth commentary on the latest developments in HER2-low breast cancer presented at this year’s ESMO Congress. From ADC design and cross-trial interpretation to differences in safety profiles, the evolution of the future treatment landscape, and directions for early-stage research, Professor Yang provides a detailed analysis and clinically valuable insights.
01
Oncology Frontier: T-DXd was the first ADC to achieve positive phase III results in HER2-low or HER2-ultralow breast cancer, providing important insights for the development of subsequent HER2 ADCs and related clinical research. Various novel HER2 ADCs are now emerging, each with distinct choices in terms of linkers, payloads, and drug-to-antibody ratios (DARs). Could you discuss the design characteristics of different HER2 ADCs currently being developed in breast cancer and share your views on the differences in their molecular design strategies?
Professor Jin Yang: The publication of phase III results for several ADCs in the treatment of HER2-low breast cancer was undoubtedly one of the highlights of this year’s ESMO Congress. Much of the discussion has focused on the differences in clinical outcomes among different HER2 ADCs in HER2-low breast cancer. Since these ADCs differ in their linkers, payloads, and DARs, I would first like to distinguish them according to their characteristics and underlying design rationale.
I. Divergent Design Strategies: Three ADCs, Three Different Answers
The central biological challenge in HER2-low breast cancer is insufficient antigen density. Each tumor cell may have only approximately 10,000–100,000 HER2 receptors on its surface, far fewer than the more than one million receptors typically found in HER2-positive disease. For an ADC to work under these conditions, it must address a common question: given the limited target density, how can a sufficient concentration of cytotoxic payload be delivered into tumor cells and then spread to neighboring antigen-negative cells? The three ADCs offer three different answers.
The strategy behind T-DXd, the first ADC to achieve positive results in HER2-low disease, is “greater payload quantity combined with a potent bystander effect.” Its DAR is as high as 8, with good homogeneity. This means that each successfully internalized antibody molecule carries a number of DXd payload molecules close to the theoretical maximum. DXd itself has high membrane permeability. Once released, it can cross cell membranes and kill neighboring tumor cells with low or even no HER2 expression, producing a strong bystander effect. The rationale is straightforward: when the target is scarce, a higher payload density and greater diffusion capacity can compensate for limited targeting efficiency.
SHR-A1811 takes a different approach: “modifying the payload and moderately reducing the DAR to secure a wider safety window.” SHR-A1811 also uses a trastuzumab antibody and a cleavable tetrapeptide linker, with SHR9265, a topoisomerase I inhibitor, as its payload. The key difference is the introduction of a chiral cyclopropyl group near the payload-linker junction, creating steric hindrance intended to inhibit premature cleavage and payload release in plasma, thereby reducing off-target toxicity. In preclinical studies, SHR-A1811 demonstrated stronger in vitro cytotoxicity and bystander killing at a DAR of 8. However, the final DAR was optimized to 6 rather than 8, which had demonstrated greater antitumor activity. This represents a clear trade-off: sacrificing some potency to improve safety.
TQB2102 adopts a third strategy: “enhancing internalization through dual-epitope targeting to compensate for insufficient antigen density at a DAR of 6.” TQB2102 is a HER2 bispecific ADC. Its antibody component simultaneously targets HER2 ECD2, the pertuzumab-binding domain, and ECD4, the trastuzumab-binding domain. Dual-epitope binding may promote HER2 receptor clustering on the cell surface and theoretically enhance internalization efficiency. Its DAR is approximately 6, and it carries a topoisomerase I inhibitor payload linked through an enzyme-cleavable linker, retaining a bystander effect.
If T-DXd compensates for low target density by carrying more payload, TQB2102 compensates by making it easier for the ADC to enter tumor cells. These two approaches pursue the same goal, but their molecular strategies are fundamentally different.
II. From Study Design to Efficacy: Cross-Trial Interpretation and Potential Causal Insights From Three Phase III Studies
At this year’s ESMO Congress, Chinese investigators presented data from two major randomized phase III studies in patients with HER2-low breast cancer: HORIZON-Breast06 (abstract LBA13) and APEX-Breast01 (abstract LBA27). These findings naturally invite comparisons with the DB-04 and DB-06 studies, prompting us to examine differences in patient characteristics, efficacy, and safety across the trials. However, in my view, these three large phase III randomized controlled studies differ in their drug design, enrolled populations, and the clinical questions they seek to answer. They should therefore not be compared in a simplistic, head-to-head manner.
The T-DXd data come from the DB-04 and DB-06 studies. In particular, a post hoc analysis of DB-06 showed that, among patients with HR-positive/HER2-low or HER2-ultralow disease, median PFS with T-DXd versus capecitabine was 13.3 versus 8.5 months (HR 0.66). Consistent benefits were also observed compared with taxanes. In the intention-to-treat (ITT) population, previously reported median PFS was 13.2 versus 8.1 months.
The unique value of DB-06 is that it is the only phase III study to include patients with HER2-ultralow expression, defined as IHC 0 with membrane staining, and achieve positive results in this population. These findings have also supported the approval of T-DXd in China for HER2-low and HER2-ultralow breast cancer.
At this year’s ESMO Congress, Professor Ying Fan presented the HORIZON-Breast06 study of SHR-A1811 on behalf of Academician Binghe Xu’s team. Among 551 patients with HER2-low breast cancer, median PFS in the overall population was 16.4 versus 6.8 months (HR 0.27). Among patients who had not received chemotherapy in the advanced-disease setting, median PFS was 16.6 versus 6.8 months (HR 0.25). The ORR was 75.5% in the SHR-A1811 group and 41.2% in the control group. It is important to note that approximately 81.5% of patients enrolled in this study had not received chemotherapy for advanced disease, indicating an overall earlier treatment-line population. This substantially contributed to the magnitude of the absolute PFS benefit.
Among the 543 patients enrolled in APEX-Breast01, 84.9% had HR-positive disease, and 91.3% had previously received CDK4/6 inhibitors. The results showed that median PFS was 16.4 versus 5.5 months in both the HR-positive/HER2-low subgroup and the overall population, with an HR of 0.29 in each population. The ORR was 68.8% versus 31.0%. Although this study also enrolled patients who had not received chemotherapy in the advanced-disease setting, the proportion previously treated with CDK4/6 inhibitors was high, indicating a population with more prominent features of endocrine resistance.
Potential causal insights linking study design to efficacy can be identified at several levels.
In DB-06, T-DXd was evaluated in patients with HER2-ultralow expression, whose antigen density was even lower. This may have contributed to a lower absolute median PFS in the overall population. Nevertheless, the HR remained stable at 0.62–0.66, suggesting that the “high DAR plus potent bystander effect” strategy maintained reliable relative efficacy even at lower antigen densities.
SHR-A1811 and TQB2102 reported longer absolute median PFS values in HER2-low populations. However, their studies excluded patients with HER2-ultralow expression and included a higher proportion of patients who had not received chemotherapy. Therefore, the higher absolute PFS values should not be interpreted as evidence of superior potency.
What is particularly noteworthy is that TQB2102 achieved PFS and HR values almost identical to those of SHR-A1811 in an HR-positive population with a very high proportion of patients previously treated with CDK4/6 inhibitors. This suggests that the strategy of enhancing internalization through dual-epitope targeting may indeed translate into measurable clinical benefit under conditions of limited antigen density.
Regarding OS, none of the three studies has yet reported mature results. At present, DB-04 is the only study to have reported positive OS results in patients with HER2-low breast cancer.
The exploratory OS HR for SHR-A1811 was 0.67 (95% CI: 0.48–0.94), while the OS HR for TQB2102 was 0.47 in patients with HR-positive/HER2-low disease and 0.44 in the overall population.
T-DXd has demonstrated a more mature OS trend in DB-04. Among patients with HR-positive/HER2-low disease, median OS was 23.9 months in the T-DXd group versus 17.5 months in the treatment of physician’s choice (TPC) group (HR=0.64). In the ITT population, median OS was 23.4 versus 16.8 months, respectively (HR=0.64).
However, it is important to emphasize that these OS findings are derived from early analyses or exploratory endpoints. Whether significant PFS differences will consistently translate into OS benefits requires longer follow-up, particularly in the HR-positive population, which generally has a relatively favorable prognosis.
III. Divergent Safety Profiles: A Direct Reflection of Design Trade-Offs
At this year’s ESMO Congress, international experts also highlighted differences in the toxicity profiles of ADCs targeting HER2 across clinical studies. These differences in safety profiles directly reflect the trade-offs made during drug design. The safety characteristics of the three ADCs can almost be viewed as an inverse reading of their respective design strategies.
For T-DXd, interstitial lung disease (ILD) is the most important design-related safety concern requiring active management. In DB-06, the incidence of ILD/pneumonitis with T-DXd was 11.3%, while the incidence of grade ≥3 drug-related adverse events was 40.6%. The high DAR and potent bystander effect enhance tumor killing at low antigen densities, but may also increase payload exposure in normal tissues, particularly the lungs. This is an inherent cost of the “greater payload quantity” strategy.
The most prominent safety concern with SHR-A1811 is hematologic toxicity. In HORIZON-Breast06, the incidence of grade ≥3 treatment-related adverse events in the SHR-A1811 group reached 84.1%, and approximately half of the patients required dose reductions. This is directly related to the 6.4 mg/kg dose used in the HER2-low phase III study, which was higher than the 4.8 mg/kg dose used in the HER2-positive setting.
To maintain sufficient payload exposure under conditions of low antigen density, the dose was increased, making hematologic toxicity a dose-limiting factor. Although payload modification improved plasma stability and reduced off-target release, it did not eliminate the intrinsic bone marrow toxicity associated with topoisomerase I inhibitors.
Among the three agents, the safety signals currently reported for TQB2102 appear the most favorable, particularly with regard to ILD. In APEX-Breast01, the incidence of treatment-related ILD was only 2.2%, with grade 3 ILD reported in 1.1% of patients.
The dual-epitope design enhances internalization without relying on an extremely high DAR and an exceptionally strong bystander effect to drive efficacy, as T-DXd does. This may provide a structural explanation for its lower ILD signal. However, it should be noted that follow-up in APEX-Breast01 remains relatively short. Since ILD can be a delayed-onset toxicity, its final incidence requires longer observation. Furthermore, current data are insufficient to rule out the possibility that bispecific antibodies may carry distinct immunogenicity or off-target binding risks with long-term use.
IV. Overall Assessment: How Design Differences Shape Clinical Positioning
When considered within the treatment landscape of HER2-low breast cancer, T-DXd currently has the most firmly established position. This is not because it leads in every individual efficacy measure, but because its evidence base is more comprehensive and its approved scope of application is currently unmatched.
DB-06 is the only phase III study to include patients with HER2-ultralow expression, making T-DXd the only ADC with positive phase III evidence and an approved indication in this population. The OS benefit from DB-04—median OS of 23.9 versus 17.5 months in the HR-positive subgroup—also provides the most mature support for its long-term survival value.
The PFS findings for SHR-A1811 and TQB2102 are impressive. However, their evidence has not yet extended to HER2-ultralow disease, their OS data remain preliminary, and further real-world experience is needed to consolidate our understanding of their safety management, particularly hematologic toxicity with SHR-A1811 and long-term ILD outcomes with TQB2102.
From the perspective of the causal relationship between drug design and clinical efficacy, the key message from the 2026 ESMO Congress is not that one ADC has definitively surpassed another in terms of potency. Rather, in HER2-low breast cancer, where antigen density is limited, ADC efficacy is no longer determined by a single structural parameter. It depends on the overall compatibility among antibody targeting strategy, payload characteristics, DAR, and linker stability.
T-DXd achieves a powerful tumor-killing effect, but also carries a pulmonary toxicity risk that requires particular vigilance. SHR-A1811 seeks to achieve a safer payload-release profile, but dose escalation has eroded some of its potential safety advantages. TQB2102 uses enhanced internalization through dual-epitope targeting to reduce its reliance on a high DAR and potent bystander effect, showing a potentially meaningful balance between efficacy and ILD risk in early data.
Whether this balance can be maintained in more mature datasets and broader patient populations will determine whether TQB2102 ultimately represents simply “another treatment option” or becomes a factor capable of reshaping the HER2-low treatment landscape.
02
Oncology Frontier: In light of the latest developments presented at this year’s ESMO Congress, how do you view the future treatment landscape for HER2-low breast cancer? In which directions would you like to see further research and exploration?
Professor Jin Yang: One of the most noteworthy signals from ESMO 2026 is that research into HER2-low breast cancer is beginning to extend from the advanced setting into earlier disease stages. However, the answers provided by early-stage data are far less definitive than those from advanced disease.
I. Differentiation by Treatment Stage: Evidence Is Moving Into Earlier Disease, but “Effective” Does Not Necessarily Mean “Sufficient to Change the Treatment Paradigm”
The first oral presentation in the breast cancer session at this year’s ESMO Congress was the TALENT study (abstract LBA23), a phase II trial evaluating T-DXd as neoadjuvant treatment for early-stage HR-positive/HER2-low breast cancer.
The final results showed a pCR rate of only 7% with T-DXd monotherapy, 3% with T-DXd plus anastrozole, and 9% with sequential T-DXd followed by chemotherapy, despite ORRs reaching 62%–72%.
These findings indicate that, in early-stage HER2-low breast cancer, T-DXd monotherapy or short-course treatment is not yet sufficient to replace neoadjuvant regimens built around chemotherapy. This result is important in defining the boundaries of its clinical application: the substantial PFS benefit demonstrated by ADCs in advanced HER2-low breast cancer cannot simply be extrapolated to imply curative potential in the early-stage setting.
Future investigations in early disease are therefore more likely to focus on rational sequencing or combination strategies involving ADCs, chemotherapy, and endocrine therapy, rather than using ADCs to eliminate chemotherapy altogether.
Meanwhile, both HORIZON-Breast06, evaluating SHR-A1811, and APEX-Breast01, evaluating TQB2102, focused on patients with advanced disease who had not received chemotherapy. Both studies reported median PFS of approximately 16.4 months in the HER2-low population, with relatively similar HRs. Together, these findings point to an emerging trend: ADCs are redefining first-line chemotherapy decisions in advanced HER2-low breast cancer. However, the role of ADCs in early-stage disease still requires a more mature body of evidence.
II. HER2-Ultralow Expression Is Emerging as a Distinct Therapeutic Entity, but Access to Treatment Remains Highly Unequal
One of the key contributions of DB-06 was establishing HER2-ultralow disease—defined as IHC 0 with incomplete, faint membrane staining in ≤10% of invasive cancer cells—as a patient population covered by phase III clinical evidence.
However, the treatment landscape for HER2-ultralow disease remains structurally asymmetric: T-DXd is currently the only drug in this population with positive phase III evidence and an approved indication.
The HORIZON-Breast06 study of SHR-A1811 explicitly enrolled patients with HER2-low expression (IHC 1+ or 2+/ISH−) but did not extend to HER2-ultralow disease. Similarly, APEX-Breast01 evaluated TQB2102 in a HER2-low population.
This means that, in the foreseeable short term, patients with HR-positive/HER2-ultralow breast cancer will effectively have only one ADC treatment option. Although HER2-ultralow expression is a biologically reasonable extension along a continuous spectrum of HER2 expression, it has not yet become a treatment subgroup parallel to HER2-low disease with multiple competing therapeutic options.
III. Treatment Sequencing Is Becoming a More Urgent Clinical Question Than First-Line Selection
In China, as T-DXd becomes more widely used in patients with HER2-low breast cancer, an increasingly important clinical question is whether another anti-HER2 ADC can remain effective after T-DXd.
A systematic review published in 2026 clearly indicated that mechanisms of resistance to T-DXd involve loss of HER2 expression, genomic alterations, DNA repair reprogramming, and remodeling of the tumor microenvironment. For subsequent treatment after T-DXd, switching targets or payloads should be prioritized to reduce the risk of cross-resistance.
This has important implications for the potential positioning of SHR-A1811 and TQB2102, because both drugs, like T-DXd, use topoisomerase I inhibitor payloads. Their differences lie primarily in their linkers, DARs, and antibody designs.
If a patient experiences disease progression during T-DXd treatment, the additional benefit of switching to another HER2 ADC carrying the same class of payload cannot be directly inferred from the phase III data of each drug against chemotherapy. This question of sequential benefit needs to be addressed through specifically designed prospective studies or high-quality real-world evidence.
IV. Standardization of Testing and Special Patient Populations: Two Foundational Issues That Will Determine Clinical Implementation
As we know, HER2-low disease requires accurate testing. However, current immunohistochemistry (IHC) methods were originally optimized to identify HER2 overexpression and have limited ability to distinguish subtle differences within the very low-expression range.
The International Expert Consensus on HER2 Testing and Reporting in Breast Cancer recommends that, for cases originally reported as “HER2 0” without distinguishing between no expression and HER2-ultralow expression, reassessment or retesting should be considered when the patient may benefit from T-DXd based on clinical evaluation.
If testing cannot reliably distinguish true HER2 negativity from HER2-ultralow expression, treatment stratification based on the HER2 expression spectrum will lack a reliable operational foundation. Future improvements in testing should therefore extend beyond standardizing IHC interpretation to include AI-assisted quantitative assessment, immunofluorescence, and supplementary evaluation at the mRNA level, with the aim of reducing dependence on subjective interpretation.
In addition, evidence remains particularly limited for patients with HER2-low breast cancer and brain metastases or leptomeningeal metastases. Prospective data on T-DXd in patients with HER2-low brain metastases are still lacking, while further evidence is also needed for SHR-A1811 and TQB2102.
For patients with HER2-low breast cancer and brain or leptomeningeal metastases, we hope that more evidence-based data will become available for different HER2 ADCs, particularly prospective clinical data, to better guide clinical practice.
Professor Jin Yang
