How Lessons from Breast Cancer Are Shaping ADC Development Across Solid Tumors

Key Takeaways
Breast cancer has generated unusually deep ADC experience across HER2 and TROP-2, helping clarify how target expression, construct design, payload choice, resistance, and sequencing affect clinical performance.
HER2-low and HER2-ultralow disease show that ADCs can make lower levels of antigen expression therapeutically relevant without making target abundance irrelevant.
Resistance can emerge through target loss, altered antibody binding, changes affecting the payload target, or other payload-related mechanisms, complicating ADC sequencing.
ADC development beyond breast cancer increasingly requires tumor-specific biomarker strategies rather than simply transferring breast-cancer thresholds into new indications.
Hepatocellular carcinoma remains an early-stage ADC field, but current GPC3 and B7-H3 programs are already testing lessons around antibody architecture, payload selection, tumor penetration, and target biology.

Breast Cancer as the ADC Proving Ground

Antibody–drug conjugates (ADCs) have become deeply embedded in the treatment of breast cancer, but their importance to the field extends beyond the individual therapies that have reached patients. Clinical experience with multiple ADCs, targets, biomarker-defined populations, and treatment sequences has made breast cancer a proving ground for questions that are increasingly relevant across solid tumors: how much target expression is enough, how tumor heterogeneity influences activity, why ADCs aimed at the same antigen can perform differently, how resistance develops, and what happens when one ADC follows another.

That experience now spans two prominent target families. Human epidermal growth factor receptor 2 (HER2)-directed ADCs include trastuzumab emtansine (T-DM1) and trastuzumab deruxtecan (T-DXd), while TROP-2-directed approaches include sacituzumab govitecan (SG) and datopotamab deruxtecan (Dato-DXd). Together, they have generated a much richer picture of ADC biology than could be drawn from any single target or drug.

As ADC development spreads across other solid tumors, the most valuable export from breast cancer may be a more sophisticated framework for designing and using the entire drug class.

Target Biology Has Become More Complicated Than Positive or Negative

Perhaps the clearest lesson from breast cancer is that the relationship between target expression and ADC activity does not always fit conventional biomarker categories. HER2 provides the strongest example. T-DXd demonstrated benefit in HER2-low metastatic breast cancer, defined in 1 as immunohistochemistry (IHC) 1⁺ or IHC 2⁺ with negative in situ hybridization. The clinically addressable population broadened again when the U.S. Food and Drug Administration (FDA) approved T-DXd in 2025 for hormone receptor (HR)-positive, HER2-low or HER2-ultralow metastatic breast cancer after progression on endocrine therapy, defining HER2-ultralow disease as IHC 0 with membrane staining.^1,2

That expansion does not mean that target abundance has ceased to matter. In the phase II DAISY trial, objective response rates differed across HER2-overexpressing, HER2-low, and HER2-nonexpressing metastatic breast cancers. The investigators concluded that HER2 expression influenced T-DXd efficacy, while also finding evidence that expression alone could not fully account for activity.³

ADCs can therefore make lower levels of antigen expression therapeutically relevant without making antigen abundance irrelevant. The actionable range becomes broader, but the relationship between expression and response remains biologically meaningful.

TROP-2 reinforces the point through a different target. SG demonstrated benefit in metastatic triple-negative breast cancer (TNBC) in the phase III ASCENT trial and later in HR-positive/HER2-negative metastatic breast cancer in TROPiCS-02, where benefit was observed across evaluated TROP-2 expression subgroups.^4,5

For developers moving into other tumors, this experience argues against treating biomarker selection as a matter of carrying an established positive/negative threshold from one cancer into another. The threshold and distribution of target expression may need to be defined for each tumor and each ADC.

The Target Is Only One Part of the Drug

Breast cancer has made clear that sharing a target does not make two ADCs pharmacologically or clinically equivalent.

T-DM1 and T-DXd both recognize HER2, yet the DESTINY-Breast03 phase III trial showed substantially different clinical outcomes between the two therapies.⁶ T-DXd produced a higher objective response rate and substantially reduced the risk of progression or death compared with T-DM1 in previously treated HER2-positive metastatic breast cancer. The trial does not establish which individual design feature produced that difference, but it demonstrates that target identity alone cannot explain ADC performance.

Antibody properties, linker behavior, payload potency and mechanism, and characteristics of the released payload can all influence how effectively a drug reaches and kills tumor cells. In heterogeneous tumors, one relevant variable is bystander activity: depending on the linker and payload, drug released from an antigen-positive cell may reach neighboring cells with less or no target expression. The extent and clinical importance of this effect vary among ADCs and remain under investigation.⁷

Breast cancer now offers a second same-target ADC family through TROP-2. SG and Dato-DXd use the same delivery target but are distinct constructs, and the phase III TROPION-Breast01 trial demonstrated improved progression-free survival with Dato-DXd compared with chemotherapy in previously treated HR-positive/HER2-negative advanced breast cancer.⁸ Although SG and Dato-DXd have not been compared head to head in the evidence reviewed here, their coexistence reinforces the principle seen with HER2: selecting an antigen does not determine the optimal drug built around it.

As multiple developers pursue the same targets across tumor types, the relevant question is not only whether an antigen is attractive but which ADC configuration best fits the biology of the disease in which it will be used.

Resistance Can Arise at Multiple Levels

The depth of ADC use in breast cancer has begun to expose how tumors adapt under treatment pressure, complicating any model of resistance based solely on loss of the surface antigen.

Changes in target expression clearly can matter. A 2026 analysis of paired breast cancer samples collected before treatment and after progression on T-DXd found major decreases in HER2 expression in 49% of evaluated cases, including complete loss of HER2 in a subset.⁹ The investigators also identified mutations in ERBB2 affecting the trastuzumab binding interface and experimentally linked these alterations to impaired T-DXd binding and resistance.

Experience with SG shows that resistance can emerge elsewhere in the therapeutic mechanism. In an intensively studied patient with metastatic TNBC, genomic analysis identified separate resistant tumor subclones carrying alterations affecting either TACSTD2, which encodes TROP-2, or TOP1, the molecular target of the SN-38 payload.¹⁰ The findings illustrate that different branches of the same cancer can escape through changes affecting delivery or susceptibility to the cytotoxic agent.

Experimental work has expanded that picture. Breast cancer models resistant to T-DXd and SG showed resistance associated predominantly with payload-related mechanisms, including increased drug-efflux activity, rather than universal antigen loss.¹¹ In those models, ADCs carrying mechanistically different payloads could restore antitumor activity.

ADC resistance can therefore emerge at several points between target recognition and cell killing. A tumor may reduce or alter the antigen, interfere with productive drug binding or processing, or become less susceptible to the payload itself. Distinguishing among those mechanisms may become increasingly important once patients have more than one ADC available.

Sequencing Creates a New ADC Problem

The success of multiple ADCs has introduced a question that was difficult to study when these therapies were isolated later-line options: how does exposure to the first ADC reshape sensitivity to the next?

Breast cancer is beginning to provide real-world evidence. In one multicenter retrospective cohort of 84 patients with HER2-low metastatic breast cancer who received T-DXd and SG sequentially, median time to treatment failure was shorter with the second ADC regardless of the order in which the drugs were given.¹² A separate analysis of 85 patients treated sequentially with the same two agents likewise found that progression-free survival on the second ADC was shorter for most patients, although some experienced prolonged benefit from an ADC administered second.¹³

The issue is especially relevant because T-DXd, SG, and Dato-DXd all use topoisomerase I inhibitor payload strategies. Preclinical evidence showing payload-associated cross-resistance raises the possibility that prior exposure could influence sensitivity to a subsequent ADC even when the surface target changes.¹¹ Switching to a mechanistically different payload restored activity in resistant experimental models, but payload switching has not been established as a clinical sequencing rule.

Future sequencing strategies may therefore need to account for the resistance pressure created by the first ADC and whether the next therapy changes the mechanism to which resistance emerged.

As additional tumors accumulate multiple ADC options, prospective trials will need to address that question directly rather than relying primarily on retrospective treatment sequences. Breast cancer is reaching that stage first.

Testing the Lessons in Other Solid Tumors

T-DXd shows how an ADC platform can move across histologies. In the phase II DESTINY-Gastric01 trial, T-DXd produced higher response rates and longer overall survival than physician's-choice chemotherapy in previously treated HER2-positive gastric cancer.¹⁴ Another phase II trial, DESTINY-PanTumor02, subsequently evaluated T-DXd across HER2-expressing endometrial, cervical, ovarian, bladder, biliary tract, pancreatic, and other solid tumors.¹⁵ Responses occurred across tumor cohorts, with the strongest activity observed among tumors with centrally confirmed HER2 IHC 3⁺ expression.

Those results contributed to the 2024 accelerated FDA approval of T-DXd for previously treated unresectable or metastatic HER2-positive IHC 3⁺ solid tumors without satisfactory alternative treatment options.¹⁶

Lung cancer shows why moving an ADC into another tumor does not necessarily mean importing the same biomarker framework. The phase II DESTINY-Lung02 trial evaluated T-DXd in metastatic non-small cell lung cancer (NSCLC) selected for activating HER2 mutations rather than the HER2-expression categories used in breast cancer.¹⁷ The target remains HER2, but patient selection is based on a different feature of HER2 biology.

Other tumors demonstrate that ADC expansion also extends beyond HER2. Dato-DXd received accelerated FDA approval in 2025 for previously treated EGFR-mutated NSCLC, carrying a TROP-2 ADC strategy into another tumor setting.¹⁸ In the EV-302 phase III trial in urothelial cancer, Nectin-4-directed enfortumab vedotin combined with pembrolizumab significantly improved progression-free and overall survival compared with platinum-based chemotherapy in previously untreated locally advanced or metastatic disease.¹⁹ That result shows an ADC moving beyond later-line use into a central first-line treatment strategy.

These examples point toward a solid-tumor ADC landscape in which targets, biomarker definitions, payloads, and treatment positions vary according to tumor biology, even as developers draw on lessons established in breast cancer.

Hepatobiliary Cancer at Two Different Stages

Hepatobiliary malignancies illustrate how unevenly ADC development is advancing across tumor types.

Biliary tract cancer (BTC) already has clinical evidence from the DESTINY-PanTumor02 program. Among 41 patients with HER2-expressing BTC included in a 2026 subgroup analysis, the investigator-assessed objective response rate was 22.0% overall. Among tumors with centrally confirmed HER2 IHC 3⁺ expression, the reported response rate was 56.3%.²⁰

The small cohort warrants caution, but the expression-dependent pattern echoes experience in breast cancer and the broader PanTumor02 program. Extending ADC treatment into a new disease does not eliminate the importance of understanding how strongly and consistently the target is expressed.

Hepatocellular carcinoma (HCC) remains at a considerably earlier stage. Clinical activity observed in BTC cannot be extrapolated to HCC, which currently has a much less mature ADC evidence base.

HCC as an Early Test Case

HCC offers an opportunity to examine whether questions that breast cancer exposed through years of clinical use can be addressed earlier in development: which target to pursue, which antibody configuration best engages that target in a solid tumor, which payload provides the most useful therapeutic profile, and how target biology interacts with tumor penetration and systemic exposure.

Glypican-3 (GPC3) has emerged as one candidate. A 2026 preclinical study described GPC3 as overexpressed in approximately 70% of HCC cases and evaluated multiple GPC3-directed ADC designs.²¹ The investigators compared three antibodies, including a biparatopic construct, and coupled them to different cytotoxic payloads, including duocarmycin SA, a pyrrolobenzodiazepine dimer, and deruxtecan.

The comparison went well beyond asking whether GPC3 could support targeted delivery. In the study's in vivo models, a moderate-affinity GPC3 antibody carrying deruxtecan outperformed a higher-affinity biparatopic deruxtecan construct. The authors discussed tumor penetration, exposure, and target-mediated drug disposition among the factors that could contribute to that difference.

The finding is preclinical, but it illustrates the type of optimization required once target identification gives way to drug design. Higher-affinity binding is not necessarily sufficient if it changes how the ADC distributes through the tumor or behaves systemically. An effective construct must balance target engagement with the pharmacologic demands imposed by the tumor environment.

Payload selection adds another layer. The same research program compared duocarmycin SA, pyrrolobenzodiazepine dimer, and deruxtecan payloads, with the resulting constructs showing different efficacy and therapeutic-window characteristics. HCC development is already treating payload choice as a central design variable, the same component of ADC biology that has become increasingly important as breast cancer confronts resistance and sequencing.

GPC3 is not the only target under consideration. Ifinatamab deruxtecan (I-DXd), a B7-H3-directed ADC, has demonstrated clinical activity across advanced solid tumors in the phase I/II IDeate-PanTumor01 study.²² HCC was not among the tumor types providing the published efficacy data from that study, but the ongoing IDeate-PanTumor02 program includes a dedicated HCC cohort and an HCC safety run-in.²³ No HCC-specific efficacy results were available in the verified trial record.

HCC is therefore not yet another established ADC success story. It offers an opportunity to apply lessons from more mature indications before a crowded treatment landscape develops, with target expression, antibody architecture, payload choice, penetration, and eventual resistance considered earlier in development.

Questions That Will Shape the Next Generation of ADCs

Future biomarker strategies will need to establish thresholds appropriate to the specific ADC and disease rather than assume that a definition developed in one setting will transfer unchanged to another. The breast cancer and pan-tumor experience already indicates that broader targetability and continued dependence on expression level can coexist.

Sequencing studies will also need to capture more than the identity of the previous target. As multiple ADCs become available within the same disease, prospective trials may need to account for prior payload mechanism and, where feasible, characterize how the tumor changed at progression. The retrospective breast cancer data and experimental resistance studies provide a rationale for testing those questions directly.

From Successful ADCs to Better ADC Strategy

Breast cancer has done more for the ADC field than demonstrate that targeted delivery of cytotoxic agents can produce substantial clinical benefit. Its increasingly crowded ADC landscape has revealed how target expression, construct design, payload biology, resistance, and treatment sequence can influence one another.

Other solid tumors are encountering those questions at different stages. Some already have clinical evidence supporting ADCs across new histologies, biomarkers, and treatment lines. HCC remains earlier, but current GPC3 and B7-H3 programs show that developers are already confronting design decisions that became increasingly visible as ADC use matured in breast cancer.

As the class expands, developers will need to determine not simply whether a tumor expresses a promising antigen, but which construct can exploit that biology, which payload remains effective in the relevant disease context, how resistance is likely to arise, and how one ADC will affect the options that follow it. The experience accumulated in breast cancer now provides a framework for asking those questions earlier as ADC development spreads across solid tumors.