Part XVII · Research methods, evidence, and emerging therapy · Chapter 96

Emerging therapeutic platforms

The platforms in development, organised by mechanism, with criteria for telling a real target from a publishable one.

1 · Next-generation endocrine agents and targeted degradation

Endocrine resistance has two distinct mechanisms and the new agents address them differently. ESR1 mutations produce ligand-independent receptor activity and are acquired under aromatase inhibitor pressure, as described in Temporal heterogeneity and clonal evolution. Ligand-independent signalling is not addressed by lowering oestrogen, which is why the mutation confers resistance to aromatase inhibition and not to receptor degradation.

Oral selective oestrogen receptor degraders were built for this. Elacestrant established the proof in ESR1-mutant disease¹, and camizestrant, imlunestrant and giredestrant extended the class^2,3. The pattern across the class is consistent. Benefit concentrates in ESR1-mutant tumours and is modest or absent in wild-type disease.

Targeted protein degradation is a different mechanism. Rather than blocking the receptor, a bifunctional molecule recruits an E3 ubiquitin ligase to it and directs it to the proteasome. Vepdegestrant is the first such agent to read out in a phase III breast cancer trial. In VERITAC-2, among 270 patients with ESR1 mutations, median progression-free survival was 5.0 months against 2.1 with fulvestrant, with a hazard ratio of 0.58. Among all 624 randomised patients it was 3.8 against 3.6 months, with a hazard ratio of 0.83 and P equal to 0.07⁴.

That split is worth teaching rather than reporting. The same drug, the same trial, one positive population and one negative overall result. The biomarker did not improve an effect that was there anyway. It identified where the effect exists.

2 · Novel conjugates, payload classes, and new targets

The conjugate field has moved on three axes at once, and keeping them separate is necessary to evaluate any individual agent. The axes are target, payload and linker.

On target, the field has moved beyond HER2 and TROP2. HER3 is the most developed new target, with a biological rationale that predates the conjugate, since HER3 upregulation is a recognised escape mechanism from HER2 and PI3K pathway inhibition^5,6. B7-H3 and LIV-1 are in development. The important lesson from DESTINY-Breast04 and DAISY is that a conjugate's activity does not require the target to be highly expressed, because payload delivered to any expressing cell can kill neighbours through bystander effect^7,8. Target expression therefore behaves less like a receptor requirement and more like a delivery address.

On payload, the shift from tubulin inhibitors to topoisomerase I inhibitors changed both efficacy and toxicity profiles. Newer payload classes under investigation include immune agonists and protein-degrader payloads.

On linker, cleavability determines bystander killing, and stability in circulation determines off-tumour toxicity. Most conjugate toxicity is payload released in the wrong place.

The open question for the class is cross-resistance. A patient who progresses on a topoisomerase I payload conjugate directed at one target may or may not respond to a different target carrying the same payload class. The sequencing evidence available is discussed in Antibody-drug conjugates and targeted delivery, and it is thinner than practice requires.

3 · Bispecific and multispecific antibodies

Bispecific antibodies serve two different purposes that are often conflated.

The first is dual pathway blockade on the tumour cell. Binding two receptors or two epitopes of one receptor can block signalling more completely than either alone, and can force internalisation. Biparatopic HER2 antibodies use this, binding two distinct HER2 epitopes to drive receptor clustering and internalisation, which also increases payload delivery when the antibody carries one. Bispecifics targeting EGFR together with HER3 address the escape biology described in Novel conjugates, payload classes, and new targets.

The second purpose is immune redirection. A T-cell engager binds a tumour antigen with one arm and CD3 with the other, forcing a synapse independent of T-cell receptor specificity. This bypasses antigen presentation entirely, which is the step most reliably lost in the immune evasion described in MHC loss and antigen presentation defects.

Redirection works well in haematological malignancy and has been difficult in solid tumours, for reasons that are specific rather than mysterious. T cells must be present in the tumour, and the excluded and desert phenotypes in Spatial organization and what it adds have few. The suppressive microenvironment inactivates those that arrive. And target antigens expressed on normal tissue produce on-target off-tumour toxicity with no therapeutic window. Breast cancer is largely an immunologically cold disease outside a subset of triple-negative tumours, which makes it a hard setting for this mechanism rather than an obvious one.

4 · Cellular therapies and therapeutic vaccines

Chimeric antigen receptor T cells transformed B-cell malignancy and have not transferred to breast cancer. The obstacles are the same three that limit T-cell engagers, plus one more. There is no breast cancer equivalent of CD19, which is to say no antigen that is uniformly expressed on tumour and dispensable on normal tissue. HER2, TROP2, mesothelin and MUC1 are each expressed somewhere it matters. Trafficking into solid tumour and persistence within a suppressive microenvironment remain unsolved. Results to date in solid tumours are modest.

Tumour-infiltrating lymphocyte therapy has a stronger rationale in breast cancer than CAR-T, because it uses the patient's own tumour-reactive repertoire rather than an engineered specificity, and therefore does not require a shared antigen. It requires resectable tumour with adequate lymphocyte infiltration, which reconnects it to the TIL biology in Tumor-infiltrating lymphocytes, biology and standardized scoring.

Therapeutic vaccines aim to generate or expand tumour-reactive T cells. Personalised neoantigen vaccines identify mutations from the patient's own tumour, predict which generate presentable peptides, and manufacture an mRNA or peptide product. The rationale is strong and the setting matters. Vaccines have their best theoretical chance in the minimal residual disease state rather than in bulky metastatic disease, where the suppressive microenvironment is established. That places them alongside the residual-disease strategies in Emerging treatment strategies. Breast cancer's relatively low tumour mutational burden constrains the approach, since fewer mutations mean fewer candidate neoantigens.

5 · Radiopharmaceuticals and novel delivery systems

Radioligand therapy delivers a radionuclide to a target-expressing cell, killing it and its neighbours within the particle's path length. The approach is established in neuroendocrine and prostate cancer and remains early in breast cancer⁹.

The distinction from conjugates matters mechanistically. A conjugate requires internalisation and intracellular payload release. A radioligand does not, because radiation acts at a distance. Path length determines the scale of the bystander effect. Beta emitters travel millimetres and kill many cells beyond the bound one. Alpha emitters travel micrometres, deposit far more energy, and are correspondingly more precise and more dependent on accurate targeting.

The contribution this class has already made in breast cancer is diagnostic rather than therapeutic, and one agent is not a prospect at all. Fluoroestradiol PET is approved and clinically available with guideline standing, and it belongs in this chapter only because the rest of the class is not. It images oestrogen receptor expression across all sites of disease at once, which is a direct measurement of the between-lesion receptor heterogeneity a single biopsy cannot capture. A meta-analysis pooled 12 studies and 308 participants. Clinical benefit from endocrine therapy followed a positive scan in 66% of patients and a negative scan in 11%, a risk ratio of 3.21 with an interval from 1.96 to 5.25¹⁰. A negative scan in a tumour reported as receptor-positive on pathology is therefore a usable result rather than an artefact to explain away.

HER2 PET does the equivalent for HER2 and has been studied for predicting conjugate benefit, and it remains investigational¹¹. Both address the sampling problem in Part VI more directly than any tissue assay can, and the limitation of the approach is sensitivity for low-level expression rather than concept.

6 · Epigenetic and transcriptional therapeutics

Epigenetic therapy in breast cancer has a clear rationale and a poor track record, and the gap between them is informative.

The rationale is that resistance is frequently non-genetic. A cell can acquire a drug-tolerant state through chromatin remodelling without any mutation, survive treatment, and later resume proliferation. That mechanism is described in Plasticity and non-genetic heterogeneity and is not addressed by any agent targeting a mutated protein. Reversing the state, rather than killing the cell, is a different and appealing strategy.

The track record is that broad epigenetic agents have narrow therapeutic windows. Histone deacetylase inhibitors and DNA methyltransferase inhibitors alter transcription globally, in normal tissue as well as tumour. The dose that reverses the tumour state is close to the dose that disrupts haematopoiesis.

More selective approaches target specific readers, writers and erasers, or the transcriptional machinery directly. CDK7 and CDK9 inhibition interferes with transcriptional elongation and has particular relevance to hormone-receptor-positive disease, where oestrogen receptor signalling depends on cofactor recruitment¹². Selectivity against the closely related cell cycle CDKs is the engineering problem.

The honest current position is that this class has strong mechanistic justification, no established role in breast cancer, and a plausible future as combination partners aimed at preventing the tolerant state rather than treating established resistance.

7 · Metabolic and mitochondrial targeting

Tumour metabolic reprogramming has been recognised for a century and has produced almost no approved therapy, which should inform expectations rather than be ignored.

The central difficulty is that the pathways are shared. Glycolysis, glutaminolysis, fatty acid oxidation and one-carbon metabolism are used by normal proliferating cells, and the therapeutic window is correspondingly narrow. Tumours also switch substrates under pressure, so blocking one pathway shifts flux to another.

Where the rationale is strongest is where a metabolic dependency is created by something else. Fatty acid oxidation supports metastatic colonisation and dormancy survival¹³, which connects to Dormancy, residual disease, and late recurrence. Metabolic adaptation contributes to endocrine and targeted therapy resistance¹⁴. A dependency generated by a treatment is a better target than a constitutive one, because the normal tissue is not under the same pressure.

The metformin experience is the cautionary case for the whole class. Epidemiological association, strong preclinical rationale and a large randomised adjuvant trial that did not improve outcomes¹⁵. The lesson is not that metabolism is unimportant. It is that population-level metabolic association does not establish a tumour-cell-autonomous dependency, and only the latter is a drug target.

8 · Microenvironment-directed therapy

Targeting the stroma rather than the tumour cell is attractive because stromal cells are genomically stable and should not acquire resistance the way tumour cells do. The clinical results have been disappointing and the reason is now reasonably well understood.

Cancer-associated fibroblasts are not one population. Single-cell work identified functionally distinct subsets with opposing roles, including subsets that restrain tumour growth alongside those that promote it^16,17,18. Broad depletion removes both. Several early stromal-depletion strategies in solid tumours accelerated disease, which is consistent with removing a restraining population.

The same logic applies to the extracellular matrix. Matrix stiffness promotes invasion, and the matrix is also a physical barrier to dissemination. Degrading it does both things.

The approaches that remain credible are selective rather than ablative. Target the specific fibroblast subset that drives immune exclusion rather than fibroblasts as a class. Normalise vasculature to improve perfusion and immune infiltration rather than destroying it. Block a specific suppressive signal such as TGF-beta in the tumour compartment rather than systemically. Each requires the ability to identify the relevant subset in a given patient, which is a spatial profiling problem before it is a therapeutic one.

9 · What separates a real target from a publishable one

This section gives criteria rather than opinions, because the distinction is assessable.

A real target satisfies six conditions. Genetic or pharmacological loss of the target impairs tumour growth in a model that contains the relevant compartments, per Experimental models and discovery platforms. The dependency is present in human tumours at a frequency worth developing against, established in clinical samples rather than lines. The target is druggable, meaning a binding site exists or a degradation or delivery strategy substitutes for one. A therapeutic window exists, which requires knowing what the target does in normal tissue, not merely that the tumour needs it. Target engagement is measurable in patients, so that a negative trial can be distinguished from a trial that never hit the target. And a biomarker identifies the patients with the dependency before treatment, not after.

A publishable target satisfies the first condition, sometimes the second, and is silent on the rest. That is not fraud. It is the reward structure of a literature that publishes mechanism and rarely publishes the absence of a therapeutic window.

Two specific tests separate them quickly. Ask whether the target's normal function is known and whether its systemic inhibition is survivable, because a target essential to haematopoiesis or cardiac function will not produce a drug however good the tumour biology. And ask what fraction of patients carry the dependency, because a dependency in 2% of tumours requires a screening strategy before it requires a drug.

References

All 18 references

1. Bidard FC, Kaklamani VG, Neven P, et al. Elacestrant (oral selective estrogen receptor degrader) versus standard endocrine therapy for estrogen receptor-positive, HER2-negative advanced breast cancer: results from the randomized phase III EMERALD trial. J Clin Oncol 2022 40:3246-3256. PMID 35584336
2. Oliveira M, Pominchuk D, Nowecki Z, et al. Camizestrant, a next-generation oral SERD, versus fulvestrant in post-menopausal women with oestrogen receptor-positive, HER2-negative advanced breast cancer (SERENA-2): a multi-dose, open-label, randomised, phase 2 trial. Lancet Oncol 2024 25:1424-1439. PMID 39481395
3. Jhaveri KL, Neven P, Casalnuovo ML, et al. Imlunestrant with or without abemaciclib in advanced breast cancer. N Engl J Med 2025 392:1189-1202. PMID 39660834
4. Campone M, De Laurentiis M, Jhaveri K, et al. Vepdegestrant, a PROTAC estrogen receptor degrader, in advanced breast cancer. N Engl J Med 2025 393:556-568. PMID 40454645
5. Sergina NV, Rausch M, Wang D, et al. Escape from HER-family tyrosine kinase inhibitor therapy by the kinase-inactive HER3. Nature 2007 445:437-441. PMID 17206155
6. Hanker AB, et al. Co-occurring gain-of-function mutations in HER2 and HER3 modulate HER2/HER3 activation, oncogenesis and HER2 inhibitor sensitivity. Cancer Cell 2021. PMID 34171264
7. Modi S, Jacot W, Yamashita T, et al. Trastuzumab deruxtecan in previously treated HER2-low advanced breast cancer. N Engl J Med 2022 387:9-20. PMID 35665782
8. Mosele F, Deluche E, Lusque A, et al. Trastuzumab deruxtecan in metastatic breast cancer with variable HER2 expression: the phase 2 DAISY trial. Nat Med 2023 29:2110-2120. PMID 37488289
9. Abdel-Wahab M, Giammarile F, Carrara M, et al. Radiotherapy and theranostics: a Lancet Oncology Commission. Lancet Oncol 2024 25:e545-e580. PMID 39362232
10. Parihar AS, et al. 18 F-Fluoroestradiol PET/CT for predicting benefit from endocrine therapy in patients with ER-positive breast cancer. J Nucl Med 2025. PMID 40081952
11. Gebhart G, Lamberts LE, Wimana Z, et al. Molecular imaging as a tool to investigate heterogeneity of advanced HER2-positive breast cancer and to predict patient outcome under trastuzumab emtansine (T-DM1): the ZEPHIR trial. Ann Oncol 2016 27:619-624. PMID 26598545
12. Toska E, et al. PI3K pathway regulates ER-dependent transcription in breast cancer through the epigenetic regulator KMT2D. Science 2017. PMID 28336670
13. Gong Y, Ji P, Yang YS, et al. Metabolic-pathway-based subtyping of triple-negative breast cancer reveals potential therapeutic targets. Cell Metab 2020 33:51-64. PMID 33181091
14. Wang YY, Attane C, Milhas D, et al. Mammary adipocytes stimulate breast cancer invasion through metabolic remodeling of tumor cells. JCI Insight 2017 2:e87489. PMID 28239646
15. Goodwin PJ, Chen BE, Gelmon KA, et al. Effect of metformin vs placebo on invasive disease-free survival in patients with breast cancer: the MA.32 randomized clinical trial. JAMA 2022 327:1963-1973. PMID 35608580
16. Costa A, Kieffer Y, Scholer-Dahirel A, et al. Fibroblast heterogeneity and immunosuppressive environment in human breast cancer. Cancer Cell 2018 33:463-479. PMID 29455927
17. Elyada E, Bolisetty M, Laise P, et al. Cross-species single-cell analysis of pancreatic ductal adenocarcinoma reveals antigen-presenting cancer-associated fibroblasts. Cancer Discov 2019 9:1102-1123. PMID 31197017
18. Kieffer Y, Hocine HR, Gentric G, et al. Single-cell analysis reveals fibroblast clusters linked to immunotherapy resistance in cancer. Cancer Discov 2020 10:1330-1351. PMID 32434947