TALENT: Enhertu Did Not Replace Chemotherapy in Preoperative Treatment of HER2-Low Breast Cancer

Could Success in Metastatic Disease Be Carried Over to the Preoperative Setting? TALENT's Answer

The antibody–drug conjugate trastuzumab deruxtecan (T-DXd, Enhertu) is a treatment that prolongs survival in hormone receptor (HR)-positive, HER2-low metastatic breast cancer. The phase 2 TALENT trial investigated whether the same drug could replace preoperative (neoadjuvant) chemotherapy in early-stage disease. According to the results presented at ESMO 2026, T-DXd shrank tumours in two-thirds of patients; however, pathological complete response, meaning that no viable cancer cells remain at surgery, was seen in only 3–9%. Adding hormone therapy did not improve response, and adding chemotherapy contributed little. The most striking finding of the trial was that the amount of HER2 in the tumour fell during treatment, and that complete response was more frequent in patients whose HER2 levels were higher at baseline. We have covered the approval of T-DXd in HER2-low metastatic breast cancer, its extension to HER2-ultralow patients and the highlights of ESMO 2026 in separate articles.

Why does this matter?

About two-thirds of breast cancers are hormone receptor-positive, and a substantial proportion of these are classified as HER2-low. For high-risk tumours, those that are large or have spread to the axillary lymph nodes, preoperative chemotherapy containing an anthracycline and a taxane is given. In HR-positive tumours, however, pathological complete response rates with this treatment are low; chemotherapy also carries a considerable burden, including hair loss, nerve damage and, rarely, permanent heart and bone marrow side effects (breast cancer stages and treatment).

In the DESTINY-Breast04 trial, T-DXd delayed disease progression and prolonged survival compared with chemotherapy in HER2-low metastatic breast cancer. This success raised the question: "Could the same drug replace chemotherapy in early-stage disease?" TALENT is one of the first trials to seek an answer. Although the result is negative, it offers important clues about how, and in whom, the drug might work better.

What did the TALENT trial show?

TALENT gave preoperative T-DXd alone, with anastrozole, or before chemotherapy to 133 patients with HR-positive, HER2-low, stage II–III breast cancer. Tumours shrank in 62–72% of patients; however, the absence of viable cancer cells at surgery (pathological complete response) was seen in only 3–9%. These rates are similar to the level expected with standard chemotherapy in this patient group.

What does HER2-low mean, and how does T-DXd work?

HER2 is a growth receptor on the cell surface. In the pathology report, HER2 staining (immunohistochemistry, IHC) is scored as 0, 1+, 2+ or 3+. Tumours scored 3+, or 2+ with gene amplification, are HER2-positive; tumours scored 1+, or 2+ without gene amplification, are called HER2-low (what is HER2?). Classic HER2 drugs do not work in HER2-low tumours, because the growth of these cells does not depend on HER2.

T-DXd, by contrast, works on a different principle: it is an antibody–drug conjugate that uses HER2 as an "address". The antibody finds even cells with few HER2 receptors, is taken up into the cell and releases the potent chemotherapy molecule it carries (DXd). Because DXd can cross the cell membrane, it can also affect surrounding cancer cells that lack HER2; this is called the "bystander effect" (what is T-DXd?, antibody–drug conjugates).

How was the trial conducted?

TALENT enrolled premenopausal and postmenopausal women, as well as men, with HR-positive, HER2-low, stage II–III operable breast cancer. The first 58 patients were randomised to two arms: arm A received T-DXd alone, and arm B received T-DXd plus the aromatase inhibitor anastrozole (with a GnRH analogue to suppress ovarian or testicular hormones in premenopausal women and in men). Because of concern that a single antibody–drug conjugate might be insufficient for high-risk disease, a third, non-randomised arm was added later: arm C received 4 cycles of T-DXd followed by 4 cycles of docetaxel and cyclophosphamide.

In the final analysis, there were 46 patients in arm A, 29 in arm B and 58 in arm C. Most patients were at high risk: in the majority, Ki-67, which reflects the rate of cell proliferation, was 30% or higher, and 44% had grade 3 tumours (what does Ki-67 mean?). In arm C, which enrolled later, the proportion of postmenopausal patients was lower and the proportion with lymph node involvement was higher. No formal statistical comparison between the arms was planned.

Results: tumours shrink but rarely disappear completely

Tumour shrinkage on imaging (objective response) was high in all three arms: 62% in arm A, 66% in arm B and 72% in arm C. However, the absence of viable cancer cells on microscopic examination of the surgical tissue, that is, pathological complete response, remained low in every arm: 7% with T-DXd alone, 3% with T-DXd plus anastrozole and 9% with T-DXd followed by chemotherapy. When complete or near-complete response (RCB 0–1) was assessed together, the rates were also similar across the three arms (about 16–17%).

Treatment duration may matter

One of the findings highlighted by Dr Sara Hurvitz was treatment duration: none of the patients in arms A and B who received only 6 cycles had a pathological complete response; all complete responses occurred in those who received 8 cycles. Because patient numbers are very small, however, this finding is not definitive but only indicative.

Side effects

Side effects were consistent with the profile of T-DXd known from metastatic disease. The most common were nausea, fatigue and loss of appetite. In arm C, where chemotherapy was added, mouth sores (mucositis) and numbness of the hands and feet (peripheral neuropathy) were, as expected, more frequent; severe side effects of grade 3 or higher and the need for dose reductions were also more common in this arm. Lung inflammation (interstitial lung disease), an important side effect of T-DXd, occurred in three patients, all of grade 2 (the lung side effect of T-DXd). In arm A, one patient experienced loss of life after severe gastrointestinal toxicity and a heart attack. The number of patients who stopped treatment early was 4 in arm A, 4 in arm B and 6 in arm C.

No formal statistical comparison was made between the arms. Objective response rates were reported with small differences in different news sources; the table uses the values from the congress report.

The trial's most interesting finding: HER2 falls, TROP2 does not change

The most valuable aspect of TALENT was the repeated biopsy, blood and imaging samples collected throughout treatment. In the central pathology assessment performed at the University of Southern California, the oestrogen receptor did not change with any treatment; Ki-67, a marker of cell proliferation, fell markedly in all arms, most of all in arm B, which received anastrozole.

At Yale University, HER2 and TROP2, another antibody–drug conjugate target, were measured with a quantitative method (TROPLEX) that is far more sensitive than IHC staining. The two proteins behaved differently: TROP2 levels did not change with treatment, whereas HER2 fell markedly during T-DXd. In arm A, the median HER2 level was 707 amol/mm² at baseline, fell to 334 after the first cycle and to 150 at surgery. In addition, the baseline amount of HER2 was associated with pathological complete response (P = 0.033); Dr Hurvitz stressed, however, that this finding came from a small sample.

These findings raise two practical questions. First, could quantitative measurement of HER2 select the patients who will benefit from T-DXd better than today's IHC scoring? Second, does the preservation of TROP2 while HER2 falls indicate that switching to TROP2-targeted drugs such as datopotamab deruxtecan or sacituzumab after T-DXd might be rational (datopotamab deruxtecan)? This is also an example of how tumours adapt to treatment (how cancer cells develop resistance to treatment).

Expert commentary: similar results, new directions

The discussant, Dr Barbara Pistilli (Gustave Roussy, France), noted that the pathological complete response rates in TALENT were similar to those historically achieved with standard chemotherapy in this patient group. The results are also consistent with the SOLTI-VALENTINE trial, in which preoperative chemotherapy was replaced entirely with another antibody–drug conjugate (the HER3-targeted patritumab deruxtecan): there, too, complete response remained low, and adding letrozole provided no benefit. In the I-SPY trials, complete response was also low in unselected HR-positive tumours in the arms that used datopotamab deruxtecan (with or without durvalumab) and T-DXd combined with a new immune drug.

According to Pistilli, HR-positive tumours rich in immune cells may benefit more from combining an antibody–drug conjugate with immunotherapy rather than with hormone therapy; this combination is worth investigating in biomarker-selected patients (immunotherapy in HR-positive breast cancer, preoperative immunotherapy). Pistilli added that this trial cannot yet answer whether the fall in HER2 results from selection pressure created by the drug or from a temporary decrease.

What do these results mean for patients?

For a patient with HR-positive, HER2-low early breast cancer, the standard approach has not changed today: in high-risk disease, chemotherapy before or after surgery, followed by long-term hormone therapy and, in suitable patients, additional drugs such as CDK4/6 inhibitors (treatment options in HR-positive breast cancer, CDK4/6 inhibitors in early-stage disease). In some low-risk patients, however, genomic tests make it possible to avoid chemotherapy (genomic testing, patients who will not need chemotherapy).

In HR-positive tumours, pathological complete response predicts long-term outcome less well than in HER2-positive or triple-negative tumours. TALENT's real answer will therefore become clear once event-free survival and breast-conserving surgery rates are reported; these analyses, together with circulating tumour DNA (ctDNA), patient-reported outcomes and ovarian function data, are ongoing (ctDNA in breast cancer).

Advice for our readers

• Find out your HER2 score (0, 1+, 2+, 3+) from your pathology report and, if needed, your ISH result; a report stating "HER2-negative" may not mean that you are not HER2-low.
• Preoperative T-DXd is not standard today in HR-positive, HER2-low early breast cancer; consider this treatment only within a clinical trial.
• Base your preoperative treatment plan on a recommendation that has been reviewed by a tumour board.
• If you are receiving T-DXd, report any new cough, shortness of breath or fever to your team immediately; lung inflammation can be treated if it is recognised early.
• If you are premenopausal, discuss the effect of treatment on ovarian function and fertility preservation options before treatment starts (menopause after cancer).
• Have your blood counts monitored during chemotherapy and seek care immediately if you develop a fever (leucopenia and neutropenia).

Frequently asked questions

What is the TALENT trial?

TALENT is a phase 2 trial that gave preoperative trastuzumab deruxtecan (T-DXd) alone, with anastrozole, or before chemotherapy in HR-positive, HER2-low, stage II–III breast cancer. Its results were presented at ESMO 2026: tumours shrank in most patients, but pathological complete response was seen in only 3–9%.

Can Enhertu be used before surgery in HER2-low breast cancer?

Not at present. In HER2-low breast cancer, Enhertu is approved only for metastatic disease. In TALENT, complete response rates with preoperative use were no better than those expected with standard chemotherapy; such use should be considered only within a clinical trial.

What does pathological complete response mean?

It is the absence of viable cancer cells on microscopic examination of the breast tissue and axillary lymph nodes removed after preoperative treatment (RCB 0). It differs from tumour shrinkage on imaging; even if a tumour shrinks, viable cells may remain within it.

What is TALENT's most important finding?

Clinically, T-DXd was not found to be effective enough to replace preoperative chemotherapy. Scientifically, however, the more frequent complete responses in tumours with a high amount of HER2 at baseline, and the fall in HER2 during treatment while TROP2 remained unchanged, have raised new questions for patient selection and sequential treatment in the future.

Glossary

HER2-low

A tumour whose HER2 staining in the pathology report is 1+, or 2+ without gene amplification.

Antibody–drug conjugate

A drug in which a potent chemotherapy molecule is attached to an antibody that recognises a target on the cancer cell.

Neoadjuvant treatment

Drug treatment given before surgery.

Pathological complete response (pCR)

The absence of viable cancer cells on microscopic examination of the surgical tissue.

RCB (residual cancer burden)

A system that grades the amount of cancer remaining after surgery from 0 to 3; RCB 0 is a complete response.

Objective response

Shrinkage of the tumour on imaging beyond a defined threshold.

TROP2

A protein found on the surface of many cancer cells and targeted by some antibody–drug conjugates.

Ki-67

A pathology marker that shows how rapidly cells are proliferating.

Anastrozole

A hormone drug from the aromatase inhibitor group that reduces oestrogen production after menopause.