Final results of TALENT: A neoadjuvant trial of T-DXd with or without anastrozole, or in sequence with chemotherapy for HER2-low, HR+ early-stage breast cancer

TALENT

Patient / Population Intervention / Exposition Vergleich Endpunkt

In patients with untreated HR+/HER2-low early breast cancer, neoadjuvant T-DXd alone, with anastrozole or followed by docetaxel plus cyclophosphamide was associated with low rates of pathologic complete response.

N

132Patienten

Design

Phase II, two-stage neoadjuvant study (NCT04553770) in patients with untreated, operable HR+/HER2-low early breast cancer, randomised 1:1 to T-DXd alone (Arm A) or T-DXd with anastrozole (Arm B), with Arm C, T-DXd followed by docetaxel plus cyclophosphamide, added after Arms A and B.

Endpunkt

Residual Cancer Burden (RCB), with safety, overall response rate and biomarkers. The abstract lists these endpoints together and does not name one as primary.

ErgebnisRCB 0/1 (pCR or near-pCR) 16% with T-DXd alone vs 17% with T-DXd plus anastrozole vs 17% (10/58) with T-DXd followed by docetaxel plus cyclophosphamide. RCB0 in evaluable patients 3/42 (7%) vs 1/29 (3%) vs 5/55 (9%). ORR 28/42 (67%) vs 19/29 (66%) vs 37/57 (65%). Grade >3 treatment-emergent AEs in Arm A 3 (20%) with 6 cycles and 8 (27%) with 8 cycles, in Arm B 3 (21%) and 2 (13%), and in Arm C 16 (28%). Interstitial lung disease, all grade 2, in 1 (7%) patient in Arm B with 6 cycles and 2 (3%) in Arm C. HER2 levels were higher with RCB0 than RCB1/2/3: 2691 vs 715-947 amol/mm2 at baseline and 1221 vs 440-608 amol/mm2 at cycle 2.

Hurvitz SA, McAndrew NP, Chan D, et al. Presented at ESMO Congress 2026, abstract LBA23, 23 October. Not yet published.

Diskussion & Kritik

This phase II study shows low pathologic complete response rates with T-DXd-based neoadjuvant therapy in HR+/HER2-low early breast cancer, and interstitial lung disease was rare. Only Arms A and B were randomised. Arm C was added later and had more node-positive disease, so comparisons across arms are not concurrent and are confounded, and treatment changed from 6 to 8 cycles partway through, which splits already small groups. There is no standard chemotherapy comparator, residual cancer burden is a short-term surrogate with no event-free or survival data, and the HER2 associations rest on very few RCB0 cases. The abstract does not describe masking, and the ER, PR and Ki67 analyses are not yet reported.