Short-term preoperative durvalumab for stage I TNBC: First results from the phase II UNLOCK 14 - POP-DURVA trial

UNLOCK 14

Patient / Population Intervention / Exposition Vergleich Endpunkt

In patients with untreated cT1N0 triple-negative breast cancer and stromal TILs of at least 5%, short-course preoperative durvalumab rarely achieved a pathological complete response, although it induced frequent immune remodelling of the tumour microenvironment.

N

71Patienten

Design

Multicentre, single-arm, open-label phase II preoperative trial of short-course durvalumab monotherapy given on days 1 and 15 in patients with untreated cT1N0 triple-negative breast cancer and locally assessed stromal TILs of 5% or more, followed by primary surgery or end-of-treatment biopsy before standard therapy (NCT05215106).

Endpunkt

Pathological complete response, with invasive residual disease at the end-of-treatment biopsy counted as non-pCR.

ErgebnispCR occurred in 1/71 (1.4%; 95% CI 0.1-7.2). Objective response rate was 11.6%. Among 46 patients who underwent surgery, 19 (41.3%) were downstaged to pT0/T1a/T1b. In 63 paired samples, median increases were 10% points for sTILs (IQR 0-30) and 9 points for PD-L1 CPS (IQR 1-20; p<0.0001). TLS emerged in 16/49 (33%) of initially negative tumours and 14/32 (44%) baseline excluded tumours became inflamed. Treatment-related adverse events occurred in 33/71 (46%). Four patients (6%) had grade 2 events and none grade ≥3. Thyroid immune-mediated events occurred in 15%, all grade 1-2.

Ribeiro JM, Arnedos M, Bazan F, et al. Presented at ESMO Congress 2026, abstract LBA26, 24 October. Not yet published.

Diskussion & Kritik

This single-arm phase II study shows that two doses of durvalumab alone in small node-negative TNBC were well tolerated and remodelled the tumour immune microenvironment, but produced a pCR in only one of 71 patients and an objective response rate of 11.6%. The immune changes in sTILs, PD-L1, TLS and spatial phenotype are biological surrogates, and the abstract gives no data linking them to pCR, recurrence or survival. With no comparator, a short window of exposure and a primary endpoint assessed partly by surgery and partly by biopsy, the study cannot show whether this window alters the benefit of later standard therapy. Selection by sTILs of 5% or more did not enrich for response, and the planned multi-omic analyses are not yet reported.