Clinicopathological evaluation of triple-negative breast cancer treated with keynote-522 regimen

Our objective is to compare the response rate between two matched cohorts of early-stage TNBC (chemotherapy: CT + immune checkpoint inhibitor: ICI vs CT alone) and to define clinicopathological variables to identify a subgroup of patients who could be spared or benefit from ICI.

Patients treated with CT + ICI according to KEYNOTE-522 (KN-522) (n = 128) were included in the study and matched 1:1 with patients treated with CT alone. Matching criteria included age range (10 years), race, clinical stage (c)-AJCC, and histological type (metaplastic [MpBC] vs no special type [IC-NST]). The following histological characteristics in the core needle biopsy (CNB) were included: histological type, Nottingham grade, degree of necrosis, and percentage of tumor-infiltrating lymphocytes (TILs). The residual cancer burden (RCB) was categorized into 0 (pCR), I, II, or III. To identify a subgroup of patients who could be spared or benefit from adding ICI, an analysis of the penalized maximum likelihood estimate was performed.

pCR was achieved in 50% of patients treated with CT + ICI vs 35.9% treated with CT alone (P = .02). In the CT group, lower TILs in CNB, non-Black race, MpBC histology, and a higher degree of necrosis were associated with non-pCR. Patients were categorized into quartiles (Q) based on the predicted risk of non-pCR to CT (Q1 represents the lowest risk and Q4 represents the highest risk of non-pCR). In Q4, the pCR rate with CT alone was only 2.9% (1/46) vs 20% (6/64) in the CT + IT cohort (P = .044). The following variables were associated with Q4 vs Q1-3: non-Black race (96.9% vs 88%, P = .04), MpBC histology (29.7% vs 1.6%, P < .01), lymph node positive disease (54.6% vs 50%, P < .01), Nottingham grade 2 (32.8% vs 6.3%, P < .01), and lower mean TILs (12.8 ± 15 vs 36.7 ± 26.5, P < .01).

The efficacy of CT + ICI regimen in the real world, although higher than CT, is lower than that observed on the KN 522 clinical trial. Our results need validation in a larger cohort.