Association of immunotherapy infusion timing with outcomes in early-stage triple-negative breast cancer.

Abstract

594 Background: Circadian regulation of immune function has been shown to influence the efficacy of immune checkpoint inhibitors (IO) in several malignancies. However, the clinical relevance of IO infusion timing in breast cancer remains unclear. We evaluated the association between IO timing, baseline characteristics, pathologic complete response (pCR), and recurrence outcomes in patients with early-stage triple-negative breast cancer (TNBC). Methods: We conducted a retrospective cohort study of 139 patients with early-stage TNBC treated at Manhattan and Mineola campuses alongside with four sites in Queens, Brooklyn, and Long Island using a KEYNOTE-522-based neoadjuvant regimen including pembrolizumab, chemotherapy, and surgery between July 2021 and June 2025. IO infusion timing was dichotomized using the cohort median time (12:22 PM); patients receiving all first three IO infusions after this cutoff were classified as late. Baseline characteristics including age, body mass index (BMI), race, ECOG performance status, clinical T/N stage, and chemotherapy backbone were compared between groups. pCR was defined as absence of residual invasive disease. Recurrence outcomes included local recurrence, distant recurrence, or death. Multivariable analyses adjusting for baseline clinical factors were performed. Results: Amongst our 139 patients, 108 received early IO and 31 received late IO. Patients receiving early IO were younger (mean age 55.9 vs 63.5, p = 0.025) and had lower BMI (27.4 vs 30.6, p = 0.017), while ECOG performance status 0-1 (early: 98.1% vs late: 90.3% respectively, p = 0.075), non-white race (52.5% vs 56%, p = 0.825), clinical T/N stage (T≥2: 88.7% vs 90.3%, p = 0.999; N≥1: 37.4% vs 38.7%, p = 0.999), carboplatin/paclitaxel chemotherapy backbone distributions (93.5% vs 90.3%, p = 0.693) were similar between groups. pCR occurred in 63% of patients receiving early IO and 45.2% receiving late IO (p = 0.098). Recurrence outcomes occurred in 3.7% of the early IO group and 16.1% of the late IO group (p = 0.026), driven primarily by distant recurrence (0.9% vs 12.9%, p = 0.009). After adjusting for age, BMI, and clinical T/N stage, late IO remained associated with distant recurrence (adjusted odds ratio 16.1; 95% CI, 1.39–186.4; p = 0.026). Conclusions: In this retrospective analysis, early IO infusion timing was associated with significantly lower recurrence risk, particularly distant recurrence. While pCR rates were not statistically significant between early vs late IO, we observed a numerical difference favoring early IO. These findings may suggest that circadian timing of IO may influence long-term outcomes beyond pathologic response. Given the retrospective design and limited number of events, these results are hypothesis-generating and warrant prospective validation.