Biological subtype discordance after neoadjuvant therapy in early breast cancer: Findings from a 1,500-patient cohort with residual disease.

604 Background: Although biomarker reassessment is standard in recurrent breast cancer, its role in early breast cancer with residual disease (RD) after neoadjuvant therapy (NAT) remains not well established, despite potential implications for adjuvant treatment decisions in the post-NAT setting. Methods: We retrospectively analyzed patients with early breast cancer who underwent surgery after NAT and had RD between January 2019 and May 2025. Only patients with estrogen receptor (ER), progesterone receptor (PR), and HER2 status assessed on both pretreatment biopsy and matched surgical specimens were included. Biological subtype discordance was analyzed across luminal, HER2-positive, and triple-negative breast cancer (TNBC). Multivariable logistic regression was used to identify clinicopathological factors associated with subtype change. Results: Among 1,504 eligible patients with residual disease, biomarker discordance involving ER, PR, and/or HER2 was observed in 502 cases (33.4%). A change in biological subtype occurred in 221 patients (14.7%). Subtype discordance was most frequent in tumors initially classified as HER2-positive (39.3%), compared with TNBC (10.6%) and luminal disease (6.2%) (p<0.001). Accordingly, patients with HER2-positive breast cancer at diagnosis were significantly more likely to experience subtype discordance after NAT compared with luminal tumors (absolute difference 33.1%, p<0.001) and TNBC (absolute difference 28.7%, p<0.001), with a smaller but statistically significant difference also observed between TNBC and luminal disease (absolute difference 4.4%, p=0.023). HER2-positive and TNBC tumors predominantly shifted toward a luminal phenotype, whereas luminal tumors more commonly converted to HER2-positive or TNBC subtypes, indicating substantial post-treatment biological heterogeneity. On multivariable analysis, baseline biological subtype remained the strongest independent predictor of subtype change, with higher odds for HER2-positive (OR = 9.14, p<0.001) and TNBC tumors (OR = 1.71, p=0.029) compared with luminal disease. Lobular histology was associated with a significantly lower likelihood of subtype discordance (OR = 0.50, p=0.040), whereas type of NAT, tumor grade, and Ki-67 were not associated with subtype change. Conclusions: In this large cohort of patients with residual disease after NAT, approximately one in seven experienced a change in biological subtype, with the highest discordance observed in initially HER2-positive tumors. These findings support routine biomarker reassessment in early breast cancer with residual disease to better inform adjuvant treatment decisions and to account for post-treatment biological heterogeneity.