Hippocampal-amygdala subregional characteristic changes predict heterogeneous post-chemotherapy neuropsychological outcomes.

[BACKGROUND] The hippocampus and amygdala are interconnected structures critical for cognition. Existing whole-region analyses show inconsistent findings, while subregional changes and their interactions remain unclear. Therefore, we aim to investigate subregional abnormalities in breast cancer patients undergoing neoadjuvant chemotherapy, focusing on both regional and network-level alterations.

[METHODS] A total of 79 breast cancer (BC) patients and 98 healthy controls (HC) were enrolled in our study. The subregional volumes and network measures in BC patients were analyzed at three time points: baseline (B0), following the first cycle of neoadjuvant chemotherapy (NAC) (B1), and at the completion of therapy (B2). These were then compared with those of the HC group at B0. Partial Least Squares Regression (PLSR) identified subregional predictors of chemotherapy-related cognitive dysfunction.

[RESULTS] Subregional analysis disclosed significant volume disparities in four hippocampal subregions and one amygdala subregion among BC patients across three time points. Compared to B0, 80 % of these regions exhibited volume increases, and 80 % of these differences were statistically significant during the B1. Network analysis demonstrated significant differences in the nodal measures of the Subregion Interconnected Networks in three hippocampal subregions and one amygdalar subregion across the three time points. Notably, all these subregions are located in the right hemisphere, displaying a lateralized distribution pattern. Eleven PLSR models for multidomain cognitive functions demonstrated predictive validity, identifying key subregions as cognitive impairment biomarkers.

[CONCLUSION] Chemotherapy induces distinct hippocampal-amygdalar subregional volumetric changes, which possess predictive significance for chemotherapy-related multidomain cognitive impairment in BC patients.