Abstract The dissemination route of brain metastases dictates their spatiotemporal distribution, but whether it alters the ultrastructural state of the blood-tumor barrier remains unresolved. Here, we compared a modified hematogenous dissemination model that minimizes extracranial signal confounds with a direct intracranial inoculation model. The hematogenous model produced multifocal brain lesions, whereas the intracranial model formed unifocal masses. Longitudinal bioluminescence imaging revealed significantly different growth rates between models, yet no statistically significant difference in overall survival was detected under the current cohort sizes and humane endpoints. Critically, exploratory transmission electron microscopy at the tumor-brain interface revealed a conserved pathological phenotype of the blood-tumor barrier in established lesions, irrespective of the seeding route. This pattern was characterized by endothelial cell swelling, a discontinuous basement membrane, and retraction of astrocytic end-feet. These findings suggest that while the seeding route determines the macroscopic pattern of disease, the established brain microenvironment imposes a stereotyped mode of neurovascular unit failure. This work provides a methodologically refined platform for studying brain metastasis and presents direct ultrastructural evidence consistent with a conserved blood-tumor barrier pathology, outlining a testable conceptual framework. These hypothesis-generating ultrastructural observations, derived from representative specimens, warrant quantitative validation.
Zhao et al. (2026) studied this question.