As breast cancer progresses to stage IV, it metastasizes to secondary organs, with a strong propensity for bone colonization. Bone metastasis results in dramatically decreased survival rates and currently lacks a definitive cure. To improve survival rates significantly, there is a need for complex and precise in vitro models that can accurately replicate advanced-stage breast cancer for drug screening purposes. Previously, we established a 3D nanoclay in vitro model of bone metastatic breast cancer using human mesenchymal stem cells in combination with either commercial breast cancer cells (MCF-7 and MDA-MB-231) or patient-derived cells (NT013 and NT023) from the primary breast cancer site. In the present study, the efficacy of the in vitro model to distinguish and differentiate between the severity of metastasis in a total of eight patient-derived cell lines representing various subtypes was evaluated. We also tested the effects of the phytochemically enriched plant extract, Rhodiola crenulata , on eight patient-derived cell lines (NT015, NT017, NT021, NT042, NT045, and NT046, in addition to NT013 and NT023) in bone metastatic (BM) culture. Our results confirmed that the cell lines maintained their subtype-specific characteristics after isolation and formed tumors within the bone microenvironment. Additionally, we assessed the impact of these cell lines on Wnt signaling pathways, identifying which lines upregulate or downregulate Wnt signaling through ET-1 and DKK-1 cytokine levels. Within each subtype, we observed differences in the severity of metastasis between patients. R. crenulata induced cytotoxicity in most patient-derived BM cultures, though NT042 BM cultures showed minimal response. In summary, our study has established a patient-derived bone-metastatic breast cancer model that is well-suited for personalized drug screening aimed at treating late-stage breast cancer. This bone metastatic testbed has the capability to evaluate the severity of metastasis within breast cancer subtypes for individual patients.