Metastasis is the cause of most triple-negative breast cancer deaths, yet anti-metastatic therapeutics remain limited. To develop new therapeutics to prevent metastasis, pathophysiologically relevant assays that recapitulate tumor microenvironment is essential for disease modeling and drug discovery. Here, we have developed a microfluidic metastasis-on-chip assay of the early stages of cancer metastasis integrated with the triple-negative breast cancer cell line (MDA-MB-231), stromal fibroblasts and a perfused microvessel. High-content imaging with automated quantification methods was optimized to assess the tumor cell invasion and intravasation within the model. Cell invasion and intravasation were enhanced when fibroblasts co-cultured with a breast cancer cell line (MDA-MB-231). However, the non-invasive breast cancer cell line, MCF7, remained non-invasive in our model, even in the presence of fibroblasts. High-content screening of a targeted anti-cancer therapy drug library was conducted to evaluate the drug response sensitivity of the optimized model. Through this screening, we identified 30 compounds that reduced the tumor intravasation by 60% compared to controls. Multi-parametric phenotypic analysis was applied by combining the data from the metastasis-on-chip, cell proliferation and 2D cell migration screens, revealing that the drug library was clustered into eight distinct groups with similar drug responses. Notably, MEK inhibitors were enriched in cluster cell invasion and intravasation. In contrast, drugs with molecular targets: ABL, KIT, PDGF, SRC, and VEGFR were enriched in the drug clusters showing a strong effect on tumor cell intravasation with less impact on cell invasion or cell proliferation, of which, Imatinib, a multi-kinase inhibitor targeting BCR-ABL/PDGFR/KIT. Further experimental analysis showed that Imatinib enhanced endothelial barrier stability as measured by trans-endothelial electrical resistance and significantly reduced the trans-endothelial invasion activity of tumor cells. Our findings demonstrate the potential of our metastasis-on-chip assay as a powerful tool for studying cancer metastasis biology, drug discovery aims, and assessing drug responses, offering prospects for personalized anti-metastatic therapies for triple-negative breast cancer patients.