Objective: To investigate the antitumor activity of different combination regimens to human breast cancer xeno graft (MCF-7) transplanted in nude mice and the effects on the expression of PCNA, and to evaluate the value of PCNA as predictive factor for the response of chemotherapy and individualized treatment. Methods: (1) 88 nude mice models of human breast cancer xenograft (MCF-7) were established, and then were randomly divided into control group and 10 chemotherapy groups (each group, n = 8). Among them, the mice of 5 chemotherapy groups were treated intraperitoneally/orally by 5 combination chemotherapy regimens (CMF, CAF, NP, TP, Xeloda) respectively at 1/3 LD_(10) dosage schedule (dose lethal to 10% of the mice), and that in another 5 chemotherapy groups were treated at 2/3 LD_(10) dosage schedule. Control animals were administered intraperitoneally with normal saline. (2) The body weight of nude mice and transplanted tumor growth were observed and recorded, then inhibition rate of tumor growth was calculated. (3) The pathological features of transplanted tumor were studied under microscope. The expression of proliferating cell nuclear antigen (PCNA) was comparatively studied in chemotherapy group and control group by SP immunohistochemical method and flow cytometry analysis. Results: (1) Body weight, tumor weight and inhibition rate of tumor growth of athymic mice bearing cancer: Body weights and tumor weights of nude mice in every 2/3 LD_(10) chemotherapy group were significantly lower than those of the control group (P < 0.05), and the inhibition rates of tumor growth were 83.1%, 75.5%, 84.6%, 87.9% and 91.0%, respectively. Body weights of athymic mice in every 1/3 LD_(10) chemotherapy group were lower than that of the control (P < 0.05). The results showed that the 2/3 LD_(10) chemotherapy groups could reflect the effect of combination chemotherapy on the nude mice and the clinical dependability was better. So the data of 2/3 LD_(10) chemotherapy groups were appropriated for successive study. (2) Immunohistochemical studies: The expressions of PCNA in every chemotherapy group were significantly lower than that of the control (P < 0.05). Moreover, the expression of PCNA in NP group was significantly lower than those of CMF, CAF, TP and Xeloda groups (P < 0.05), while the expressions of TP and Xeloda groups were significantly lower than those of CMF and CAF groups (P < 0.05). (3) FCM analysis: Fl values of PCNA in every chemotherapy group were significantly lower than that of the control (P < 0.05). Fl values of PCNA in TP and Xeloda groups were significantly lower than those of CMF and CAF groups (P < 0.05), while the value of NP group was significantly lower than that of CMF group (P < 0.05). (4) Relationship between PCNA expression and pathologic response: The expression of PCNA was significantly correlated with pathological therapeutic response of transplanted breast carcinoma (P = 0.001). Conclusion: In vivo chemosensitivity testing with 2/3 LD_(10) dosage combinations in nude mice bearing cancer can reflect the effects of chemotherapeutics and affects of organism exactly. Various chemotherapy regimens all can decrease the expression of PCNA in breast cancer. The PCNA can be regarded as the factor to judge the response to chemotherapy, and it become possibly one of the prospective factors in the selection of chemotherapy regimen and play a rule in individualized therapy in the clinic.