〔Abstract〕 Objective To investigate the effect of18F-deoxythymidine nucleoside(18F-FLT) positron emission computed tomography(PET/CT) imaging to evaluate the radiosensitization effect of Olaparib on breast cancer model in nude mice. Methods According to the random number table method, twenty-four BALB/C nude mice MCF-7 breast cancer models were established and divided into four groups with 6 mice in each group, namely the control group, radiotherapy group, Olaparib group and Olaparib + radiotherapy group.18F-FLT micro PET/CT imaging was performed on nude mice before and 48 h after treatment, respectively. The changes of maximum standardized uptake value(SUVmax), total proliferation volume(TPV) and tumor volume before and after tumor treatment in four groups were compared. The tumors were extracted and weighed to observe the changes of tumor weight, and the expression of Ki-67 and PCNA was analyzed by immunohistochemistry staining. The correlation of tumor SUVmaxwith Ki-67 and PCNA was analyzed. Results Before treatment, there were no significant differences in SUVmax, TPV and tumor volume among the 4 groups(F=0.041, 0.061, 0.045,P>0.05). 48 h after treatment, SUVmaxin the control and Olaparib groups increased significantly(t=-12.111,P<0.001;t=-3.001,P=0.03), SUVmaxwas reduced in the radiotherapy and Olaparib + radiotherapy groups(t=5.829,P<0.01;t=4.448,P<0.01), while SUVmax, TPV and tumor volume of tumors in the Olaparib + radiotherapy group were lower than those in the radiotherapy group(t=3.388, 5.884, 5.990,P<0.01).Tumor weight was significantly lower in the Olaparib + radiotherapy group than in the other three groups(F=44.405,P<0.001). Immunohistochemical staining showed that Ki-67 and PCNA were the least expressed in the Olaparib + radiotherapy group than in the other three groups(F=16.289,39.645,P<0.001). SUVmaxwas positively correlated with Ki-67 and PCNA expression(r=0.920,0.918,P<0.01). Conclusion 18F-FLT Micro PET/CT imaging can evaluate the radiosensitizing effect of Olaparib on nude mouse breast cancer model.