OBJECTIVE To explore and evaluate the machine learning model for prediction of bacterial bloodstream infections established based on routine test data.METHODS By means of retrospective survey,a total of 5 421 pa-tients who were hospitalized in 3 medical institutions from Jan.2015 to Dec.2022 were recruited as the research subjects,1 914 of whom were assigned as the bloodstream infection group,and 3 507 were assigned as the non-bloodstream infection group.The baseline data including gender and age and the results of routine laboratory tests were collected from the enrolled patients.The 3 types of machine learning algorithms,logistic regression,support vector machine and random forest,were respectively used for the screening of the optimal prediction model;the contribution of feature variables to the predictive capability of the model was interpreted through SHAP.The fea-ture variables of the model were optimized by using recursive feature elimination method,and the predictive effi-ciency of the model was evaluated by the area under the curve(AUC)of receiver operating characteristic(ROC)curves.RESULTS Totally 26 variables involving age,gender and blood routine test indexes were included.The random forest was chosen as the optimal machine learning algorithm for the establishment of prediction model for bloodstream infections,and the accuracy of the model was 0.709,with the AUC 0.706.The result of SHAP ex-planation indicated that the age,hematokrit and erythrocyte volume distribution width-CV had remarkable effect on the model's making right decisions.17 variables of the prediction model showed more remarkable effect than 26 variable on distinguishing from the gram-positive bacteria bloodstream infections from the gram-negative bacteria bloodstream infections,with the AUC 0.715,the sensitivity 0.701,the specificity 0.632.CONCLUSIONS The prediction model that is established based on the blood routine test indexes by machine learning algorithm can pre-dict the bacterial bloodstream infection.Meanwhile,the feature selection strategy can further improve the predic-tive efficiency of the model on basis of lowering the dimensionality.

OBJECTIVE To explore and evaluate the machine learning model for prediction of bacterial bloodstream infections established based on routine test data.METHODS By means of retrospective survey,a total of 5 421 pa-tients who were hospitalized in 3 medical institutions from Jan.2015 to Dec.2022 were recruited as the research subjects,1 914 of whom were assigned as the bloodstream infection group,and 3 507 were assigned as the non-bloodstream infection group.The baseline data including gender and age and the results of routine laboratory tests were collected from the enrolled patients.The 3 types of machine learning algorithms,logistic regression,support vector machine and random forest,were respectively used for the screening of the optimal prediction model;the contribution of feature variables to the predictive capability of the model was interpreted through SHAP.The fea-ture variables of the model were optimized by using recursive feature elimination method,and the predictive effi-ciency of the model was evaluated by the area under the curve(AUC)of receiver operating characteristic(ROC)curves.RESULTS Totally 26 variables involving age,gender and blood routine test indexes were included.The random forest was chosen as the optimal machine learning algorithm for the establishment of prediction model for bloodstream infections,and the accuracy of the model was 0.709,with the AUC 0.706.The result of SHAP ex-planation indicated that the age,hematokrit and erythrocyte volume distribution width-CV had remarkable effect on the model's making right decisions.17 variables of the prediction model showed more remarkable effect than 26 variable on distinguishing from the gram-positive bacteria bloodstream infections from the gram-negative bacteria bloodstream infections,with the AUC 0.715,the sensitivity 0.701,the specificity 0.632.CONCLUSIONS The prediction model that is established based on the blood routine test indexes by machine learning algorithm can pre-dict the bacterial bloodstream infection.Meanwhile,the feature selection strategy can further improve the predic-tive efficiency of the model on basis of lowering the dimensionality.

Objective To explore the effects of different dose rates of X-ray under the same dose on cell clonogenic formation in non-small-cell lung cancer cell line A549 in order to provide experimental basis for clinical radiotherapy plan. Methods The A549 cells were cultured at low density and irradiated with X-rays at dose of 4 Gy and selected dose rates of 1, 2, 4 and 6 Gy/min, respectively, from a linear accelerator. The 8th day after irradiation, the cells were fixed and stained with Giemsa solution, and colonies containing at least 50 cells were counted. The plating efficiency and surviving fraction were calculated. Results The clonogenic number in non-irradiated cells was 88.6±4.6. The numbers were significantly reduced in irradiated cells at dose rate 1, 2, 4 and 6 Gy/min (12.3±3.4, 9.0±0.8, 5.6±1.0, 11.5±1.7, respectively) than that in non-irradiated control cells (F=678.799, P<0.05). The plating efficiencies were decreased in irradiated cells, especially in 4 Gy/min irradiated cells, which was lower than that in any of the other three dose rate groups (P< 0.05). Conclusions Though at same radiation dose, cancer cells have different clonogenic formation efficiency when irradiation with X-ray at different dose rates. Thus, treatment with optimal dose rate may improve the radiotherapy efficacy.

Objective To explore the effect of dose rate of X-rays on migration of non-small cell lung cancer (NSCLC) cells and provide the experimental basis for developing radiotherapy scheme. Methods Human NSCLC cell line A549 was cultured and irradiated with X-rays at dose of 6 Gy from a linear accelerator. The dose rates of 1, 2, 4 and 6 Gy/min were selected. Monolayer adherent cells were scratched and photographed at 0 hour and 24 hours under a microscope to measure the scratch width. Results After 24 hours, the scratch width of nonirradiated control cells was (640.7±8.1)μm. The scratch widths of cells were different when cells were irradiated with X-rays of various dose rates. Scratch widths were the largest in cells irradiated at dose rates of 1 Gy/min [(691.4±7.6)μm] and 6 Gy/min [(691.8±12.1)μm]. The scratch width was (666.2±1.3) μm of X-rays at 4 Gy/min, and there were significant differences compared with nonirradiated group (all P< 0.01), which suggested that inhibitory effect of X-rays at dose rates on A549 cell migration was obvious. However, the scratch width of cells irradiated at 2 Gy/min [(643.5 ±6.8) μm] had no difference compared with the control cells (t=-0.336, P=0.742). Conclusions The effect of X-rays irradiation on cell migration of human NSCLC cell line A549 is related with irradiated dose rate. The effect of different dose rates on cell migration is significantly different. Selecting appropriate dose rates for irradiation may help to improve the efficacy of radiotherapy.