Search

Aim: To develop an automatic tool on screening diabetic retinopathy (DR) from diabetic patients., Methods: We extracted textures from eye fundus images of each diabetes subject using grey level co-occurrence matrix method and trained a Bayesian model based on these textures. The receiver operating characteristic (ROC) curve was used to estimate the sensitivity and specificity of the Bayesian model., Results: A total of 1000 eyes fundus images from diabetic patients in which 298 eyes were diagnosed as DR by two ophthalmologists. The Bayesian model was trained using four extracted textures including contrast, entropy, angular second moment and correlation using a training dataset. The Bayesian model achieved a sensitivity of 0.949 and a specificity of 0.928 in the validation dataset. The area under the ROC curve was 0.938, and the 10-fold cross validation method showed that the average accuracy rate is 93.5%., Conclusion: Textures extracted by grey level co-occurrence can be useful information for DR diagnosis, and a trained Bayesian model based on these textures can be an effective tool for DR screening among diabetic patients.

Objective: To design a blood ordering schedule and explore the influencing factors of blood utilizing and ordering for tumor surgical patients., Methods: For a total of 58 306 tumor surgical patients, 22 643 applications of blood ordering and 7430 person-times of blood utilization from October 2002 to May 2012 were retrospectively analyzed at Cancer Hospital of Chinese Academy of Medical Sciences. Their clinical profiles and test results were analyzed., Results: The operative transfusion rate was 32.81%. According to the operation position and the blood transfusion and preparation data, the surgical blood ordering schedule of tumor patients was established. Patient gender, hemoglobin, hematocrit, platelet, total protein and albumin level test results had significant effect on the transfusion of red blood cells (OR = 0.797, 9.614, 1.949, 0.437,0.444, 2.038, all P < 0.05). Patient gender, hemoglobin, hematocrit, activated partial thromboplastin time, prothrombin time, albumin and total protein level test results had significant effect on the transfusion of plasma (OR = 0.851, 1.367, 1.801, 1.652, 2.922, 2.224, 1.362, all P < 0.05) ., Conclusion: For surgical tumor patients, blood ordering should be based upon the test results of routine blood, blood coagulation and protein level test results to ensure that blood transfusion is both rational and safe.

Objective: To explore the influencing factors of platelet transfusion effects in cancer patients., Methods: A total of 363 cases of cancer patients undergoing platelet transfusion at our hospital from January 1, 2010 to June 30, 2011 were enrolled. Only one transfusion was randomly selected for each patient for single-factor analysis. Then a binary Logistic regression analysis was performed with transfusion effects as a dependent variable and the variables with P ≤ 0.2 as covariates in single-factor analysis., Results: Among them, there were 280 effective and 83 ineffective transfusions with an overall effective rate of 77.1%. The single-factor analysis revealed that the P values were < 0.05 for age, platelet transfusion history, fever and antibiotic. Logistic analysis of multiple variables showed that platelet transfusion history (1-4 times compared to none, OR = 1.969,95%CI: 1.135 - 3.417; ≥ 5 times compared to none, OR = 5.260,95% CI:2.344 - 11.806), antibiotic (OR = 2.020,95%CI: 1.139 - 3.583), fever(OR = 1.789,95%CI: 1.015 - 3.153) and storage days of platelets (OR = 1.559,95%CI: 1.112 - 2.186) were independent risk factors., Conclusions: Influenced by multiple factors, the effects of platelet transfusion may be improved through reducing unreasonable transfusions, using fewer storage days of platelets and avoiding transfusions during fever and antibiotic uses.