Influencing factors of survival and prognosis of patients with bladder urothelial carcinoma after surgical treatment and recommendation of an artificial intelligence algorithm / 现代泌尿外科杂志

【Objective】 To explore the factors influencing the survival and prognosis of patients with bladder urothelial carcinoma (BUC) after surgical treatment, and to establish an artificial intelligence algorithm to predict the effects of different surgical regimens. 【Methods】 BUC patients treated with surgery during Jan.2007 and Jan.2019 in The Second Hospital of Dalian Medical University and Nanfang Hospital of Southern Medical University were enrolled. The complete clinical and follow-up data were collected. Deep neural network (DNN) was used to establish an artificial intelligence algorithm model. A prediction model of survival and prognosis was established, and the influencing factors of survival were explored and ranked by the artificial intelligence algorithm. 【Results】 A total of 832 patients were involved, including 438 (52.64%) treated in The Second Hospital of Dalian Medical University, and 394 (47.36%) treated in Nanfang Hospital of Southern Medical University. Of all cases, 579 (69.6%) were non-muscle invasive bladder cancer, and 253 (30.4%) were muscle invasive bladder cancer. Transurethral resection of bladder tumor was conducted in 539 (64.8%) cases, partial cystectomy in 66 (7.9%) cases, and total cystectomy in 227 (27.3%) cases. The data of patients treated in Second Hospital of Dalian Medical University were used for DNN modeling, and the data of patients treated in Nanfang Hospital of Southern Medical University were used for external verification after modeling. Finally, it was concluded that the factors affecting survival and prognosis were T stage, pathological grade, hypertension or cardiovascular and cerebrovascular disease, hemoglobin, blood calcium, smoking, albumin, lymphocytes, age, ratio of albumin/globulin, operation method, N stage, and creatinine clearance rate in descending order. The model could be used for preoperative prediction. 【Conclusion】 Through DNN modeling and external verification, the influencing factors of postoperative survival can be predicted for patients with bladder cancer, and the surgical effects can also be predicted before operation. The model can provide artificial intelligence algorithm support for the selection of surgical methods and postoperative follow-up plans.

Using machine learning algorithm to predict the risk of post-traumatic stress disorder among firefighters in Changsha / 中南大学学报(医学版)

OBJECTIVES@#Firefighters are prone to suffer from psychological trauma and post-traumatic stress disorder (PTSD) in the workplace, and have a poor prognosis after PTSD. Reliable models for predicting PTSD allow for effective identification and intervention for patients with early PTSD. By collecting the psychological traits, psychological states and work situations of firefighters, this study aims to develop a machine learning algorithm with the aim of effectively and accurately identifying the onset of PTSD in firefighters, as well as detecting some important predictors of PTSD onset.@*METHODS@#This study conducted a cross-sectional survey through convenient sampling of firefighters from 20 fire brigades in Changsha, which were evenly distributed across 6 districts and Changsha County, with a total of 628 firefighters. We used the synthetic minority oversampling technique (SMOTE) to process data sets and used grid search to finish the parameter tuning. The predictive capability of several commonly used machine learning models was compared by 5-fold cross-validation and using the area under the receiver operating characteristic curve (ROC-AUC), accuracy, precision, recall, and F1 score.@*RESULTS@#The random forest model achieved good performance in predicting PTSD with an average AUC score at 0.790. The mean accuracy of the model was 90.1%, with an F1 score of 0.945. The three most important predictors were perseverance, forced thinking, and reflective deep thinking, with weights of 0.165, 0.158, and 0.152, respectively. The next most important predictors were employment time, psychological power, and optimism.@*CONCLUSIONS@#PTSD onset prediction model for Changsha firefighters constructed by random forest has strong predictive ability, and both psychological characteristics and work situation can be used as predictors of PTSD onset risk for firefighters. In the next step of the study, validation using other large datasets is needed to ensure that the predictive models can be used in clinical setting.

Resampling combined with stacking learning for prediction of blood-brain barrier permeability of compounds / 生物医学工程学杂志

It is a significant challenge to improve the blood-brain barrier (BBB) permeability of central nervous system (CNS) drugs in their development. Compared with traditional pharmacokinetic property tests, machine learning techniques have been proven to effectively and cost-effectively predict the BBB permeability of CNS drugs. In this study, we introduce a high-performance BBB permeability prediction model named balanced-stacking-learning based BBB permeability predictor(BSL-B3PP). Firstly, we screen out the feature set that has a strong influence on BBB permeability from the perspective of medicinal chemistry background and machine learning respectively, and summarize the BBB positive(BBB+) quantification intervals. Then, a combination of resampling algorithms and stacking learning(SL) algorithm is used for predicting the BBB permeability of CNS drugs. The BSL-B3PP model is constructed based on a large-scale BBB database (B3DB). Experimental validation shows an area under curve (AUC) of 97.8% and a Matthews correlation coefficient (MCC) of 85.5%. This model demonstrates promising BBB permeability prediction capability, particularly for drugs that cannot penetrate the BBB, which helps reduce CNS drug development costs and accelerate the CNS drug development process.

Construction of blood consumption prediction model for emergency trauma patients based on machine learning algorithm / 中华急诊医学杂志

Objective:To establish a blood consumption prediction model for emergency trauma patients based on machine learning algorithm, so as to guide blood collection and blood supply institutions to prepare for the early blood demand of mass casualties in public emergencies.Methods:A retrospective analysis was conducted on trauma patients in the emergency system database of 12 hospitals in Zhejiang Province from January 2018 to December 2020. Patients with chronic medical history such as hematological diseases and tumors, and transferred from other hospitals after external treatment were excluded. The patients were divided into the transfusion group and non-transfusion group according to whether they received blood transfusion. The differences in demographic and clinical characteristics between the two groups were compared, and the computer learning algorithm (XGBoost) was used to build the blood consumption prediction model and blood consumption volume prediction model of emergency trauma patients.Results:Totally 2025 patients were included in this study, including 1146 patients in the transfusion group and 879 patients in the non-transfusion group. The blood demand of emergency trauma patients mainly occurred within 3 days of admission (60%). The main variables affecting the blood consumption prediction model of emergency trauma patients were shock index, hematocrit, systolic blood pressure, abdominal injury, pelvic injury, ascites and hemoglobin. Compared with the traditional prediction model, XGBoost model had the highest hit rate of 59.0%. The accuracy of blood consumption prediction model was the highest when seven levels of blood volume were adopted, and the deviation fluctuated between [0~1] U. According to the prediction model, the blood consumption prediction formula was∑ nw× c. Conclusions:The preliminarily constructed prediction model of blood transfusion and blood consumption for emergency trauma patients has better performance than the traditional prediction model of blood transfusion, which provides reference for optimizing the decision-making ability of blood demand assessment of hospitals and blood supply institutions under public emergencies.

Research on Self-perception and Active Warning Model of Medical Equipment Operation and Maintenance Status Based on Machine Learning Algorithm / 中国医疗器械杂志

The panoramic perception of medical equipment operation and maintenance status is the basic guarantee for the implementation of smart medical care, the machine learning algorithm-based autonomous perception and active early warning model of medical equipment operation and maintenance status is proposed. Introduce deep learning multi-dimensional perception of medical equipment multi-source heterogeneous fault data training sample characteristics to realize autonomous perception of medical equipment operation and maintenance status, introduce reinforcement learning to realize autonomous decision-making of test sample fault characteristics, and build the active early warning mechanism for medical equipment faults. Taking the equipment department of hospital as the carrier of model effectiveness verification, the effectiveness simulation of the model was carried out, the results show that the model has the advantages of comprehensive fault information perception, strong compatibility of medical equipment, high efficiency of active early warning.

Research status and prospect of deep learning algorithm-based artificial intelligence in assisted diagnosis of diabetic retinopathy / 中华实验眼科杂志

Based on deep learning algorithm, big stride development has been made about artificial intelligence ( AI) technology,both in its basic theory and clinical ophthalmic image analysis. AI can diagnose diabetic retinopathy ( DR) automatically by using color fundus photography. Compared with other ophthalmic diseases, DR assisted diagnosis with AI might be far more advanced technic. Benefited from advantage of fast diagnostic speed,high accuracy and accordingly saved human resources, great potential can be expected in AI-assisted DR screening and grading. However,as a recently developed interdisciplinary technology,deep learning-based AI-aided DR screening system still needs multidisciplinary cooperation and resources sharing to get further development,such as overcoming data standardization, real-world verification and productization issues. Although challenges coexist, AI applied in ophthalmology clinical practice can be realized with technical development and widespread concern of society.

Research on convolutional neural network and its application on medical image / 生物医学工程学杂志

Recent years, convolutional neural network (CNN) is a research hot spot in machine learning and has some application value in computer aided diagnosis. Firstly, this paper briefly introduces the basic principle of CNN. Secondly, it summarizes the improvement on network structure from two dimensions of model and structure optimization. In model structure, it summarizes eleven classical models about CNN in the past 60 years, and introduces its development process according to timeline. In structure optimization, the research progress is summarized from five aspects (input layer, convolution layer, down-sampling layer, full-connected layer and the whole network) of CNN. Thirdly, the learning algorithm is summarized from the optimization algorithm and fusion algorithm. In optimization algorithm, it combs the progress of the algorithm according to optimization purpose. In algorithm fusion, the improvement is summarized from five angles: input layer, convolution layer, down-sampling layer, full-connected layer and output layer. Finally, CNN is mapped into the medical image domain, and it is combined with computer aided diagnosis to explore its application in medical images. It is a good summary for CNN and has positive significance for the development of CNN.

Use of a Machine Learning Algorithm to Predict Individuals with Suicide Ideation in the General Population

OBJECTIVE: In this study, we aimed to develop a model predicting individuals with suicide ideation within a general population using a machine learning algorithm. METHODS: Among 35,116 individuals aged over 19 years from the Korea National Health & Nutrition Examination Survey, we selected 11,628 individuals via random down-sampling. This included 5,814 suicide ideators and the same number of non-suicide ideators. We randomly assigned the subjects to a training set (n=10,466) and a test set (n=1,162). In the training set, a random forest model was trained with 15 features selected with recursive feature elimination via 10-fold cross validation. Subsequently, the fitted model was used to predict suicide ideators in the test set and among the total of 35,116 subjects. All analyses were conducted in R. RESULTS: The prediction model achieved a good performance [area under receiver operating characteristic curve (AUC)=0.85] in the test set and predicted suicide ideators among the total samples with an accuracy of 0.821, sensitivity of 0.836, and specificity of 0.807. CONCLUSION: This study shows the possibility that a machine learning approach can enable screening for suicide risk in the general population. Further work is warranted to increase the accuracy of prediction.

A feasibility study of the automatic cystocele severity grading software for quantitative evaluation of prolapse of bladder posterior wall by transperineal ultrasound / 中华超声影像学杂志

Objective To investigate the feasibility of the automatic cystocele severity grading software for quantitative evaluation of prolapse of bladder posterior wall by transperineal ultrasound . Methods One hundred and seventy transperineal ultrasound video clips were recorded when the female patients performing the Valsalva maneuver and those clips were divided into training group ( 85 cases) and test group ( 85 cases) randomly ,then the ralated structures of the images from the training group offline were marked . Through machine learning algorithm ,the computer had learned and was able to analyzed the marking information ,then the automatic cystocele severity grading software was obtained . And later the software was ran to mark the structures and get the cystocele severity grading in the images from the test group . Meanwhile , the same structures of the same images manually were marked and after an interval of more than two weeks the process were repeated by 3 doctors . Finally the grading results obtained from the software and the measurers of the 3 doctors were compared . Results The intelligent identification and automatic measurement software obtained from the machine learning algorithm was able to identify the related structures . The grading results of each measurer were of good consistency ( κ :0 .72 －0 .78 ;ICC :0 .980－0 .990) . The grading results between different measurers were of good consistency ( κ :0 .65－0 .75 ;ICC :0 .985－0 .992) . The grading results between automatic software and three different measurers were of good consistency ( κ :0 .63－0 .67 ;ICC :0 .967－0 .969 ; r ＝0 .936 ,0 .943 ,0 .936 ,all P ＜0 .01) . Conclusions The automatic cystocele severity grading software is able to identify the related structures in the images and reliable to apply the software in pelvic floor ultrasound .

Classifying syndromes in Chinese medicine using multi-label learning algorithm with relevant features for each label / 中国结合医学杂志

<p><b>OBJECTIVE</b>To develop an effective Chinese Medicine (CM) diagnostic model of coronary heart disease (CHD) and to confifirm the scientifific validity of CM theoretical basis from an algorithmic viewpoint.</p><p><b>METHODS</b>Four types of objective diagnostic data were collected from 835 CHD patients by using a self-developed CM inquiry scale for the diagnosis of heart problems, a tongue diagnosis instrument, a ZBOX-I pulse digital collection instrument, and the sound of an attending acquisition system. These diagnostic data was analyzed and a CM diagnostic model was established using a multi-label learning algorithm (REAL).</p><p><b>RESULTS</b>REAL was employed to establish a Xin (Heart) qi defificiency, Xin yang defificiency, Xin yin defificiency, blood stasis, and phlegm fifive-card CM diagnostic model, which had recognition rates of 80.32%, 89.77%, 84.93%, 85.37%, and 69.90%, respectively.</p><p><b>CONCLUSIONS</b>The multi-label learning method established using four diagnostic models based on mutual information feature selection yielded good recognition results. The characteristic model parameters were selected by maximizing the mutual information for each card type. The four diagnostic methods used to obtain information in CM, i.e., observation, auscultation and olfaction, inquiry, and pulse diagnosis, can be characterized by these parameters, which is consistent with CM theory.</p>