1.Establishment and verification of risk prediction model of acute exacerbation of chronic obstructive pulmonary disease based on regression analysis
Minghang WANG ; Kunkun CAI ; Dingli SHI ; Xinmin TU ; Huanhuan ZHAO ; Suyun LI ; Jiansheng LI
Chinese Critical Care Medicine 2021;33(1):64-68
Objective:To establish a risk prediction model for acute exacerbation of chronic obstructive pulmonary disease (AECOPD) using regression analysis and verify the model.Methods:The risk factors and acute exacerbation of 1 326 patients with chronic obstructive pulmonary disease (COPD) who entered the stable phase and followed up for 6 months in the four completed multi-center large-sample randomized controlled trials were retrospectively analyzed. Using the conversion-random number generator, about 80% of the 1 326 cases were randomly selected as the model group ( n = 1 074), and about 20% were the verification group ( n = 252). The data from the model group were selected, and Logistic regression analysis was used to screen independent risk factors for AECOPD, and an AECOPD risk prediction model was established; the model group and validation group data were substituted into the model, respectively, and the receiver operating characteristic (ROC) curve was drawn to verify the effectiveness of the risk prediction model in predicting AECOPD. Results:There were no statistically significant differences in general information (gender, smoking status, comorbidities, education level, etc.), body mass index (BMI) classification, lung function [forced expiratory volume in 1 second (FEV1), forced vital capacity (FVC), etc.], disease status (the number and duration of acute exacerbation in the past year, duration of disease, etc.), quality of life scale [COPD assessment test (CAT), etc.] and clinical symptoms (cough, chest tightness, etc.) between the model group and the validation group. It showed that the two sets of data had good homogeneity, and the cases in the validation group could be used to verify the effectiveness of the risk prediction model established through the model group data to predict AECOPD. Logistic regression analysis showed that gender [odds ratio ( OR) = 1.679, 95% confidence interval (95% CI) was 1.221-2.308, P = 0.001], BMI classification ( OR = 0.576, 95% CI was 0.331-1.000, P = 0.050), FEV1 ( OR = 0.551, 95% CI was 0.352-0.863, P = 0.009), number of acute exacerbation ( OR = 1.344, 95% CI was 1.245-1.451, P = 0.000) and duration of acute exacerbation ( OR = 1.018, 95% CI was 1.002-1.034, P = 0.024) were independent risk factors for AECOPD. A risk prediction model for AECOPD was constructed based on the results of regression analysis: probability of acute exacerbation ( P) = 1/(1+ e- x), x = -3.274 + 0.518×gender-0.552×BMI classification + 0.296×number of acute exacerbation + 0.018×duration of acute exacerbation-0.596×FEV1. The ROC curve analysis verified that the area under ROC curve (AUC) of the model group was 0.740, the AUC of the verification group was 0.688; the maximum Youden index of the model was 0.371, the corresponding best cut-off value of prediction probability was 0.197, the sensitivity was 80.1%, and the specificity was 57.0%. Conclusion:The AECOPD risk prediction model based on the regression analysis method had a moderate predictive power for the acute exacerbation risk of COPD patients, and could assist clinical diagnosis and treatment decision in a certain degree.
2.Construction and verification of the risk prediction model for acute exacerbation within 6 months in patients with chronic obstructive pulmonary disease: a secondary analysis based on previous research data
Minghang WANG ; Kunkun CAI ; Dingli SHI ; Lichan BI ; Jiansheng LI
Chinese Critical Care Medicine 2022;34(4):373-377
Objective:To construct the risk prediction model of acute exacerbation of chronic obstructive pulmonary disease (AECOPD) and verify its effectiveness based on deep learning and back propagation algorithm neural network (BP neural network).Methods:Based on the relevant data of 1 326 patients with chronic obstructive pulmonary disease (COPD) in the team's previous clinical study, the acute exacerbation, and its risk factors during the stable period and 6 months of follow-up were recorded and analyzed. Combined with previous clinical research data and expert questionnaire results, the independent risk factors of AECOPD after screening and optimization by multivariate Logistic regression including gender, body mass index (BMI) classification, number of acute exacerbation, duration of acute exacerbation and forced expiratory volume in one second (FEV1) were used to build the BP neural network by Python 3.6 programming language and Tensorflow 1.12 deep learning framework. The patients were randomly selected according to the ratio of 4∶1 to generate the training group and the test group, of which, the training group had 1 061 sample data while the test group had 265 pieces of sample data. The training group was used to establish the prediction model of neural network, and the test group was used for back-substitution test. When using the training group data to construct the neural network model, the training group was randomly divided into training set and verification set according to the ratio of 4∶1. There were 849 training samples in the training set and 212 verification samples in the verification set. The optimal model was screened by adjusting the parameters of the neural network and combining the area under the receiver operator characteristic curve (AUC), and the sample data of the test group was substituted into the model for verification.Results:The independent risk factors including gender, BMI classification, number of acute exacerbation, duration of acute exacerbation and FEV1 were collected from the team's previous clinical research, and the AECOPD risk prediction model was constructed based on deep learning and BP neural network. After 10 000 training sessions, the accuracy of the AECOPD risk prediction model in the validation set of the training group was 83.09%. When the number of training times reached 8 000, the accuracy basically tended to be stable and the prediction ability reached the upper limit. The AECOPD risk prediction model trained for 10 000 times was used to predict the risk of the validation set data, and the receiver operator characteristic curve (ROC curve) analysis showed that the AUC was 0.803. When using this model to predict the risk of the data of the test group, the accuracy rate was 81.69%.Conclusion:The risk prediction model based on deep learning and BP neural network has a medium level of prediction efficiency for acute exacerbation within 6 months in COPD patients, which can evaluate the risk of AECOPD and assist the clinic in making accurate treatment decisions.
3.Osteosarcoma with bone metastasis or pulmonary metastasis show distinct genomic manifestations
Zhenyu CAI ; Yanchun SHE ; Lu XIE ; Han WANG ; Zhiye DU ; Yuan LI ; Tingting REN ; Jie XU ; Xin SUN ; Kunkun SUN ; Danhua SHEN ; Xiaodong TANG ; Wei GUO
Chinese Journal of Orthopaedics 2023;43(9):581-590
Objective:To investigate the genomic manifestation and pathogenesis of osteosarcoma with different relapse pattens, which were respectively initially presented with bone metastasis or pulmonary metastasis.Methods:From May 1, 2021 to October 1, 2021, 38 fresh tumor specimens and some paraffin-embedded specimens of high-grade osteosarcoma were collected in Peking University People's Hospital, including 29 males and 9 females, aged 19.6±2.2 years (range, 6-61 years). Among the 38 cases, 12 cases had initial bone metastasis (group A) and 26 cases had initial lung metastasis (group B), of which 15 cases (40%, 15/38) had paired specimens of primary and metastatic lesions. Based on Illumina NovaSeq 6000, we analyzed whole-exome sequencing (WES) as well as transcriptome for osteosarcoma with paired samples in different relapse patterns. During all their treatment courses, we also collected their paired samples to reveal these tumors' evolution. We sought to redefine disease subclassifications for osteosarcoma based on genetic alterations and correlate these genetic profiles with clinical treatment courses to elucidate potential evolving cladograms.Results:We found that osteosarcoma in group A mainly carried single-nucleotide variations (83%, 10/12), displaying higher tumor mutation burden [4.9 (2.8, 12.0) & 2.4 (1.4, 4.5), P=0.010] and neoantigen load [743.0 (316.5, 1,034.5) & 128.5 (49.0, 200.5), P=0.003], while those in group B mainly exhibit structural variants (58%, 15/26). The mutation spectrum showed that there was a significant difference in age-related gene imprinting 1 between the bone metastasis group and the lung metastasis group ( P=0.005). Samples were randomly selected from group A (3 patients) to investigate immunologic landscape by multiplex immunohistochemistry, from which we noticed tertiary lymphatic structure from one patient from group A. High conservation of reported genetic sequencing over time was found in their evolving cladograms. Conclusion:Osteosarcoma with mainly single-nucleotide variations other than structural variants might exhibit biological behavior predisposing toward bone metastases with older in age as well as better immunogenicity in tumor microenvironment.