ObjectiveTo analyze the relevant risk factors affecting postoperative relapse-free survival (RFS) in the primary gastrointestinal stromal tumors (GIST) and develop a Nomogram predictive model of postoperative RFS for the GIST patients. MethodsThe patients diagnosed with GIST by postoperative pathology from January 2011 to December 2020 at the First Hospital of Lanzhou University and Gansu Provincial People’s Hospital were collected, and then were randomly divided into a training set and a validation set at a ratio of 7∶3 using R software function. The univariate and multivariate Cox regression analysis were used to identify the risk factors affecting the RFS for the GIST patients after surgery, and then based on this, the Nomogram predictive model was constructed to predict the probability of RFS at 3- and 5-year after surgery for the patients with GIST. The effectiveness of the Nomogram was evaluated using the area under the receiver operating characteristic curve (AUC), consistency index (C-index), and calibration curve, and the clinical utility of the Nomogram and the modified National Institutes of Health (M-NIH) classification standard was evaluated using the decision curve analysis (DCA). ResultsA total of 454 patients were included, including 317 in the training set and 137 in the validation set. The results of multivariate Cox regression analysis showed that the tumor location, tumor size, differentiation degree, American Joint Committee onCancer TNM stage, mitotic rate, CD34 expression, treatment method, number of lymph node detection, and targeted drug treatment time were the influencing factors of postoperative RFS for the GIST patients (P<0.05). The Nomogram predictive model was constructed based on the influencing factors. The C-index of the Nomogram in the training set and validation set were 0.731 [95%CI (0.679, 0.783)] and 0.685 [95%CI (0.647, 0.722)], respectively. The AUC (95%CI) of distinguishing the RFS at 3- and 5-year after surgery were 0.764 (0.681, 0.846) and 0.724 (0.661, 0.787) in the training set and 0.749 (0.625, 0.872) and 0.739 (0.647, 0.832) in the validation set, respectively. The calibration curve results showed that a good consistency of the 3-year and 5-year recurrence free survival rates between the predicted results and the actual results in the training set, while which was slightly poor in the validation set. There was a higher net benefit for the 3-year recurrence free survival rate after GIST surgery when the threshold probability range was 0.19 to 0.57. When the threshold probability range was 0.44 to 0.83, there was a higher net benefit for the 5-year recurrence free survival rate after GIST surgery. And within the threshold probability ranges, the net benefit of the Nomogram was better than the M-NIH classification system at the corresponding threshold probability. ConclusionsThe results of this study suggest that the patients with GIST located in the other sites (mainly including the esophagus, duodenum, and retroperitoneum), with tumor size greater than 5 cm, poor or undifferentiated differentiation, mitotic rate lower than 5/50 HPF, negative CD34 expression, ablation treatment, number of lymph nodes detected more than 4, and targeted drug treatment time less than 3 months need to closely pay attentions to the postoperative recurrence. The discrimination and clinical applicability of the Nomogram predictive model are good.
ObjectiveTo develop and validate a Nomogram for predicting severe immune-related adverse events (irAEs) in patients with advanced non-small cell lung cancer (NSCLC) undergoing immunotherapy based on clinical features and inflammatory indicators. MethodsA total of 423 patients with advanced NSCLC treated with immunotherapy between January 2023 and January 2025 at Tianjin Fourth Center Hospital and Tianjin Cancer Hospital Airport Hospital were enrolled. Patients were divided into a severe irAEs group (≥grade 3, n=76) and a non-severe irAEs group (n=347), then randomly allocated into training and validation cohorts (7:3 ratio) . Clinical data, neutrophil-to-lymphocyte ratio (NLR), and interleukin-6/C-reactive protein (IL-6/CRP) levels were collected. Independent risk factors for severe irAEs during immunotherapy in advanced NSCLC were identified through logistic regression analysis, and a nomogram model was constructed accordingly. The discriminative ability, accuracy, and clinical utility of the model were evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). ResultsAmong the 423 included patients [274 males, 149 females, aged 44-78 (60.77±5.91) years], the overall incidence of irAEs was 57.92% (245/423), with severe irAEs occurring in 17.97% (76/423). Multivariate analysis revealed that Eastern Cooperative Oncology Group (ECOG) performance score ≥2, programmed death-ligand 1 (PD-L1) expression [tumor proportion score (TPS) ≥50%], combination therapy regimen, low NLR values, and high IL-6/CRP ratio were independent risk factors for severe irAEs during immunotherapy in advanced NSCLC (P<0.05). The area under the ROC curve (AUC) was 0.948 [95%CI (0.912, 0.985)] in the training cohort and 0.946 [95%CI (0.917, 0.976)] in the validation cohort. Calibration curves and DCA demonstrated good consistency and clinical net benefit of the model. ConclusionThe nomogram integrating clinical features and inflammatory markers effectively predicts the risk of severe irAEs in advanced NSCLC patients receiving immunotherapy, exhibiting excellent discrimination, calibration, and clinical practicality.
ObjectiveTo predict the risk factors affecting postoperative recurrence of granulomatous lobular mastitis (GLM) in the mass stage by machine learning algorithm, and to provide a reference for the early identification and prevention of postoperative recurrence of GLM in the mass stage. MethodsThe electronic medical records and follow-up data of patients with GLM in the Department of Breast Disease Unit, the First Affiliated Hospital of Henan University of Traditional Chinese Medicine from October 2020 to January 2023 were selected. A total of 340 patients with GLM in the mass stage who met the inclusion and exclusion criteria were selected as the research subjects. According to whether the patients relapsed after surgery, they were divided into recurrence group and non-recurrence group. The collected cases were randomly divided into training set and test set according to the ratio of 7:3. In the training set, the recurrence prediction model was constructed by using traditional logistic regression and three machine learning algorithms: artificial neural network, random forest and XGBoost (extrem gradient boosting). In the test set, the performance of the model was evaluated by sensitivity, specificity, accuracy,positive predictive value, negative predictive value, F1 value and area under the curve (AUC) value. The Shapley Additive exPlanation (SHAP) method was used to explore the important variables that affect the optimal model in identifying postoperative recurrence in the GLM mass phase. The optimal risk cutoff value of the prediction model was determined by the Youden index. Based on this, the postoperative patients in the GLM mass phase of the external test set were divided into high-risk and low-risk groups. ResultsA total of 392 patients who met the GLM mass stage were included, and 52 cases were excluded according to the exclusion criteria, and 340 cases were finally included, including 60 cases in the recurrence group and 280 cases in the non-recurrence group. Based on the results of univariate analysis, correlation analysis and clinically meaningful influencing factors, 12 non-zero coefficient characteristic variables were screened for the construction of the prediction model, and these 12 characteristic variables included other disease history, number of miscarriages, breastfeeding duration of the affected breast, history of milk stasis, lesion location, nipple indentation, fluctuation sensation, low-density lipoprotein, testosterone, previous antibiotic therapy, previous oral hormone medication, and perioperative traditional Chinese medicine treatment duration. The logistic regression prediction model, artificial neural network, random forest and XGBoost prediction models were constructed, and the results showed that the accuracy, positive predictive value and negative predictive value of the four prediction models were all >75%, among which the XGBoost model had the best performance, with accuracy, specificity, sensitivity, AUC, positive predictive value, negative predictive value and F1 values of 0.93, 0.99, 0.65, 0.87, 0.92, 0.93 and 0.76, respectively. SHAP method found that the duration of traditional Chinese medicine treatment during perioperative period, the duration of breast-feeding on the affected side, low density lipoprotein, testosterone and previous hormone drugs were the top five factors affecting XGBoost model to identify postoperative recurrence of GLM in mass stage. ConclusionsCompared with the traditional Logistic regression prediction model, the models based on machine learning for identifying postoperative recurrence in the GLM mass phase showed better performance, among which the XGBoost model performed best. Targeted preventive measures can be given based on the above risk factors to improve the postoperative prognosis of the GLM mass phase.
ObjectiveTo systematically review the research progress on risk factors and predictive models for postoperative pulmonary infection (POPI) in gastric cancer patients, aiming to provide a reference for the early identification of high-risk patients and the optimization of clinical interventions. MethodsBy reviewing relevant domestic and international studies in recent years, the key risk factors for POPI in gastric cancer were summarized. And the construction methodologies, efficacy, and clinical application value of the latest predictive models developed in the last three years were evaluated. ResultsIn addition to traditional risk factors, recent studies had further confirmed the significant predictive value of novel factors for POPI following gastric cancer surgery, including nutritional-immune-inflammatory markers (such as prognostic nutritional index, C-reactive protein to albumin ratio, C-reactive protein-albumin-lymphocyte index), preoperative frailty, sarcopenia, and specific surgical approaches (e.g., differences between totally laparoscopic and laparoscopically assisted gastrectomy). Regarding predictive models, nomogram models developed based on multivariate logistic regression analysis and risk scoring systems had demonstrated favorable performance in both internal and partial external validations, with the area under the receiver operating characteristic curve mostly ranging from 0.74 to 0.97. Notably, composite models that integrate nutritional and immune-inflammatory markers with frailty assessments had shown superior predictive accuracy and clinical applicability. ConclusionsThis review provides a novel predictive perspective based on emerging biomarkers and functional assessments for the early identification of high-risk populations of POPI following gastric cancer surgery. Future research should prioritize the validation and refinement of existing models through multicenter collaboration, ultimately transforming them into more effective clinical risk assessment tools to guide precision prevention.
Objective To evaluate the clinical application value of four inflammatory indices [monocyte-to-lymphocyte ratio (MLR), platelet-to-lymphocyte ratio (PLR), systemic immune-inflammation index (SII), neutrophil-to-lymphocyte ratio (NLR)] in predicting postoperative mortality risk in patients with Siewert type Ⅱ esophagogastric junction adenocarcinoma, and to explore the predictive performance of four inflammatory indices. Methods This retrospective study collected clinical data from 310 patients with Siewert typeⅡ esophagogastric junction adenocarcinoma who were admitted to the Second Hospital of Lanzhou University between October 2016 and March 2023, and met the inclusion and exclusion criteria. Univariate analysis was used to initially screen variables related to postoperative mortality risk. The variance inflation factor (VIF) analysis was performed to assess multicollinearity issues, and multivariate regression analysis was used to further reveal the independent effects of key variables on postoperative mortality risk. The performance of the predictive models was evaluated using receive operatior characteristic curves and Kaplan-Meier survival analysis, and the effects of different inflammatory indices on patient survival were explored. Finally, machine learning methods such as Light GBM, random forest, support vector machine (SVM), and XGBoost were used to evaluate the predictive performance of the four inflammatory indices. Results The four inflammatory indices were significantly associated with postoperative mortality risk in patients with Siewert type Ⅱ esophagogastric junction adenocarcinoma (MLR: HR=2.6884, 95% CI 1.4559 to 4.9642, P=0.002; PLR: HR=1.0022, 95% CI1.0001 to 1.0043, P=0.041; SII: HR=1.0003, 95% CI1.0001 to 1.0006, P=0.002; NLR: HR=1.0697, 95% CI 1.0277 to 1.1134, P=0.001). Machine learning model results showed that NLR had the best performance in the random forest model, with an AUC of 0.863 in the training set and an AUC of 0.834 in the test set. Conclusion Preoperative clinical indicators, especially the NLR inflammatory factor, are of significant importance in predicting the postoperative mortality risk of patients with Siewert typeⅡ esophagogastric junction adenocarcinoma.
Objective To explore the independent risk factors for hospital infections in tertiary hospitals in Gansu Province, and establish and validate a prediction model. Methods A total of 690 patients hospitalized with hospital infections in Gansu Provincial Hospital between January and December 2021 were selected as the infection group; matched with admission department and age at a 1∶1 ratio, 690 patients who were hospitalized during the same period without hospital infections were selected as the control group. The information including underlying diseases, endoscopic operations, blood transfusion and immunosuppressant use of the two groups were compared, the factors influencing hospital infections in hospitalized patients were analyzed through multiple logistic regression, and the logistic prediction model was established. Eighty percent of the data from Gansu Provincial Hospital were used as the training set of the model, and the remaining 20% were used as the test set for internal validation. Case data from other three hospitals in Gansu Province were used for external validation. Sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) were used to evaluate the model effectiveness. Results Multiple logistic regression analysis showed that endoscopic therapeutic manipulation [odds ratio (OR)=3.360, 95% confidence interval (CI) (2.496, 4.523)], indwelling catheter [OR=3.100, 95%CI (2.352, 4.085)], organ transplantation/artifact implantation [OR=3.133, 95%CI (1.780, 5.516)], blood or blood product transfusions [OR=3.412, 95%CI (2.626, 4.434)], glucocorticoids [OR=2.253, 95%CI (1.608, 3.157)], the number of underlying diseases [OR=1.197, 95%CI (1.068, 1.342)], and the number of surgical procedures performed during hospitalization [OR=1.221, 95%CI (1.096, 1.361)] were risk factors for hospital infections. The regression equation of the prediction model was: logit(P)=–2.208+1.212×endoscopic therapeutic operations+1.131×indwelling urinary catheters+1.142×organ transplantation/artifact implantation+1.227×transfusion of blood or blood products+0.812×glucocorticosteroids+0.180×number of underlying diseases+0.200×number of surgical procedures performed during the hospitalization. The internal validation set model had a sensitivity of 72.857%, a specificity of 77.206%, an accuracy of 76.692%, and an AUC value of 0.817. The external validation model had a sensitivity of 63.705%, a specificity of 70.934%, an accuracy of 68.669%, and an AUC value of 0.726. Conclusions Endoscopic treatment operation, indwelling catheter, organ transplantation/artifact implantation, blood or blood product transfusion, glucocorticoid, number of underlying diseases, and number of surgical cases during hospitalization are influencing factors of hospital infections. The model can effectively predict the occurrence of hospital infections and guide the clinic to take preventive measures to reduce the occurrence of hospital infections.
With the increasing availability of clinical and biomedical big data, machine learning is being widely used in scientific research and academic papers. It integrates various types of information to predict individual health outcomes. However, deficiencies in reporting key information have gradually emerged. These include issues like data bias, model fairness across different groups, and problems with data quality and applicability. Maintaining predictive accuracy and interpretability in real-world clinical settings is also a challenge. This increases the complexity of safely and effectively applying predictive models to clinical practice. To address these problems, TRIPOD+AI (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis+artificial intelligence) introduces a reporting standard for machine learning models. It is based on TRIPOD and aims to improve transparency, reproducibility, and health equity. These improvements enhance the quality of machine learning model applications. Currently, research on prediction models based on machine learning is rapidly increasing. To help domestic readers better understand and apply TRIPOD+AI, we provide examples and interpretations. We hope this will support researchers in improving the quality of their reports.
ObjectiveTo explore the risk factors which affect the severity of acute pancreatitis by using machine learning algorithms. MethodsA retrospective review was conducted of medical records from 262 patients hospitalized for acute pancreatitis at the Second Affiliated Hospital of Zhengzhou University between October 2022 and February 2024. Patients were classified according to the revised edition Atlanta Classification into mild cases (n=146) and non-mild cases (n=116). LASSO analysis was employed to identify predictors for non-mild acute pancreatitis. Six machine learning algorithms, including extreme gradient boosting, random forest, logistic regression, decision tree, support vector machine, and K-nearest neighbors were integrated to construct predictive models. Model performance was evaluated by comparing the following metrics: area under the curve (AUC), sensitivity, specificity, accuracy, F1 score, calibration curves, and decision curves. ResultsThrough LASSO regression analysis, six feature variables, including heart rate, white blood cell count, neutrophil count, C-reactive protein, albumin, and calcium ion were selected to train and test machine learning models. Results showed that extreme gradient boosting achieved the highest AUC value of 0.93 on the test set, making it the optimal model. The sensitivity, specificity, accuracy, Brier score, and F1 score of the extreme gradient boosting model were 0.97, 0.70, 0.85, 0.108, and 0.84. ConclusionThe prediction model developed using extreme gradient boosting has high clinical utility value, helps to predict the severity of acute pancreatitis at an early stage and is valuable in guiding clinical decision-making.
ObjectiveTo investigate the risk factors for lymph node metastasis in resectable lung adenocarcinoma by combining spatial location, clinical, and imaging features, and to construct a lymph node metastasis prediction model. MethodsA retrospective study on patients who underwent chest CT at the First Affiliated Hospital of Nanjing Medical University from June 2016 to June 2020 and were surgically confirmed to have invasive lung adenocarcinoma with or without lymph node metastasis was conducted. Patients were divided into a positive group and a negative group based on the presence or absence of lymph node metastasis. Clinical and imaging data of the patients were collected, and the independent risk factors for lymph node metastasis in resectable lung adenocarcinoma were analyzed using univariate and multivariate logistic regression. A combined spatial location-clinical-imaging feature prediction model for lymph node metastasis was established and compared with the traditional lymph node metastasis prediction model that does not include spatial location features. ResultsA total of 611 patients were included, with 333 in the positive group, including 172 males and 161 females, with an average age of (58.9±9.7) years; and 278 in the negative group, including 127 males and 151 females, with an average age of (60.1±11.4) years. Univariate and multivariate logistic regression analyses showed that the spatial relationship of the lesion to the lung hilum, nodule type, pleural changes, and serum carcinoembryonic antigen (CEA) levels were independent risk factors for lymph node metastasis. Based on this, the combined spatial location-clinical-imaging feature prediction model had a sensitivity of 91.67%, specificity of 74.05%, accuracy of 87.88%, and area under the curve (AUC) of 0.885. The traditional lymph node metastasis prediction model, which did not include spatial location features, had a sensitivity of 76.40%, specificity of 72.10%, accuracy of 53.86%, and AUC of 0.827. The difference in AUC between the two prediction methods was statistically significant (P=0.026). Compared with the traditional prediction model, the predictive performance of the combined spatial location-clinical-imaging feature prediction model was significantly improved. ConclusionIn patients with resectable lung adenocarcinoma, those with central/inner spatial location, solid density, pleural changes with wide base depression, and elevated serum CEA levels have a higher risk of lymph node metastasis.
ObjectiveTo investigate the prognosis and satisfaction of the R2 intervention procedure and develop related predictive models. Methods The clinical data of 64 patients with primary craniofacial hyperhidrosis who underwent R2 intervention surgery at the First Affiliated Hospital of Fujian Medical University from November 2018 to October 2022 were retrospectively analyzed. By statistically analyzing the risk factors for compensatory hyperhidrosis (CH) and satisfaction, and conducting feature screening, a relevant prediction model was established. ResultsFinally, 51 patients were collected, including 43 (84.3%) males and 8 (15.7%) females, with an average age of (30.27±7.22) years. Overall postoperative satisfaction was high, with only 5.9% of patients expressing regret about the surgery. However, 92.2% of patients experienced CH. The onset of postoperative CH was most prominent within the first 3 months postoperatively, with the incidence rate stabilizing thereafter. Preoperative heart rate and R2 sympathetic nerve clipping were identified as independent risk factors for severe CH. The preoperative body mass index, the degree of sweating in the chest and abdomen, are significantly correlated with postoperative satisfaction. Conclusion The R2 intervention surgery effectively alleviates the symptoms of primary craniofacial hyperhidrosis, and patient satisfaction is high.