Objective:To develop and validate a predictive model for massive hemorrhage during brain tumor resection in pediatric patients.Methods:A retrospective analysis was performed on the clinical data from pediatric patients who underwent elective brain tumor resection under general anesthesia at the Women and Children′s Medical Center Affiliated to Guangzhou Medical University from December 2016 to October 2023. The patients were randomly divided into model group and internal validation group in a ratio of 8∶2. Pediatric patients who underwent elective brain tumor resection under general anesthesia at Qilu Hospital of Shandong University from January 2021 to July 2024 were selected and served as external validation group. Relevant characteristic variables were screened through Lasso regression. A multivariate logistic regression was used to develop the model and plot the nomogram for intraoperative massive hemorrhage. The performance of the model was evaluated using the area under the receiver operating characteristic curve and calibration curve.Results:Through Lasso regression and multivariate logistic regression analyses, 11 independent influencing factors were identified: age ( OR=0.323, 95% confidence interval [ CI]: 0.280-0.374, P<0.001), weight ( OR=0.164, 95% CI: 0.135-0.199, P<0.001), activated partial thromboplastin time ( OR=1.133, 95% CI: 1.036-1.239, P=0.006), thrombin time ( OR=1.141, 95% CI: 1.048-1.243, P=0.002), red blood cell count ( OR=0.941, 95% CI: 0.888-0.996, P=0.035), hemoglobin concentration ( OR=0.873, 95% CI: 0.822-0.926, P<0.001), platelet count ( OR=1.062, 95% CI: 1.001-1.127, P=0.048), maximum tumor diameter ( OR=2.384, 95% CI: 2.241-2.536, P<0.001), tumor invasiveness ( OR=2.376, 95% CI: 2.071-2.726, P<0.001), hydrocephalus ( OR=2.409, 95% CI: 2.139-2.713, P<0.001), and centered midline structure ( OR=0.509, 95% CI: 0.465-0.557, P<0.001). Based on this, a nomogram prediction model was established. The receiver operating characteristic curve showed that the area under the curve of this model in predicting the risk of massive hemorrhage during brain tumor resection was 0.936 (95% CI: 0.90-0.959) in model group, 0.863 (95% CI: 0.744-0.948) in internal validation group, and 0.855 (95% CI: 0.726-0.955) in external validation group. The calibration curve indicated good model consistency, and the Hosmer-Lemeshow goodness-of-fit test result showed a P value of 0.979 ( P>0.05). Conclusions:Age, body weight, activated partial thromboplastin time, thrombin time, red blood cell count, hemoglobin concentration, platelet count, maximum tumor diameter, tumor invasiveness, hydrocephalus and midline structure are independent influencing factors for major bleeding during brain tumor resection in pediatric patients, and the prediction model established based on this histogram has high accuracy.

Objective:To develop and validate a predictive model for massive hemorrhage during brain tumor resection in pediatric patients.Methods:A retrospective analysis was performed on the clinical data from pediatric patients who underwent elective brain tumor resection under general anesthesia at the Women and Children′s Medical Center Affiliated to Guangzhou Medical University from December 2016 to October 2023. The patients were randomly divided into model group and internal validation group in a ratio of 8∶2. Pediatric patients who underwent elective brain tumor resection under general anesthesia at Qilu Hospital of Shandong University from January 2021 to July 2024 were selected and served as external validation group. Relevant characteristic variables were screened through Lasso regression. A multivariate logistic regression was used to develop the model and plot the nomogram for intraoperative massive hemorrhage. The performance of the model was evaluated using the area under the receiver operating characteristic curve and calibration curve.Results:Through Lasso regression and multivariate logistic regression analyses, 11 independent influencing factors were identified: age ( OR=0.323, 95% confidence interval [ CI]: 0.280-0.374, P<0.001), weight ( OR=0.164, 95% CI: 0.135-0.199, P<0.001), activated partial thromboplastin time ( OR=1.133, 95% CI: 1.036-1.239, P=0.006), thrombin time ( OR=1.141, 95% CI: 1.048-1.243, P=0.002), red blood cell count ( OR=0.941, 95% CI: 0.888-0.996, P=0.035), hemoglobin concentration ( OR=0.873, 95% CI: 0.822-0.926, P<0.001), platelet count ( OR=1.062, 95% CI: 1.001-1.127, P=0.048), maximum tumor diameter ( OR=2.384, 95% CI: 2.241-2.536, P<0.001), tumor invasiveness ( OR=2.376, 95% CI: 2.071-2.726, P<0.001), hydrocephalus ( OR=2.409, 95% CI: 2.139-2.713, P<0.001), and centered midline structure ( OR=0.509, 95% CI: 0.465-0.557, P<0.001). Based on this, a nomogram prediction model was established. The receiver operating characteristic curve showed that the area under the curve of this model in predicting the risk of massive hemorrhage during brain tumor resection was 0.936 (95% CI: 0.90-0.959) in model group, 0.863 (95% CI: 0.744-0.948) in internal validation group, and 0.855 (95% CI: 0.726-0.955) in external validation group. The calibration curve indicated good model consistency, and the Hosmer-Lemeshow goodness-of-fit test result showed a P value of 0.979 ( P>0.05). Conclusions:Age, body weight, activated partial thromboplastin time, thrombin time, red blood cell count, hemoglobin concentration, platelet count, maximum tumor diameter, tumor invasiveness, hydrocephalus and midline structure are independent influencing factors for major bleeding during brain tumor resection in pediatric patients, and the prediction model established based on this histogram has high accuracy.