Research on grading prediction model of traumatic hemorrhage volume based on deep learning / 中华危重病急救医学

Objective:To develop a grading prediction model of traumatic hemorrhage volume based on deep learning and assist in predicting traumatic hemorrhage volume.Methods:A retrospective observational study was conducted based on the experimental data of pig gunshot wounds in the time-effect assessment database for experiments on war-traumatized animals constructed by the General Hospital of the Chinese People's Liberation Army. The hemorrhage volume data of the study population were extracted, and the animals were divided into 0-300 mL, 301-600 mL, and > 600 mL groups according to the hemorrhage volume. Using vital signs indexes as the predictive variables and hemorrhage volume grading as the outcome variable, trauma hemorrhage volume grading prediction models were developed based on four traditional machine learning and ten deep learning methods. Using laboratory test indexes as predictive variables and hemorrhage volume grading as outcome variables, trauma hemorrhage volume grading prediction models were developed based on the above fourteen methods. The effect of the two groups of models was evaluated by accuracy and area under the receiver operator characteristic curve (AUC), and the optimal models in the two groups were mixed to obtain hybrid model 1. Feature selection was conducted according to the genetic algorithm, and hybrid model 2 was constructed according to the best feature combination. Finally, hybrid model 2 was deployed in the animal experiment database system.Results:Ninety-six traumatic animals in the database were enrolled, including 27 pigs in the 0-300 mL group, 40 in the 301-600 mL group, and 29 in the > 600 mL group. Among the fourteen models based on vital signs indexes, fully convolutional network (FCN) model was the best [accuracy: 60.0%, AUC and 95% confidence interval (95% CI) was 0.699 (0.671-0.727)]. Among the fourteen models based on laboratory test indexes, recurrent neural network (RNN) model was the best [accuracy: 68.9%, AUC (95% CI) was 0.845 (0.829-0.860)]. After mixing the FCN and RNN models, the hybrid model 1, namely RNN-FCN model was obtained, and the performance of the model was improved [accuracy: 74.2%, AUC (95% CI) was 0.847 (0.833-0.862)]. Feature selection was carried out by genetic algorithm, and the hybrid model 2, namely RNN-FCN* model, was constructed according to the selected feature combination, which further improved the model performance [accuracy: 80.5%, AUC (95% CI) was 0.880 (0.868-0.893)]. The hybrid model 2 contained ten indexes, including mean arterial pressure (MAP), hematocrit (HCT), platelet count (PLT), lactic acid, arterial partial pressure of carbon dioxide (PaCO 2), Total CO 2, blood sodium, anion gap (AG), fibrinogen (FIB), international normalized ratio (INR). Finally, the RNN-FCN* model was deployed in the database system, which realized automatic, continuous, efficient, intelligent, and grading prediction of hemorrhage volume in traumatic animals. Conclusion:Based on deep learning, a grading prediction model of traumatic hemorrhage volume was developed and deployed in the information system to realize the intelligent grading prediction of traumatic animal hemorrhage volume.

Analysis of risk factors of early acute kidney injury after liver transplantation from DCD donor liver / 器官移植

Objective To analyze the risk factors of early acute kidney injury (AKI) after liver transplantation from donation after cardiac death(DCD) donor liver. Methods Clinical data of 184 donors and recipients undergoing liver transplantation from DCD donor liver were retrospectively analyzed. According to the incidence of early AKI, all participants were divided into the AKI and non-AKI groups. The patients in the AKI group were subject to AKI staging. Baseline data, preoperative, intraoperative and postoperative related parameters were statistically compared between two groups. The cumulative survival rate and clinical prognosis of patients in non-AKI group and AKI group with different staging were statistically analyzed by Kaplan-Meier curve analysis. Results Among 184 patients, 68 cases (37.0%) presented with early AKI after liver transplantation including 31 stage 1 AKI, 26 stage 2 AKI and 11 stage 3 AKI, mainly occurring within postoperative 3 d. Univariate analysis revealed that preoperative levels of albumin <35 g/L, preoperative levels of serum sodium ≤137 mmol/L, operation time>7.5 h, intraoperative hemorrhage volume>3 000 mL, intraoperative red cell infusion volume>15 U and intraoperative urine amount ≤100 mL/h were the risk factors of early AKI after liver transplantation (all P<0.05). Multi-variate Logistic regression analysis demonstrated that intraoperative red cell infusion >15 U was an independent risk factor of early AKI after liver transplantation [odds ratio(OR) 1.061, 95% confidence interval(CI)1.008-1.118,P=0.024].Result of Kaplan-Meier survival curve suggested that the cumulative survival rate was gradually reduced along with the aggravation of AKI with statistical significance (all P<0.05). Conclusions The incidence of early AKI following liver transplantation is relatively high. The severity of early AKI is intimately correlated with the short- and long-term prognosis of the recipients. A large quantity of intraoperative red blood cell infusion is an independent risk factor of AKI.

Analysis on cause,prevention and treatment of intra-abdominal hemorrhage after liver transplantation:a report of 82 cases / 器官移植

Objective To investigate the cause,prevention and treatment of intra-abdominal hemorrhage after liver transplantation. Methods Clinical data of 82 patients undergoing liver transplantation were retrospectively analyzed. All participants were divided into the intra-abdominal hemorrhage (n =12)and control groups (n =70). Preoperative parameters including age,model for end-stage liver disease (MELD)score,prothrombin time (PT),prothrombin time international normalized ratio (PT-INR),fibrinogen (FIB),activated partial thromboplastin time (APTT),platelet (Plt) were statistically compared between two groups. Intraoperative hemorrhage volume,cold ischemia time of donor liver, anhepatic phase time and operation time were also compared between two groups. Postoperatively,the mortality rate was compared between two groups. Results Among 82 patients,1 2 (1 5%)presented with intra-abdominal hemorrhage and required twice surgical hemostasis. In the intra-abdominal hemorrhage group,4 cases (33%)died,and 8 (1 1%)died in the control group. No statistical significance was documented between two groups (P>0. 05 ). Age,MELD score,PT-INR, FIB,APTT and PLT did not significantly differ between two groups (all P>0. 05 ). Compared with patients in the control group,those in the intra-abdominal hemorrhage group yielded significantly more blood loss intraoperatively,longer operation time and longer cold ischemia time of donor liver (all P<0. 05 ). Anhepatic phase time did not significantly differ between two groups (P>0. 05 ). Conclusions After liver transplantation,intra-abdominal hemorrhage is associated with longer cold ischemia time of donor liver,more intraoperative blood loss and longer operation time. In order to decrease the incidence of postoperative intra-abdominal hemorrhage,coagulation function should be completely corrected prior to surgery and the surgical skills should also be enhanced.

Relationship between different surgical methods, hemorrhage position, hemorrhage volume, surgical timing, and treatment outcome of hypertensive intracerebral hemorrhage / 世界急诊医学杂志（英文）

BACKGROUND: The present study aimed to explore the relationship between surgical methods, hemorrhage position, hemorrhage volume, surgical timing and treatment outcome of hypertensive intracerebral hemorrhage (HICH). METHODS: A total of 1310 patients, who had been admitted to six hospitals from January 2004 to January 2008, were divided into six groups according to different surgical methods: craniotomy through bone flap (group A), craniotomy through a small bone window (group B), stereotactic drilling drainage (group C1 and group C2), neuron-endoscopy operation (group D) and external ventricular drainage (group E) in consideration of hemorrhage position, hemorrhage volume and clinical practice. A retrospective analysis was made of surgical timing and curative effect of the surgical methods. RESULTS: The effectiveness rate of the methods was 74.12％ for 1310 patients after one-month follow-up. In this series, the disability rate was 44.82％ 3–6 months after the operation. Among the 1310 patients, 241 (18.40％) patients died after the operation. If hematoma volume was >80 mL and the operation was performed within 3 hours, the mortality rate of group A was significantly lower than that of groups B, C, D, and E (P<0.05). If hematoma volume was 50–80 mL and the operation was performed within 6–12 hours, the mortality rate of groups B and D was lower than that of groups A, C and E (P<0.05). If hematoma volume was 20–50 mL and the operation was performed within 6–24 hours, the mortality rate of group C was lower than that of groups A, B and D (P<0.05). CONCLUSIONS: Craniotomy through a bone flap is suitable for patients with a large hematoma and hernia of the brain. Stereotactic drilling drainage is suggested for patients with hematoma volume less than 80 mL. The curative effect of HICH individualized treatment would be improved via the suitable selection of operation time and surgical method according to the position and volume of hemorrhage.

A multicentre retrospective analysis of surgical effects of the 1310 Hypertensive intracerebral hemorrhage / 中华急诊医学杂志

Objective To explore the relationship between different hemorrhage position,hemorrhage volume,surgical time and outcome of treatment with surgical methods of HICH.Methods A total of 1310 patients were admitted from six hospitals from January 2004 to January 2008,the 1310 patients were divided into six groups according to different operation:craniotomy through bone flap (group A),craniotomy through small bone window (group B),stereotactic drilling drainage (group C1 and group C2),neuron-endoscopy operation (group D) and external ventricular drainage (group E),considering hemorrhage position,hemorrhage volume,surgical time and result of surgical methods were reviewed and analyzed.Results ①Craniotomy through bone flap should be selected with the case of superficial or deep hematoma volume (＞ 80 mL),median line structure distinct motion,metaphase or advanced stage of hernia of brain.②Craniotomy through small bone window and neuron-endoscopy should be selected with the case of moderate hematoma volume (50-80 mL) ③Drilling drainage should be selected with the case of small hematoma volume in superficial or deep hematoma volume (20-50 mL) ④Extemal drainage should be selected in dealing with ventricular hemorrhage.Small bone window or neuron-endoscopy should be selected in ventricular casting mould.⑤The appropriate operation time for patients with hematoma volume less than 80 mL should be 6-12 hours and large hematoma should be immediately operated to save lives.The operation time should depend on patients detail condition.Conclusions Craniotomy through bone flap was suitable for large hematoma and hernia of brain; Stereotactic drilling drainage should be selected in patients with hematoma volume less than 80mL; and the operation results in dealing with HICH would be improved via suitable operation time and surgical methods and adividual according to Hemorrhage position and Hemorrhage volume.

The Relationship Between Subarachnoid Hemorrhage Volume and Development of Cerebral Vasospasm

OBJECTIVE: The objective of this study is to verify the relationship between subarachnoid hemorrhage (SAH) volume (not Fisher grade) and development of cerebral vasospasm prospectively. METHODS: Patients who visited our hospital with a diffuse or localized thick subarachnoid blood clot seen on computed tomography (CT), taken within 48 hours after SAH and the aneurysm was confirmed by CT Angiogram (CTA) from March 2010 to July 2011 were enrolled in this study. CTA was checked at least twice after admission. Angiographic vasospasm (AVS) on CTA was defined as irregularity or narrowing of intracranial vessels on follow up CTA compared with initial CTA. Total intracranial hemorrhage (ICH) volume (subdural, SAH, intracerebral and intraventricular) was calculated and SAH volume (all supratentorial and infratentorial cisterns) was also calculated using the MIPAV software package. RESULTS: A total of 55 patients were included in our study. Thirty six patients did not show AVS on CTA or clinical deterioration (non vasospasm group: NVS). AVS without ischemic neurologic symptoms was observed in four patients and development of symptomatic vasospasm (SVS), defined as AVS with ischemic symptoms, was observed in 15 patients. SAH volume in SVS patients was statistically larger than that in NVS patients (p < 0.05). Total ICH volume in SVS patients was larger than that in NVS patients. However, the difference was not statistically significant. CONCLUSION: Results of this study indicate an association of development of vasospasm with the SAH volume, not intracranial hemorrhage.