Objective To develop and compare the predictive performance of five machine learning models for adverse postoperative outcomes in cardiac surgery patients, and to identify key decision factors through SHapley Additive exPlanations (SHAP) interpretability analysis. Methods A retrospective collection of perioperative data (including demographic information, preoperative, intraoperative, and postoperative indicators) with 88 variables was conducted from adult cardiac surgery patients at the First Affiliated Hospital of Xinjiang Medical University in 2023. Adverse postoperative outcomes were defined as the occurrence of acute kidney injury and/or in-hospital mortality during the postoperative hospitalization period following cardiac surgery. Patients were divided into an adverse outcome group and a favorable outcome group based on the presence of adverse postoperative outcomes. After screening feature variables using the least absolute shrinkage and selection operator (LASSO) regression method, five machine learning models were constructed: eXtreme gradient boosting (XGBoost), random forest (RF), gradient boosting machine (GBM), light gradient boosting machine (LightGBM), and generalized linear model (GLM). The dataset was randomly divided into a training set and a test set at a 7 : 3 ratio using stratified sampling, with postoperative outcome as the stratification factor. Model performance was evaluated using receiver operating characteristic curves, decision curve analysis, and F1 Score. The SHAP method was applied to analyze feature contribution. Results A total of 639 patients were included, comprising 395 males and 244 females, with a median age of 62 (55, 69) years. The adverse outcome group consisted of 191 patients, while the favorable outcome group included 448 patients, resulting in an adverse postoperative outcome incidence of 29.9%. Univariate analysis showed no significant differences between the two groups for any variables (P>0.05). Using LASSO regression, 16 feature variables were selected (including cardiopulmonary bypass support time, blood glucose on postoperative day 3, creatine kinase-MB isoenzyme, systemic inflammatory response index, etc.), and five machine learning models (GLM, RF, GBM, LightGBM, XGBoost) were constructed. Evaluation results demonstrated that the XGBoost model exhibited the best predictive performance on both the training set (n=447) and test set (n=192), with area under the curve values of 0.761 [95%CI (0.719, 0.800) ] and 0.759 [95%CI (0.692, 0.818) ], respectively. It also significantly outperformed other models in positive predictive value, and balanced accuracy in the test set. Decision curve analysis further confirmed its clinical utility across various risk thresholds. SHAP analysis indicated that variables such as cardiopulmonary bypass support time, blood glucose on postoperative day 3, creatine kinase-MB isoenzyme, and inflammatory markers (SIRI, NLR, CAR) had high contributions to the prediction. Conclusion The XGBoost model effectively predicts adverse postoperative outcomes in cardiac surgery patients. Clinically, attention should be focused on cardiopulmonary bypass support time, postoperative blood glucose control, and monitoring of inflammatory levels to improve patient prognosis.
The clinical performance and failure issues are significantly influenced by prosthetic malposition in unicompartmental knee arthroplasty (UKA). Uncertainty exists about the impact of the prosthetic joint line height in UKA on tibial insert wear. In this study, we combined the UKA musculoskeletal multibody dynamics model, finite element model and wear model to investigate the effects of seven joint line height cases of fixed UKA implant on postoperative insert contact mechanics, cumulative sliding distance, linear wear depth and volumetric wear. As the elevation of the joint line height in UKA, the medial contact force and the joint anterior-posterior translation during swing phase were increased, and further the maximum von Mises stress, contact stress, linear wear depth, cumulative sliding distance, and the volumetric wear also were increased. Furthermore, the wear area of the insert gradually shifted from the middle region to the rear. Compared to 0 mm joint line height, the maximum linear wear depth and volumetric wear were decreased by 7.9% and 6.8% at –2 mm joint line height, and by 23.7% and 20.6% at –6 mm joint line height, the maximum linear wear depth and volumetric wear increased by 10.7% and 5.9% at +2 mm joint line height, and by 24.1% and 35.7% at +6 mm joint line height, respectively. UKA prosthetic joint line installation errors can significantly affect the wear life of the polyethylene inserted articular surfaces. Therefore, it is conservatively recommended that clinicians limit intraoperative UKA joint line height errors to –2?+2 mm.
ObjectiveTo explore the effect of metabolic syndrome (MS) on postoperative pulmonary infection in patients with colorectal cancer (CRC) and to construct a risk prediction model for postoperative pulmonary infection in CRC patients. MethodsRetrospective collection of clinical data from 291 CRC patients who underwent surgical treatment at Department of General Surgery, Suzhou Ninth People’s Hospital in the period of January 2020 to August 2024. To explore the risk factors of postoperative pulmonary infection in patients with CRC and to establish a nomogram model. ResultsAmong the 291 CRC patients enrolled, there were 58 MS patients (19.93%) and 233 non-MS patients (80.07%). Compared with patients without MS, CRC patients with MS had longer surgery time (P<0.001) and higher incidence of postoperative pulmonary infection (P<0.001). The results of multiple logistic regression analysis showed that smoking history [OR=2.184, 95%CI (1.097, 4.345), P=0.026], body mass index (BMI)≥25 kg/m2 [OR=2.662, 95%CI (1.241, 5.703), P=0.012], MS [OR=2.770, 95%CI (1.415, 5.425), P=0.003], increased surgical time [OR=4.039, 95%CI (1.774, 9.197), P<0.001] and increased intraoperative bleeding [OR=2.398, 95%CI (1.246, 4.618), P=0.009] were all risk factors for postoperative pulmonary infection in CRC patients. Based on these risk factors, a nomogram model was constructed. The area under the curve (AUC) was 0.845 [95%CI (0.769, 0.906)], and the sensitivity and specificity were 84.2% and 87.5% respectively. The internal verification of Bootstrap test showed that the simulated curve and the actual curve had good consistency. The clinical decision curve analysis showed that when the threshold probability was in the range of 8%–84%, the net benefit of the model for patient diagnosis was higher. ConclusionsMS increases the risk of postoperative pulmonary infection in CRC patients. At the same time, smoking history, BMI≥25 kg/m2, long operation time, and more intraoperative blood loss are also risk factors for postoperative pulmonary infection in patients with CRC. Building a model based on this can effectively evaluate the risk of postoperative pulmonary infection in CRC patients.
Pathological images of gastric cancer serve as the gold standard for diagnosing this malignancy. However, the recurrence prediction task often encounters challenges such as insignificant morphological features of the lesions, insufficient fusion of multi-resolution features, and inability to leverage contextual information effectively. To address these issues, a three-stage recurrence prediction method based on pathological images of gastric cancer is proposed. In the first stage, the self-supervised learning framework SimCLR was adopted to train low-resolution patch images, aiming to diminish the interdependence among diverse tissue images and yield decoupled enhanced features. In the second stage, the obtained low-resolution enhanced features were fused with the corresponding high-resolution unenhanced features to achieve feature complementation across multiple resolutions. In the third stage, to address the position encoding difficulty caused by the large difference in the number of patch images, we performed position encoding based on multi-scale local neighborhoods and employed self-attention mechanism to obtain features with contextual information. The resulting contextual features were further combined with the local features extracted by the convolutional neural network. The evaluation results on clinically collected data showed that, compared with the best performance of traditional methods, the proposed network provided the best accuracy and area under curve (AUC), which were improved by 7.63% and 4.51%, respectively. These results have effectively validated the usefulness of this method in predicting gastric cancer recurrence.
ObjectiveTo explore the predictive effect of the femoral neck strength composite indexes on femoral head collapse in non-traumatic osteonecrosis of the femoral head (ONFH) compared with bone turnover marker.MethodsThe non-traumatic ONFH patients who were admitted and received non-surgical treatment between January 2010 and December 2016 as the research object. And 96 cases (139 hips) met the selection criteria and were included in the study. There were 54 males (79 hips) and 42 females (60 hips), with an average age of 40.2 years (range, 22-60 years). According to whether the femoral head collapsed during follow-up, the patients were divided into collapsed group and non-collapsed group. The femoral neck width, hip axis length, height, body weight, and bone mineral density of femoral neck were measured. The femoral neck strength composite indexes, including the compressive strength index (CSI), bending strength index (BSI), and impact strength index (ISI), were calculated. The bone turnover marker, including the total typeⅠcollagen amino terminal elongation peptide (t-P1NP), β-crosslaps (β-CTx), alkaline phosphatase (ALP), 25 hydroxyvitamin D [25(OH)D], and N-terminal osteocalcin (N-MID), were measured. The age, gender, height, body weight, body mass index (BMI), bone mineral density of femoral neck, etiology, Japanese Osteonecrosis Investigation Committee (JIC) classification, femoral neck strength composite indexes, and bone turnover marker were compared between the two groups, and the influencing factors of the occurrence of femoral head collapse were initially screened. Then the significant variables in the femoral neck strength composite indexes and bone turnover marker were used for logistic regression analysis to screen risk factors; and the receiver operating characteristic (ROC) curve was used to determine the significant variables’ impact on non-traumatic ONFH. ResultsAll patients were followed up 3.2 years on average (range, 2-4 years). During follow-up, 46 cases (64 hips) had femoral head collapse (collapsed group), and the remaining 50 cases (75 hips) did not experience femoral head collapse (non-collapsed group). Univariate analysis showed that the difference in JIC classification between the two groups was significant (Z=–7.090, P=0.000); however, the differences in age, gender, height, body weight, BMI, bone mineral density of femoral neck, and etiology were not significant (P>0.05). In the femoral neck strength composite indexes, the CSI, BSI, and ISI of the collapsed group were significantly lower than those of the non-collapsed group (P<0.05); in the bone turnover marker, the t-P1NP and β-CTx of the collapsed group were significantly lower than those of the non-collapsed group (P<0.05); there was no significant difference in N-MID, 25(OH)D or ALP between groups (P>0.05). Multivariate analysis showed that the CSI, ISI, and t-P1NP were risk factors for femoral collapse in patients with non-traumatic ONFH (P<0.05). ROC curve analysis showed that the cut-off points of CSI, BSI, ISI, t-P1NP, and β-CTx were 6.172, 2.435, 0.465, 57.193, and 0.503, respectively, and the area under the ROC curve (AUC) were 0.753, 0.642, 0.903, 0.626, and 0.599, respectively. ConclusionThe femoral neck strength composite indexes can predict the femoral head collapse in non-traumatic ONFH better than the bone turnover marker. ISI of 0.465 is a potential cut-off point below which future collapse of early non-traumatic ONFH can be predicted.
ObjectiveTo discuss the risk factors of acute respiratory distress syndrome (ARDS) in patients with severe pneumonia.MethodsData of 80 patients with severe pneumonia admitted in our ICU were analyzed retrospectively, and they were divided into two groups according to development of ARDS, which was defined according to the Berlin new definition. The age, gender, weight, Acute Physiology and Chronic Health EvaluationⅡscore, lactate, PSI score and LIPS score, etc. were collected. Statistical significance results were evaluated by multivariate logistic regression analysis after univariate analysis. Receiver operating characteristic (ROC) curve was plotted to analyze the predictive value of the parameter for ARDS after severe pneumonia.ResultsForty patients with severe pneumonia progressed to ARDS, there were 4 moderate cases and 36 severe cases according to diagnostic criteria. Univariate analysis showed that procalcitonin (t=4.08, P<0.001), PSI score (t=10.67, P<0.001), LIPS score (t=5.14, P<0.001), shock (χ2=11.11, P<0.001), albumin level (t=3.34, P=0.001) were related to ARDS. Multivariate logistic regression analysis showed that LIPS ［odds ratio (OR) 0.226, 95%CI=4.62-5.53, P=0.013］ and PSI (OR=0.854, 95%CI=132.2-145.5, P=0.014) were independent risk factors for ARDS. The predictive value of LIPS and PSI in ARDS occurrence was significant. The area under ROC curve (AUC) of LIPS was 0.901, the cut-off value was 7.2, when LIPS ≥7.2, the sensitivity and specificity were both 85.0%. AUC of PSI was 0.947, the cut-off value was 150.5, when PSI score ≥150.5, the sensitivity and specificity were 87.5% and 90.0% respectively.ConclusionsPSI and LIPS are independent risk factors of ARDS in patients with severe pneumonia, which may be references for guiding clinicians to make an early diagnosis and treatment plan.
ObjectiveTo analyze prognostic ability of inflammation-based Glasgow prognostic score (GPS) in patients with ST-segment elevation myocardial infarction (STEMI).MethodsWe retrospectively analyzed the clinical data of 289 patients with STEMI admitted to the Department of Emergency in West China Hospital from April 2015 to January 2016. All study subjects were divided into three groups: a group of GPS 0 (190 patients including 150 males and 40 females aged 62.63±12.98 years), a group of GPS 1 (78 patients including 58 males and 20 females aged 66.57±15.25 years), and a group of GPS 2 (21 patients including 16 males and 5 females aged 70.95±9.58 years). Cox regression analysis was conducted to analyze the independent risk factors of predicting long-term mortality of patients with STEMI.ResultsThere was a statistical difference in long-term mortality (9.5% vs. 23.1% vs. 61.9%, P<0.001) and in-hospital mortality (3.7% vs. 7.7% vs. 23.8%, P<0.001) among the three groups. The Global Registry of Acute Coronary Events (GRACE) scores and Gensini scores increased in patients with higher GPS scores, and the differences were statistically different (P<0.001). Multivariable Cox regression analysis showed that the GPS was independently associated with STEMI long-term all-cause mortality (1 vs. 0, HR: 2.212, P=0.037; 2 vs. 0, HR: 8.286, P<0.001).ConclusionGPS score is helpful in predicting the long-term and in-hospital prognosis of STEMI patients, and thus may guide clinical precise intervention by early risk stratification.
ObjectiveProlonged mechanical ventilation (PMV) is a prognostic marker for short-term adverse outcomes in patients after lung transplantation.The risk of prolonged mechanical ventilation after lung transplantation is still not clear. The study to identify the risk factors of prolonged mechanical ventilation (PMV) after lung transplantation.Methods This retrospective observational study recruited patients who underwent lung transplantation in Wuxi People’s Hospital from January 2020 to December 2022. Relevant information was collected from patients and donors, including recipient data (gender, age, BMI, blood type, comorbidities), donor data (age, BMI, time of endotracheal intubation, oxygenation index, history of smoking, and any comorbidity with multidrug-resistant bacterial infections), and surgical data (surgical mode, incision type, operation time, cold ischemia time of the donor lung, intraoperative bleeding, and ECMO support), and postoperative data (multi-resistant bacterial lung infection, multi-resistant bacterial bloodstream infection, and mean arterial pressure on postoperative admission to the monitoring unit). Patients with a duration of mechanical ventilation ≤72 hours were allocated to the non-prolonged mechanical ventilation group, and patients with a duration of mechanical ventilation>72 hours were allocated to the prolonged mechanical ventilation group. LASSO regression analysis was applied to screen risk factors., and a clinical prediction model for the risk of prolonged mechanical ventilation after lung.ResultsPatients who met the inclusion criteria were divided into the training set and the validation set. There were 307 cases in the training set group and 138 cases in the validation set group. The basic characteristics of the training set and the validation set were compared. There were statistically significant differences in the recipient’s BMI, donor’s gender, CRKP of the donor lung swab, whether the recipient had pulmonary infection before the operation, the type of transplantation, the cold ischemia time of the donor lung, whether ECMO was used during the operation, the duration of ECMO assistance, CRKP of sputum, and the CRE index of the recipient's anal test (P<0.05). 2. The results of the multivariate logistic regression model showed that female recipients, preoperative mechanical ventilation in recipients, preoperative pulmonary infection in recipients, intraoperative application of ECMO, and the detection of multi-drug resistant Acinetobacter baumannii, multi-drug resistant Klebsiella pneumoniae and maltoclomonas aeruginosa in postoperative sputum were independent risk factors for prolonged mechanical ventilation after lung transplantation. The AUC of the clinical prediction model in the training set and the validation set was 0.838 and 0.828 respectively, suggesting that the prediction model has good discrimination. In the decision curves of the training set and the validation set, the threshold probabilities of the curves in the range of 0.05-0.98 and 0.02-0.85 were higher than the two extreme lines, indicating that the model has certain clinical validity.ConclusionsFemale patients, Preoperative pulmonary infection, preoperative mechanical ventilation,blood type B, blood type O, application of ECMO assistance, multi-resistant Acinetobacter baumannii infection, multi-resistant Klebsiella pneumoniae infection, and multi-resistant Stenotrophomonas maltophilia infection are independent risk factors for PMV (prolonged mechanical ventilation) after lung transplantation.
Objective To establish and verify the early prediction model of critical illness patients with influenza. Methods Critical illness patients with influenza who diagnosed with influenza in the emergency departments from West China Hospital of Sichuan University, Shangjin Hospital of West China Hospital of Sichuan University, and Panzhihua Central Hospital between January 1, 2017 and June 30, 2020 were selected. According to K-fold cross validation method, 70% of patients were randomly assigned to the model group, and 30% of patients were assigned to the model verification group. The patients in the model group and the model verification group were divided into the critical illness group and the non-critical illness group, respectively. Based on the modified National Early Warning Score (MEWS) and the Simplified British Thoracic Society Score (confusion, uremia, respiratory, BP, age 65 years, CRB-65 score), a critical illness influenza early prediction model was constructed and its accuracy was evaluated. Results A total of 612 patients were included. Among them, there were 427 cases in the model group and 185 cases in the model verification group. In the model group, there were 304 cases of non-critical illness and 123 cases of critical illness. In the model verification group, there were 152 cases of non-critical illness and 33 cases of critical illness. The results of binary logistic regression analysis showed that age, hypertension, the number of days between the onset of symptoms and presentation at the emergency department, consciousness state, white blood cell count, and lymphocyte count, oxygen saturation of blood were the independent risk factors for critical illness influenza. Based on these 7 risk factors, an early prediction model for critical illness influenza was established. The correct percentages of the model for non-critical illness and critical illness patients were 95.4% and 77.2%, respectively, with an overall correct prediction percentage of 90.2%. The results of the receiver operator characteristic curve showed that the sensitivity and specificity of the early prediction model for critical illness influenza in predicting critical illness patients were 0.909, 0.921, and the area under the curve and its 95% confidence interval were 0.931 (0.860, 0.999). The sensitivity, specificity, and area under the curve (0.935, 0.865, 0.942) of the early prediction model for critical illness influenza were higher than those of MEWS (0.642, 0.595, 0.536) and CRB-65 (0.628, 0.862, 0.703). Conclusions The conclusion is that age, hypertension, the number of days between the onset of symptoms and presentation at the emergency department, consciousness, oxygen saturation, white blood cell count, and absolute lymphocyte count are independent risk factors for predicting severe influenza cases. The early prediction model for critical illness patients with influenza has high accuracy in predicting severe influenza cases, and its predictive value and accuracy are superior to those of the MEWS score and CRB-65 score.
Objective To analyze the epidemic trend of prostate cancer in China from 1992 to 2021, and predict its epidemic trends from 2022 to 2032. Methods Based on the data of Chinese population and prostate cancer incidence and mortality from Global Burden of Disease Database, the Joinpoint log-linear model was used to analyze the trends of prostate cancer incidence and mortality, use the age-period-cohort model to analyze the effects of age, period and cohort on changes in incidence and mortality, and the gray prediction model was used to predict the trends of prostate cancer. Results From 1992 to 2021, the incidence and mortality of prostate cancer in China showed an upward trend, with AAPC of 5.652% (P<0.001) and 3.466% (P<0.001), and the AAPC of age-standardized incidence decreased to 1.990% (P<0.001), the age-standardized mortality showed a downward trend and was not statistically significant. The results of the age-period-cohort model showed that the net drift values of prostate cancer incidence and mortality were 3.03% and ?1.06%, respectively, and the risk of incidence and mortality gradually increased with age and period. The results of the grey prediction model showed that the incidence and mortality of prostate cancer showed an upward trend from 2022 to 2032, and the incidence trend was more obvious. Conclusion The incidence and mortality of prostate cancer in China showed an increasing trend, with a heavy disease burden and severe forms of prevention and control, so it is necessary to do a good job in monitoring the incidence and mortality of prostate cancer, and strengthen the efficient screening, early diagnosis and treatment of prostate cancer.