| 61. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
A Comparative Study of Machine Learning Algorithms in Predicting Severe Complications after Bariatric Surgery
- 2019
-
Ingår i: Journal of Clinical Medicine. - : MDPI. - 2077-0383. ; 8:5
-
Tidskriftsartikel (refereegranskat)abstract
- Background: Severe obesity is a global public health threat of growing proportions. Accurate models to predict severe postoperative complications could be of value in the preoperative assessment of potential candidates for bariatric surgery. So far, traditional statistical methods have failed to produce high accuracy. We aimed to find a useful machine learning (ML) algorithm to predict the risk for severe complication after bariatric surgery.Methods: We trained and compared 29 supervised ML algorithms using information from 37,811 patients that operated with a bariatric surgical procedure between 2010 and 2014 in Sweden. The algorithms were then tested on 6250 patients operated in 2015. We performed the synthetic minority oversampling technique tackling the issue that only 3% of patients experienced severe complications.Results: Most of the ML algorithms showed high accuracy (>90%) and specificity (>90%) in both the training and test data. However, none of the algorithms achieved an acceptable sensitivity in the test data. We also tried to tune the hyperparameters of the algorithms to maximize sensitivity, but did not yet identify one with a high enough sensitivity that can be used in clinical praxis in bariatric surgery. However, a minor, but perceptible, improvement in deep neural network (NN) ML was found.Conclusion: In predicting the severe postoperative complication among the bariatric surgery patients, ensemble algorithms outperform base algorithms. When compared to other ML algorithms, deep NN has the potential to improve the accuracy and it deserves further investigation. The oversampling technique should be considered in the context of imbalanced data where the number of the interested outcome is relatively small.
|
|
| 62. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Comparison of piecewise structural equation modeling and Bayesian network for de novo construction of a quantitative adverse outcome pathway network
- 2023
-
Ingår i: Altex. - : Springer Spektrum. - 1868-596X .- 1868-8551. ; 40:2, s. 287-298
-
Tidskriftsartikel (refereegranskat)abstract
- Quantitative adverse outcome pathway network (qAOPN) is gaining momentum due to the predictive nature, alignment with quantitative risk assessment and great potential as a computational new approach methodology (NAM) to reduce laboratory animal tests. The present work aimed to demonstrate two advanced modeling approaches, piecewise structural equation modeling (PSEM) and Bayesian network (BN), for de novo qAOPN model construction based on routine ecotoxicological data. A previously published AOP network comprised of four linear AOPs linking excessive reactive oxygen species production to mortality in aquatic organisms was employed as a case study. The demonstrative case study intended to answer: Which linear AOP in the network contributed the most to the AO? Can any of the upstream KEs accurately predict the AO? What are the advantages and limitations of PSEM or BN in qAOPN development? The outcomes from the two approaches showed that both PSEM and Bayesian network were suitable for constructing a complex qAOPN based on limited experimental data. Besides quantification of response-response relationships, both approaches were capable of identifying the most influencing linear AOP in a complex network and evaluating the predictive ability of the AOP, albeit some discrepancies in predictive ability were identified for the two approaches using this specific dataset. The PROs and CONs of the two approaches for qAOPN construction were discussed in detail and suggestions on optimal workflows of PSEM and BN were provided to guide future qAOPN development.
|
|
| 63. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Deep Learning Neural Networks to Predict Serious Complications After Bariatric Surgery : Analysis of Scandinavian Obesity Surgery Registry Data
- 2020
-
Ingår i: JMIR Medical Informatics. - : JMIR Publications. - 2291-9694. ; 8:5
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND: Obesity is one of today's most visible public health problems worldwide. Although modern bariatric surgery is ostensibly considered safe, serious complications and mortality still occur in some patients.OBJECTIVE: This study aimed to explore whether serious postoperative complications of bariatric surgery recorded in a national quality registry can be predicted preoperatively using deep learning methods.METHODS: Patients who were registered in the Scandinavian Obesity Surgery Registry (SOReg) between 2010 and 2015 were included in this study. The patients who underwent a bariatric procedure between 2010 and 2014 were used as training data, and those who underwent a bariatric procedure in 2015 were used as test data. Postoperative complications were graded according to the Clavien-Dindo classification, and complications requiring intervention under general anesthesia or resulting in organ failure or death were considered serious. Three supervised deep learning neural networks were applied and compared in our study: multilayer perceptron (MLP), convolutional neural network (CNN), and recurrent neural network (RNN). The synthetic minority oversampling technique (SMOTE) was used to artificially augment the patients with serious complications. The performances of the neural networks were evaluated using accuracy, sensitivity, specificity, Matthews correlation coefficient, and area under the receiver operating characteristic curve.RESULTS: In total, 37,811 and 6250 patients were used as the training data and test data, with incidence rates of serious complication of 3.2% (1220/37,811) and 3.0% (188/6250), respectively. When trained using the SMOTE data, the MLP appeared to have a desirable performance, with an area under curve (AUC) of 0.84 (95% CI 0.83-0.85). However, its performance was low for the test data, with an AUC of 0.54 (95% CI 0.53-0.55). The performance of CNN was similar to that of MLP. It generated AUCs of 0.79 (95% CI 0.78-0.80) and 0.57 (95% CI 0.59-0.61) for the SMOTE data and test data, respectively. Compared with the MLP and CNN, the RNN showed worse performance, with AUCs of 0.65 (95% CI 0.64-0.66) and 0.55 (95% CI 0.53-0.57) for the SMOTE data and test data, respectively.CONCLUSIONS: MLP and CNN showed improved, but limited, ability for predicting the postoperative serious complications after bariatric surgery in the Scandinavian Obesity Surgery Registry data. However, the overfitting issue is still apparent and needs to be overcome by incorporating intra- and perioperative information.
|
|
| 64. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Determination of clinically acceptable cut-offs for hemolysis index : an application of bootstrap method using real-world data
- 2021
-
Ingår i: Clinical Biochemistry. - : Elsevier. - 0009-9120 .- 1873-2933. ; 94, s. 74-79
-
Tidskriftsartikel (refereegranskat)abstract
- Objectives: To assess the impact of hemolysis on laboratory results under local conditions and to verify the hemolysis index cut-off for potassium using real-world data.Methods: The statistical bootstrapping method was performed on 54,125 samples collected at the University Hospital of Örebro (USÖ) and the results were compared to a method based on stratification of samples according to hemolysis level, and on paired difference testing.Results: Setting the acceptable allowable limit of error to 10 %, the three assessed strategies yielded comparable results with respect to the impact of haemolytic interference on test results for potassium. The suggested cut-offs were 111 mg Hb/dL for the bootstrapping method, between 100-125 mg Hb/dL for the method based on stratification, and around 150 mg/dL for the paired difference testing strategy. The impact of hemolysis on potassium measurement is likely different between primary care patients and inpatients.Conclusions: Using the effect of hemolysis on potassium measurement as a model, a novel approach towards finding clinically acceptable limits for analytical interference is presented, that relies on the bootstrapping method and on actual patient data from routine laboratory operation, hence incorporating local population characteristics, equipment and instrumental settings.
|
|
| 65. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Development and Validation of an XGBoost-Algorithm-Powered Survival Model for Predicting In-Hospital Mortality Based on 545,388 Isolated Severe Traumatic Brain Injury Patients from the TQIP Database
- 2023
-
Ingår i: Journal of Personalized Medicine. - : MDPI. - 2075-4426. ; 13:9
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND: Traumatic brain injury (TBI) represents a significant global health issue; the traditional tools such as the Glasgow Coma Scale (GCS) and Abbreviated Injury Scale (AIS) which have been used for injury severity grading, struggle to capture outcomes after TBI.AIM AND METHODS: This paper aims to implement extreme gradient boosting (XGBoost), a powerful machine learning algorithm that combines the predictions of multiple weak models to create a strong predictive model with high accuracy and efficiency, in order to develop and validate a predictive model for in-hospital mortality in patients with isolated severe traumatic brain injury and to identify the most influential predictors. In total, 545,388 patients from the 2013-2021 American College of Surgeons Trauma Quality Improvement Program (TQIP) database were included in the current study, with 80% of the patients used for model training and 20% of the patients for the final model test. The primary outcome of the study was in-hospital mortality. Predictors were patients' demographics, admission status, as well as comorbidities, and clinical characteristics. Penalized Cox regression models were used to investigate the associations between the survival outcomes and the predictors and select the best predictors. An extreme gradient boosting (XGBoost)-powered Cox regression model was then used to predict the survival outcome. The performance of the models was evaluated using the Harrell's concordance index (C-index). The time-dependent area under the receiver operating characteristic curve (AUC) was used to evaluate the dynamic cumulative performance of the models. The importance of the predictors in the final prediction model was evaluated using the Shapley additive explanations (SHAP) value.RESULTS: On average, the final XGBoost-powered Cox regression model performed at an acceptable level for patients with a length of stay up to 250 days (mean time-dependent AUC = 0.713) in the test dataset. However, for patients with a length of stay between 20 and 213 days, the performance of the model was relatively poor (time-dependent AUC < 0.7). When limited to patients with a length of stay ≤20 days, which accounts for 95.4% of all the patients, the model achieved an excellent performance (mean time-dependent AUC = 0.813). When further limited to patients with a length of stay ≤5 days, which accounts for two-thirds of all the patients, the model achieved an outstanding performance (mean time-dependent AUC = 0.917).CONCLUSION: The XGBoost-powered Cox regression model can achieve an outstanding predictive ability for in-hospital mortality during the first 5 days, primarily based on the severity of the injury, the GCS on admission, and the patient's age. These variables continue to demonstrate an excellent predictive ability up to 20 days after admission, a period of care that accounts for over 95% of severe TBI patients. Past 20 days of care, other factors appear to be the primary drivers of in-hospital mortality, indicating a potential window of opportunity for improving outcomes.
|
|
| 66. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Does background postnatal methyl mercury exposure in toddlers affect cognition and behavior?
- 2010
-
Ingår i: Neurotoxicology. - : Elsevier. - 0161-813X .- 1872-9711. ; 31:1, s. 1-9
-
Tidskriftsartikel (refereegranskat)abstract
- Because the toxicological effects of mercury (Hg) are more serious in the developing central nervous system of children than adults, there are growing concerns about prenatal and early childhood Hg exposure. This study examined postnatal methylmercury (MeHg) exposure and cognition and behavior in 780 children enrolled in the Treatment of Lead (Pb)-exposed Children clinical trial (TLC) with 396 children allocated to the succimer and 384 to the placebo groups. Mercury exposure was determined from analyses of blood drawn 1 week before randomization and 1 week after treatment began when succimer had its maximal effect on blood Pb (PbB). The baseline MeHg concentrations were 0.54 microg/L and 0.52 microg/L and post-treatment concentrations were 0.51 microg/L and 0.48 microg/L for placebo and succimer groups, respectively. Because the baseline characteristics in the two groups were balanced and because succimer had little effect on MeHg concentration and no effect on the cognitive or behavioral test scores, the groups were combined in the analysis of MeHg and neurodevelopment. The children's IQ and neurobehavioral performance were tested at age 2, 5 and 7 years. We saw weak, non-significant but consistently positive associations between blood MeHg and IQ test scores in stratified, spline regression and generalized linear model data analyses. The behavioral problem scores were constant or decreased slightly with increasing MeHg concentration. Additional adjustment for PbB levels in multivariable models did not alter the conclusion for MeHg and IQ scores, but did confirm that concurrent PbB was strongly associated with IQ and behavior in TLC children. The effects of MeHg on neurodevelopmental indices did not substantially differ by PbB strata. We conclude that at the present background postnatal MeHg exposure levels of US children, adverse effects on children's IQ and behavior are not detectable.
|
|
| 67. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Efficacy of succimer chelation of mercury at background exposures in toddlers : a randomized trial
- 2011
-
Ingår i: The Journal of Pediatrics. - : Elsevier. - 0022-3476 .- 1097-6833. ; 158:3, s. 480-485.e1
-
Tidskriftsartikel (refereegranskat)abstract
- OBJECTIVE: To examine whether succimer, a mercaptan compound known to reduce blood lead concentration in children, reduces blood mercury concentration.STUDY DESIGN: We used samples from a randomized clinical trial of succimer chelation for lead-exposed children. We measured mercury levels in pre-treatment samples from 767 children. We also measured mercury levels in blood samples drawn 1 week after treatment began (n = 768) and in a 20% random sample of the children who received the maximum 3 courses of treatment (n = 67). A bootstrap-based isotonic regression method was used to compare the trend with time in the difference between the adjusted mean mercury concentrations in the succimer group and that in the placebo group.RESULTS: The adjusted mean organic mercury concentration in the succimer group relative to the placebo group fell from 99% at baseline to 82% after 3 courses of treatment (P for trend = .048), but this resulted from the prevention of the age-related increase in the succimer group.CONCLUSION: Succimer chelation for low level organic mercury exposure in children has limited efficacy.
|
|
| 68. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Goitrogenic anions, thyroid-stimulating hormone, and thyroid hormone in infants
- 2010
-
Ingår i: Journal of Environmental Health Perspectives. - : National Institute of Environmental Health Science. - 0091-6765 .- 1552-9924. ; 118:9, s. 1332-1337
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND: Environmental exposure of infants to perchlorate, thiocyanate, nitrate, might interfere with thyroid function. U.S. women with higher background perchlorate exposure have higher thyroid-stimulating hormone (TSH) and lower thyroxine (T4). There are no studies with individual measures of thyroid function and these goitrogens available in infants.OBJECTIVE: We examined the association of urinary perchlorate, nitrate, iodide, and thiocyanate with urinary T4 and TSH in infants and whether that association differed by sex or iodide status.METHODS: We used data and samples from the Study of Estrogen Activity and Development, which assessed hormone levels of full-term infants over the first 12 months of life. The study included 92 full-term infants between birth and 1 year of age seen up to four times. Perchlorate, thiocyanate, nitrate, and iodide were measured in 206 urine samples; TSH and T4 and were measured in urines and in 50 blood samples.RESULTS: In separate mixed models, adjusting for creatinine, age, sex, and body mass index, infants with higher urinary perchlorate, nitrate or thiocyanate had higher urinary TSH. With all three modeled, children with higher nitrate and thiocyanate had higher TSH, but higher perchlorate was associated with TSH only in children with low iodide. Unexpectedly, exposure to the three chemicals was generally associated with higher T4.CONCLUSIONS: The association of perchlorate exposure with increased urinary TSH in infants with low urinary iodide is consistent with previous findings. Higher thiocyanate and nitrate exposure were also associated with higher TSH in infants.
|
|
| 69. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Is Olfactory Impairment Associated With 10-year Mortality Mediating by Neurodegenerative Diseases in Older Adults? The Four-Way Decomposition Analysis
- 2021
-
Ingår i: Frontiers In Public Health. - : Frontiers Media S.A.. - 2296-2565. ; 9
-
Tidskriftsartikel (refereegranskat)abstract
- Background: Literature shows that olfactory impairment (OI) is associated not only with neurodegenerative diseases (NDDs), but also with increased mortality. In this study, we analyzed data collected from the prospective phase of the 10-year follow-up of the Shanghai Aging Study (SAS) to explore the mediation effect of NDDs on the OI-mortality relationship.Methods: We analyzed data collected from the prospective phase of the 10-year follow-up of the SAS. We included 1,811 participants aged 60 years or older who completed both an olfactory identification test and a cognitive assessment at baseline (2010-2011). Survival status of the participants from baseline to December 31, 2019 was obtained from the local mortality surveillance system. We used the four-way decomposition method to attribute effects to interaction and mediation and to explore the mediation effect of NDDs on the OI-mortality relationship.Results: The four-way decomposition method revealed a statistically significant association of OI with death. Overall, 43% higher risk for death was associated with OI [excess relative risk (ERR) = 0.43, 95% CI: 0.06-0.80, p = 0.023]. Excluding the mediation from NDDs and interaction between OI and NDDs, the controlled direct effect of OI on death was even higher in NDDs participants, with an ERR of 77% (95% CI: 0.00-1.55, p = 0.050). Statistically significant association was found for failure to identify coffee (ERR = 0.77, 95% CI: 0.18-1.36, p = 0.010) and marginally significant associations were found for failure to identify cinnamon (ERR = 0.33, 95% CI: -0.02-0.68, p = 0.068) and rose (ERR = 0.33, 95% CI: -0.01-0.67, p = 0.054) with death.Conclusion: OI was associated with the long-term mortality in older adults and the association was even stronger in those with NDDs. Failure to identify coffee or rose was associated with a higher mortality risk, and the association was mediated by NDDs.
|
|
| 70. |
- Cao, Yang, Associate Professor, 1972-, et al.
(författare)
-
Postnatal cadmium exposure, neurodevelopment, and blood pressure in children at 2, 5, and 7 years of age
- 2009
-
Ingår i: Journal of Environmental Health Perspectives. - : National Institute of Environmental Health Science. - 0091-6765 .- 1552-9924. ; 117:10, s. 1580-1586
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND: Adverse health effects of cadmium in adults are well documented, but little is known about the neuropsychological effects of cadmium in children, and no studies of cadmium and blood pressure in children have been conducted.OBJECTIVE: We examined the potential effects of low-level cadmium exposure on intelligence quotient, neuropsychological functions, behavior, and blood pressure among children, using blood cadmium as a measure of exposure.METHODS: We used the data from a multicenter randomized clinical trial of lead-exposed children and analyzed blood cadmium concentrations using the whole blood samples collected when children were 2 years of age. We compared neuropsychological and behavioral scores at 2, 5, and 7 years of age by cadmium level and analyzed the relationship between blood cadmium levels at 2 years of age and systolic and diastolic blood pressure at 2, 5, and 7 years of age.RESULTS: The average cadmium concentration of these children was 0.21 microg/L, lower than for adults in the National Health and Nutrition Examination Survey (NHANES), but comparable to concentrations in children < 3 years of age in NHANES. Except for the California Verbal Learning Test for Children, there were no differences in test scores among children in different cadmium categories. For children with detectable pretreatment blood cadmium, after adjusting for a variety of covariates, general linear model analyses showed that at none of the three age points was the coefficient of cadmium on Mental Development Index or IQ statistically significant. Spline regression analysis suggested that behavioral problem scores at 5 and 7 years of age tended to increase with increasing blood cadmium, but the trend was not significant. We found no significant associations between blood cadmium levels and blood pressure.CONCLUSION: We found no significant associations between background blood cadmium levels at 2 years of age and neurodevelopmental end points and blood pressure at 2, 5, and 7 years of age. The neuropsychological or hypertensive effects from longer background exposures to cadmium need further study.
|
|
