Dataset Information

Application of data mining for predicting hemodynamics instability during pheochromocytoma surgery.

ABSTRACT:

Background

Surgical resection of pheochromocytoma may lead to high risk factors for intraoperative hemodynamic instability (IHD), which can be life-threatening. This study aimed to investigate the risk factors that could predict IHD during pheochromocytoma surgery by data mining.

Method

Relief-F was used to select the most important features. The accuracies of seven data mining models (CART, C4.5, C5.0, and C5.0 boosted), random forest algorithm, Naive Bayes and logistic regression were compared, the cross-validation, hold-out, and bootstrap methods were used in the validation phase. The accuracy of these models was calculated independently by dividing the training and the test sets. Receiver-Operating Characteristic curves were used to obtain the area under curve (AUC).

Result

Random forest had the highest AUC and accuracy values of 0.8636 and 0.8509, respectively. Then, we improved the random forest algorithm according to the classification of imbalanced data. Improved random forest model had the highest specificity and precision among all algorithms, including relatively higher sensitivity (recall) and the highest f1-score integrating recall and precision. The important attributes were body mass index, mean age, 24?h urine vanillylmandelic acid/upper normal limit value, tumor size and enhanced computed tomography difference.

Conclusions

The improved random forest algorithm may be useful in predicting IHD risk factors in pheochromocytoma surgery. Data mining technologies are being increasingly applied in clinical and medical decision-making, and provide continually expanding support for the diagnosis, treatment, and prevention of various diseases.

SUBMITTER: Zhao Y

PROVIDER: S-EPMC7370474 | biostudies-literature | 2020 Jul

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Application of data mining for predicting hemodynamics instability during pheochromocytoma surgery.

Zhao Yueyang Y Fang Li L Cui Lei L Bai Song S

BMC medical informatics and decision making 20200720 1

<h4>Background</h4>Surgical resection of pheochromocytoma may lead to high risk factors for intraoperative hemodynamic instability (IHD), which can be life-threatening. This study aimed to investigate the risk factors that could predict IHD during pheochromocytoma surgery by data mining.<h4>Method</h4>Relief-F was used to select the most important features. The accuracies of seven data mining models (CART, C4.5, C5.0, and C5.0 boosted), random forest algorithm, Naive Bayes and logistic regressio ...[more]

PMID: 32690077

Similar Datasets

Project description:BackgroundMaintaining intraoperative haemodynamic stability can reduce cardiovascular complications during surgery for pheochromocytoma and paraganglioma (PPGL). Risk factors such as tumour size and catecholamine levels are reported to predict haemodynamic responses during surgery for PPGL. We hypothesized that additional factors including body composition and genetic information could further improve prediction.MethodsConsecutive patients with PPGL confirmed by surgical pathology between June 2010 and June 2019 were retrospectively included. Cross-sectional computed tomography images at the L3 level were used to assess body composition parameters including skeletal muscle area and visceral fat area. Next-generation sequencing was performed using a panel containing susceptibility genes of PPGL. Differences in clinical-genetic characteristics and body composition parameters were analysed and compared in patients with and without intraoperative haemodynamic instability (HDI).ResultsWe included 221 patients with PPGL (median age 47 [38-56] years, and 52% male). Among them, 49.8% had Cluster 2 mutations (related to kinase signalling pathways), 44.8% had sarcopenia, and 52.9% experienced intraoperative HDI. Compared with patients without HDI, more patients with HDI had Cluster 2 mutations (59.8% vs. 38.5%, P = 0.002) and less had sarcopenia (35.9% vs. 54.8%, P = 0.005). Multivariate analysis showed that urine vanillylmandelic acid ≥ 58 μmol/day (adjusted odds ratio [OR] = 1.840, 95% confidence interval [CI] = 1.012-3.347, P = 0.046), tumour size ≥ 4 cm (adjusted OR = 2.278, 95% CI = 1.242-4.180, P = 0.008), and Cluster 2 mutations (adjusted OR = 2.199, 95% CI = 1.128-4.285, P = 0.021) were independent risk factors for intraoperative HDI, while sarcopenia (adjusted OR = 0.475, 95% CI = 0.266-0.846, P = 0.012) decreased the risk.ConclusionsBody composition and genotype were associated with intraoperative haemodynamics in patients with PPGL. Our results indicated that inclusion of body composition and genotype in the overall assessment of patients with PPGL helped to predict HDI during surgery, which could assist in implementing preoperative and intraoperative measures to reduce perioperative complications.

Project description:Congenital heart defects (CHD) is one of the most common birth defects in China. Many studies have examined risk factors for CHD, but their predictive abilities have not been evaluated. In particular, few studies have attempted to predict risks of CHD from, necessarily unbalanced, population-based cross-sectional data. Therefore, we developed and validated machine learning models for predicting, before and during pregnancy, women's risks of bearing children with CHD. We compared the results of these models in a large-scale, comprehensive population-based retrospective cross-sectional epidemiological survey of birth defects in six counties in Shanxi Province, China, covering 2006 to 2008. This contained 78 cases of CHD among 33831 live births. We constructed nine synthetic variables to use in the models: maternal age, annual per capita income, family history, maternal history of illness, nutrition and folic acid deficiency, maternal illness in pregnancy, medication use in pregnancy, environmental risk factors in pregnancy, and unhealthy maternal lifestyle in pregnancy. The machine learning algorithms Weighted Support Vector Machine (WSVM) and Weighted Random Forest (WRF) were trained on, and a logistic regression (Logit) was fitted to, two-thirds of the data. Their predictive abilities were then tested in the remaining data. True positive rate (TPR), true negative rate (TNR), accuracy (ACC), area under the curves (AUC), G-means, and Weighted accuracy (WTacc) were used to compare the classification performance of the models. Median values, from repeating the data partitioning 1000 times, were used in all comparisons. The TPR and TNR of the three classifiers were above 0.65 and 0.93, respectively, better than any reported in the literature. TPR, wtACC, AUC and G were highest for WSVM, showing that it performed best. All three models are precise enough to identify groups at high risk of CHD. They should all be considered for future investigations of other birth defects and diseases.

Dataset Information

Application of data mining for predicting hemodynamics instability during pheochromocytoma surgery.

Background

Method

Result

Conclusions

Publications

Application of data mining for predicting hemodynamics instability during pheochromocytoma surgery.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets