JMIR Publications

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 the present study. The patients who underwent a bariatric procedure between 2010 and 2014 were used as training data, and those 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 of the patient 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). Except for the outcome variable, 16 features were used for machine learning, including five continuous features and eleven categorical features. The continuous features were standardized to have the mean of 0 and a standard deviation of 1 before they enter the model. The synthetic minority oversampling technique (SMOTE) was used to artificially augment the patients with serious complications. The K-fold cross-validation method was used during the training phase. The performances of the neural networks were evaluated using accuracy, sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve. The exhaustive grid search method was used for hyperparameter tuning.

Results:

In total, 37,811 and 6,250 patients were used as the training data and test data, with the incidence rates of serious complication 3.2% and 3.0%, respectively. When trained using the SMOTE data, the MLP appeared to have a desirable performance with an area under curve (AUC) of 0.84. However, its performance was low for the test data with an AUC of 0.54. The performance of CNN appeared similar to MLP. It generated AUCs of 0.79 and 0.57 for the SMOTE data and test data, respectively. Compared with the MLP and CNN, the RNN showed worse performance. AUCs of RNN for the SMOTE data and test data were 0.65 and 0.55, respectively.

Conclusions:

DLNNs such as MLP and CNN may improve the predictability of the postoperative serious complications after bariatric surgery using SOReg data. However, an overfitting issue is still apparent and needs to be overcome by incorporating intra- and perioperative information. Clinical Trial: NA.