Table 2.

Summary of the application of AI in the renal ultrasound for prediction of the kidney function.

Task	Algorithms	Data source	Size	Results	Ref.
Prediction of the kidney function	Decisive area-proportional, textural features and SVM techniques	non-diabetics, non- acute renal failure, non-polycystic kidney disease, non-hydronephrosis, non-inpatient, and age between 18 to 75 years old	798 images	left-kidney: ACC=71% right-kidney: ACC=76% combining: ACC=70%	(33)
	The neural network architecture consists of 33 residual blocks as CNN-based feature extractors, and three fully connected layers of 512, 512, and 256 neurons as regressors	the view of the maximum observable kidney length	4505 images	ACC=86%	(34)
	5 ML algorithms (Nu-Support Vector Classification, C- Support Vector Classification, Random Forest, Adaptive boosting, and Xtreme gradient boosting) based on radiomics	clinical medical records of patients who underwent renal transplantation	233 patients	AUC=0.79-0.84	(35)

SVM, support vector machine; CNN, convolutional neural network; ML, machine learning; ACC, accuracy; AUC, the area under the receiver operating characteristic curve.