Skip to main content

Table 6 The validation results of the proposed method by randomly using a proportion of the patient data as the independent testing data and using the rest of the patient data to establish the corresponding machine-learning model

From: Ultrasonic liver steatosis quantification by a learning-based acoustic model from a novel shear wave sequence

Proportion of the independent testing data in the entire patient data (%)

Correlation coefficients between the model’s estimation results and the liver fat proportion from MR mDIXON quantification

P value

5

0.8249 (95% CI 0.7223 to 0.8920)

< 0.00001

10

0.8049 (95% CI 0.6925 to 0.8791)

< 0.00001

15

0.7816 (95% CI 0.7173 to 0.8326)

< 0.00001