A New Multimodel Machine Learning Framework to Improve Hepatic Fibrosis Grading Using Ultrasound Elastography Systems from Different Vendors.

The purpose of the work described here was to determine if the diagnostic performance of point and 2-D shear wave elastography (pSWE; 2-DSWE) using shear wave velocity (SWV) with a new machine learning (ML) technique applied to systems from different vendors is comparable to that of magnetic resonance elastography (MRE) in distinguishing non-significant (<F2) from significant (≥F2) fibrosis. We included two patient groups with liver disease: (i) 144 patients undergoing pSWE (Siemens) and MRE; and (ii) 60 patients undergoing 2-DSWE (Philips) and MRE. Four ML algorithms using 10 SWV measurements as inputs were trained with MRE. Results were validated using twofold cross-validation. The performance of median SWV in binary grading of fibrosis was moderate for pSWE (area under the curve [AUC]: 0.76) and 2-DSWE (0.84); the ML algorithm support vector machine (SVM) performed particularly well (pSWE: 0.96, 2-DSWE: 0.99). The results suggest that the multivendor ML-based algorithm SVM can binarily grade liver fibrosis using ultrasound elastography with excellent diagnostic performance, comparable to that of MRE.