Added value of shear wave viscoelasticity imaging, homodyned-K tissue imaging and acoustic attenuation to assess liver cancer at ultrasound: A multiparametric machine learning approach


Hepatocellular carcinoma (HCC) accounts for 85% of cancers that arise from the liver. HCC is the sixth most common cancer and the second leading cause of cancer mortality worldwide. Risk factors for HCC include cirrhosis, chronic infection with hepatitis B and C viruses, excess alcohol consumption, nonalcoholic fatty liver disease, family history of HCC, obesity, type 2 diabetes mellitus, and smoking. HCC detection at an early stage, through surveillance, and curative therapy has considerably improved the 5-year survival. Therefore, medical societies recommend systematic surveillance with ultrasound every 6 Page 13 / 51 months in at-risk cirrhotic patients or non-cirrhotic hepatitis B carriers. Ultrasound is recommended on the basis of its cost-effectiveness and wide availability for large-scale screening. However, a key challenge is that ultrasound may miss 20-40% of liver cancers and diagnosis must rely on additional tests (e.g., magnetic resonance imaging or biopsy). Hence, there is an urgent need to improve characterization and early detection of HCC. Members of our team have developed advanced ultrasound methods that provide unique information not available on clinical ultrasound scanners. This research proposes approaches that are based on liver mechanical and structural properties obtained by analysing experimental ultrasound echoes with novel algorithms. Improving detection of HCC is critical because HCC can be treated curatively only if detected at an early stage and leads to improvements in survival rates in patients enrolled in surveillance programs. This study is timely and of high importance because HCC is the only cancer in Canada for which mortality is increasing.