Clinical Decision Biological Biomarker and Prognosis Prediction of Hepatocellular Carcinoma by Deep Learning

A Deep Learning Model Based on Contrast-enhanced Ultrasound to Aid Clinical Decisions and Predict Biological Biomarker and Prognosis of Hepatocellular Carcinoma

Developing a deep learning model based on contrast-enhanced ultrasound (CEUS) to predict the prognosis of hepatocellular carcinoma (HCC) and aid choose operation decisions

Collecting CEUS and clinical data of HCC from different institutions retrospectively.

Developing a deep learning model based on CEUS to predict the prognosis of HCC. Developing a deep learning model based on CEUS to choose a better operation (ablation or surgery) of HCC patients.

Then, validating the deep learning model in the prospective data.

Inclusion Criteria:

• patients with HCC (Ia, Ib, IIa stage) China liver cancer staging who underwent resection or ablation
• without macro-vascular invasion
• Child-Pugh A/B grade
• HCC is proved by pathological examination or two enhanced imaging
• CEUS (Sonovue or Sonozoid) images are performed two weeks before the operation
• Invasive biomarker or prognosis of HCC available
• CEUS images are included in at least three stages (Arterial phase, Portal phase, and Late phase)

Exclusion Criteria:

• postop follow-up loss or expired less than 3 months
• patients with co-malignancy
• poor images quality for analyzing