HCC - Hepatocellular Carcinoma, Microvascular Invasion (MVI)
Conditions
Keywords
hepatocellular carcinoma, microvascular invasion, contrast-enhanced ultrasound, deep learning
Brief summary
An artificial intelligence (AI) model to predict MVI of HCC using contrast-enhanced ultrasound was constructed. This model also has biological explainability. The investigators named it as MAPUSE (MVI AI prediction via contrast-enhanced ultrasound with explainability). The goal of MAPUSE study is to prospectively test the performance of MAPUSE model on MVI prediction and its biological correlation in different geographical areas of China.
Detailed description
The presence of microvascular invasion (MVI) in hepatocellular carcinoma (HCC) is a critical prognostic indicator, but its preoperative diagnosis remains challenging. Contrast-enhanced ultrasound (CEUS), with its dynamic microvascular imaging capability, holds promise in prediction of MVI. The investigators constructed an artificial intelligence (AI) model to predict MVI using contrast-enhanced ultrasound. This model also has biological explainability. We named it as MAPUSE (MVI AI prediction via contrast-enhanced ultrasound with explainability). The goal of MAPUSE study is to prospectively test the performance of MAPUSE model on MVI prediction and its biological correlation in different geographical areas of China. The performance of MAPUSE is to be tested in two prospective testing cohorts from two centers in southern and northern China. Before surgery, patient CEUS videos will be collected and analysed by MAPUSE model to generate an MVI risk score. According to the postoperative pathological diagnosis of MVI (golden criterion), the result of MAPUSE will be evaluated. Parameters include area under curve (AUC), accuracy (ACC), sensitivity, specificity and F1-score.
Interventions
DIAGNOSTIC_TESTthe MAPUSE model
Using the MAPUSE model to predict MVI status before surgical resection for HCC patients
Sponsors
Chinese Academy of Sciences
Chinese PLA General Hospital
Study design
Observational model
COHORT
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
1. Age \>18 years old. 2. The HCC diagnosis and the presence of MVI were confirmed by surgical pathology. 3. Complete and clear CEUS videos obtained within two weeks preoperatively.
Exclusion criteria
1. Unqualified CEUS images. 2. Missing surgical pathological diagnosis. 3. Lesions underwent local treatments. 4. Non-HCC diagnosis
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| area under operating characteristic curves (AUC) | From preoperative enrollment to the postoperative confirmation of pathological diagnosis (7-15 days postopertively) | the area under operating characteristic curves (AUC) to evaluate the performance of MAPUSE model in predicting MVI in HCC patients |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| ACC (accuracy) | From preoperative enrollment to the postoperative confirmation of pathological diagnosis (7-15 days postopertively) | The ratio of the number of samples correctly predicted by the model to the total number of samples |
| Specificity | From preoperative enrollment to the postoperative confirmation of pathological diagnosis (7-15 days postopertively) | Proportion of all patients without MVI who are predicted negative by MAPUSE |
| Sensitivity | From preoperative enrollment to the postoperative confirmation of pathological diagnosis (7-15 days postopertively) | The proportion of patients with MVI that MAPUSE correctly identifies |
Countries
China
Outcome results
None listed