Skip to content

A Multi-Task Deep Learning Model Based on Reinforcement Learning and Image Segmentation for Differentiating Benign and Malignant Liver Nodules, Identifying Microinvasion in Liver Cancer, and Predicting Liver Cancer Prognosis

A Multi-Task Deep Learning Model Based on Reinforcement Learning and Image Segmentation for Differentiating Benign and Malignant Liver Nodules, Identifying Microinvasion in Liver Cancer, and Predicting Liver Cancer Prognosis

Status
Active, not recruiting
Phases
Unknown
Study type
Observational
Source
ChiCTR
Registry ID
ChiCTR2600119850
Enrollment
Unknown
Registered
2026-03-04
Start date
2025-06-10
Completion date
Unknown
Last updated
2026-03-09

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Liver Nodules

Interventions

Case group:This is an observational study without interventions.

Sponsors

Sun Yat-sen University Cancer Center
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to No maximum

Inclusion criteria

Inclusion criteria: Inclusion criteria: 1. Adults aged 18 years or older; 2. Patients with liver nodules detected by imaging or diagnosed with liver cirrhosis; 3. Availability of at least one analyzable liver MR image; 4. Diagnosis of complex liver nodules that are difficult to differentiate on imaging, including high-grade dysplastic nodules (HGDN), well-differentiated hepatocellular carcinoma (WD-HCC), low-grade dysplastic nodules (low-grade DN), focal nodular hyperplasia (FNH), and hepatocellular adenoma.

Exclusion criteria

Exclusion criteria: Exclusion criteria: 1. Liver MR images that do not meet the minimum quality requirements; 2. Patients with isolated cirrhotic nodules without other types of liver nodules or relevant imaging features; 3. Lack of contemporaneous clinical information or medical history close to the imaging time; 4. Presence of other major comorbidities that may interfere with liver disease assessment; 5. History of liver transplantation or major liver surgery; 6. Patients unwilling or unable to comply with follow-up procedures.

Design outcomes

Primary

MeasureTime frame
The accuracy of the AI model in differentiating complex liver nodules.;

Countries

China

Contacts

Public ContactZhongguo Zhou

Sun Yat-sen University Cancer Center

tianxp@sysucc.org.cn+86 20 8734 3355

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Mar 14, 2026