The precise treatment of primary hepatocellular carcinoma (HCC) highly depends on accurate disease staging (CNLC, TNM, BCLC) and scientific treatment decision-making, which necessitate the integration of both imaging and clinical baseline data. This study prospectively recruits HCC patients and clinical physicians across different hospital tiers to evaluate the clinical value of a self-developed artificial intelligence (AI) model in assisting multi-dimensional comprehensive assessment and treatment decision-making. Utilizing a Multi-Rater Multi-Case (MRMC) crossover balanced design, the study compares the accuracy of clinical evaluations performed by physicians under "unassisted (without AI)" versus "AI-assisted" conditions. A key focus is to explore whether AI can significantly enhance the comprehensive assessment capabilities of physicians in primary/secondary care hospitals, thereby prospectively reducing diagnostic and therapeutic heterogeneity across different institutional levels.
Age >= 18 years.
Patients prospectively presenting with suspected or newly diagnosed primary hepatocellular carcinoma (HCC) later confirmed by pathology or meeting the China Liver Cancer (CNLC) guidelines.
Complete baseline clinical data acquired during the prospective enrollment period, including complete history of present/past illness, ECOG PS score, comprehensive laboratory tests (liver function, coagulation, tumor markers such as AFP, etc.), and baseline abdominal contrast-enhanced CT.
Patients (or their legal representatives) must provide written informed consent for their clinical data to be used in this trial.
Patients with secondary (metastatic) liver cancer or concurrent severe malignancies of other systems.
Patients who fail to complete the required baseline imaging or laboratory tests, preventing accurate staging calculation (e.g., missing data for Child-Pugh score).
Patients who have previously received anti-tumor therapies for liver cancer prior to enrollment.
The study population comprises 108 prospectively and consecutively enrolled adult patients with newly diagnosed primary hepatocellular carcinoma (HCC). Following enrollment and confirmation of complete baseline clinical and imaging data, these 108 patient cases are randomly divided into two equal datasets: Set A (54 cases) and Set B (54 cases). Randomization is stratified to ensure no statistically significant differences between the two sets regarding baseline characteristics such as tumor burden, liver function grading, and staging distribution. In the context of this multi-rater multi-case (MRMC) crossover design, these patient cases are allocated to distinct evaluation conditions. Set A cases are assigned to be evaluated by the first group of reviewing physicians without AI assistance (control condition) and by the second group of physicians with AI assistance (experimental condition). Conversely, Set B cases are evaluated by the second group of physicians without AI
Cohort
Prospective
Unassisted Independent Evaluation
Diagnostic testPhysicians independently evaluate the HCC cases and provide staging and treatment decisions using only complete clinical baseline data and imaging data, without any assistance from the AI model.
AI-Assisted Evaluation
Diagnostic testPhysicians evaluate the HCC cases and provide final staging and treatment decisions after reviewing the initial predictions and related evidence generated by the self-developed artificial intelligence (AI) model, alongside the clinical baseline and imaging data.
Improvement in Overall Accuracy
The difference in average accuracy across the 4 classification tasks between AI-assisted evaluation (experimental) and independent evaluation (control). Accuracy is determined by comparing physicians' predictions against the reference standard (Ground Truth) established by the independent expert panel
Homogenization Effect on Evaluation Accuracy
Assessment of whether the difference in clinical evaluation accuracy between physicians in the primary/secondary hospital group and the tertiary hospital groups is significantly reduced under AI assistance compared to unassisted independent evaluation.
Evaluation Efficiency (Average Time per Case)
Comparison of the average evaluation time (e.g., measured in minutes) per case required by participating physicians when utilizing AI assistance versus performing unassisted independent evaluation.
Inter-rater Agreement
Comparison of the consistency of evaluation results (staging and treatment decisions) among all participating physicians, assessed using appropriate statistical measures (e.g., Kappa statistics), under AI-assisted versus unassisted conditions.
Click on the lead sponsor to view all of their trials.
Beijing Tsinghua Chang Gung Hospital
Lead sponsor
Affiliated Hospital of Hebei University
Collaborator
Meng Chao Hepatobiliary Hospital of Fujian Medical University
Collaborator
Zhongnan Hospital
Collaborator
Xingtai People's Hospital
Collaborator
Second Affiliated Hospital of Xi'an Jiaotong University
Collaborator
Xinan hospital of Army Medical University
Collaborator
