A Refined Prognostic Model for Postoperative Overall Survival in Hepatocellular Carcinoma Based on CODEX-Based...

Area of Science:

• Oncology
• Medical Imaging
• Proteomics

Background:

• Hepatocellular carcinoma (HCC) is a major global health concern.
• Accurate prognosis prediction is crucial for effective patient management after surgical resection.
• Current prognostic models often lack comprehensive integration of multi-modal data.

Purpose of the Study:

• To develop and validate a predictive model for overall survival (OS) in HCC patients post-resection.
• To compare the performance of clinicopathological, radiomics, proteomics, and combined models.
• To establish a foundation for personalized HCC diagnosis and treatment.

Main Methods:

• Eighty-two HCC patients were divided into training (n=62) and validation (n=20) cohorts.
• Clinicopathological, CO-Detection by Indexing (CODEX)-based multiproteomics, and MRI-derived radiomics features were utilized.
• Four models (clinicopathological, radiomics, proteomics, combined) were constructed and evaluated using C-index, ROC curves, and DCA.

Main Results:

• The combined model integrating all features showed superior OS prediction in both cohorts (Training C-index=0.821, Validation C-index=0.791).
• Calibration curves indicated high accuracy for the combined nomogram in predicting OS.
• The combined model outperformed single-modality models in prognostic prediction.

Conclusions:

• The integrated model combining clinicopathological, radiomics, and CODEX-based multiproteomics data offers enhanced prognostic prediction for HCC.
• This multi-modal approach provides a stronger basis for personalized HCC patient care.
• Further large-scale, multi-center validation is necessary to confirm clinical utility.