Noninvasive imaging-based machine learning algorithm to identify progressive disease in advanced hepatocellular
1Department of Medical Oncology, Nanyang Second People's Hospital, Nanyang, China.
Scientific Reports
|July 1, 2023
Summary
Radiomics analysis accurately predicts treatment efficacy for advanced hepatocellular carcinoma (HCC) using tyrosine kinase inhibitors plus anti-PD-1 antibodies. The support vector machine model demonstrated high predictive accuracy, aiding clinical decision-making.
Area of Science:
- Medical Imaging
- Oncology
- Radiomics
Background:
- Advanced hepatocellular carcinoma (HCC) presents treatment challenges.
- Predicting treatment response to tyrosine kinase inhibitors (TKI) plus anti-PD-1 antibodies is crucial for second-line therapy.
Purpose of the Study:
- To predict the efficacy of TKI plus anti-PD-1 antibodies (TKI-PD-1) in advanced HCC using radiomics.
- To develop and validate predictive models based on pre-treatment CT imaging features.
Main Methods:
- Radiomic features were extracted from pre-treatment CT scans of 55 advanced HCC patients.
- Feature selection was performed using intraclass correlation coefficients (ICCs) and least absolute shrinkage and selection operator (LASSO).
- Ten machine learning algorithms were developed and validated, with performance assessed by AUC, Kaplan-Meier, and Cox regression analyses.
Main Results:
- The support vector machine (SVM) model achieved the highest predictive accuracy with an AUC of 0.933 (training) and 0.792 (testing).
- Selected radiomic features were significantly associated with patient overall survival.
- 18 out of 55 patients (32.7%) experienced disease progression.
Conclusions:
- Radiomics analysis, particularly using the SVM algorithm, effectively predicts TKI-PD-1 efficacy in advanced HCC.
- Pre-treatment imaging features hold significant value for predicting treatment outcomes in HCC.
- This approach can aid in personalized treatment strategies for advanced HCC.
