Could 18-FDG PET-CT Radiomic Features Predict the Locoregional Progression-Free Survival in Inoperable or...

Area of Science:

• Oncology
• Radiology
• Medical Imaging Analysis

Background:

• Esophageal cancer poses significant challenges, particularly for inoperable or unresectable cases.
• Predictive biomarkers are crucial for tailoring treatment and improving patient outcomes.
• Positron-emission tomography−computed tomography (PET-CT) offers valuable imaging insights.

Purpose of the Study:

• To evaluate the predictive value of pre-treatment PET-CT-based radiomic features for locoregional progression-free survival (LR-PFS).
• To identify specific radiomic signatures associated with treatment response in esophageal cancer.

Main Methods:

• Retrospective analysis of 46 patients with inoperable/unresectable esophageal cancer.
• Extraction and principal component analysis (PCA) of 230 radiomic parameters from pre-treatment PET-CT scans.
• Correlation analysis between radiomic features and LR-PFS, followed by hierarchical clustering.

Main Results:

• Two-year LR-PFS and overall PFS rates were 35.9% and 21.6%, respectively.
• Hierarchical clustering identified two patient groups with significantly different median LR-PFS (22.8 vs. 9.9 months).
• Two radiomic features, 'F_rlm_rl_entr_per' and 'F_rlm_2_5D_rl_entr', were significantly associated with LR-PFS, with lower values correlating to better outcomes.

Conclusions:

• Pre-treatment radiomic features derived from PET-CT are valuable predictors of LR-PFS in esophageal cancer.
• Specific radiomic signatures ('F_rlm_rl_entr_per' and 'F_rlm_2_5D_rl_entr') can identify patients at lower risk of locoregional relapse after chemoradiotherapy (CRT).
• These findings support the potential integration of radiomics into treatment planning for esophageal cancer.