2075 - CT-Based Radiomic T-cell Inflamed Signature Refines Risk Stratification of Local Recurrence and Guides Immunotherapy Selection in Esophageal Cancer Treated with Definitive Chemoradiotherapy
Presenter(s)
J. Gong, M. Liu, and L. Zhao; Department of Radiation Oncology, First Affiliated Hospital of Air Force Medical University, Xi'an, Shaanxi, China
Purpose/Objective(s): The tumor immune microenvironment (TIME) critically influences cancer outcomes. While a previous study developed a CT-based radiomic signature of T-cell inflamed TIME (CT-TIME) based on pan-cancer cohorts, its prognostic value in esophageal squamous cell carcinoma (ESCC) remains unproven. Given ESCC’s high recurrence rates after definitive chemoradiotherapy (dCRT) and emerging immunotherapy combinations, validating CT-TIME could promote personalized treatment strategies.
Materials/Methods: We evaluated a pre-established CT-TIME in 244 ESCC patients undergoing dCRT (dCRT cohort) and 13 patients receiving combined immunotherapy and dCRT (I-dCRT cohort). Dynamic CT-TIME subtypes (Type1-3) were derived from pretreatment and posttreatment contrast-enhanced CT (CECT) in dCRT cohort. Associations with local recurrence-free survival (LRFS) were assessed using Kaplan-Meier analysis and Cox regression. An integrated clinic-radiomic nomogram (CliTIME) was developed and validated for predicting LRFS in ESCC treated with dCRT. CT-TIME status was derived from pretreatment CECT in I-dCRT cohort to evaluate its association with LRFS in ESCC treated with I-dCRT.
Results: In dCRT cohort, patients with persistent immune-inflamed CT-TIME subtype (Type3) had significantly prolonged LRFS (3-year rate: 74.2% vs. 35.2% in uninflamed; P < 0.0001). The CliTIME nomogram, combining dynamic CT-TIME subtype and clinical factors (T-stage, tumor length, concurrent chemotherapy), improved LRFS prediction (C-index: 0.701 vs. 0.668 for clinical model; P < 0.05). In I-dCRT cohort, pretreatment immune-inflamed CT-TIME status predicted superior outcomes (3-year LRFS: 60.0% vs. 12.5%; P = 0.019).
Conclusion: The CT-TIME signature noninvasively predicts LRFS and may identify ESCC patients likely to benefit from immunotherapy escalation. Dynamic monitoring of TIME could guide personalized therapeutic strategies.