In Vivo T2 mapping for early treatment response assessment in esophageal cancer: a prospective study

To evaluate the feasibility and reproducibility of in vivo T2 mapping in esophageal cancer (EC) and explore its role in early treatment response assessment during nonsurgical multimodal therapy. This prospective study consecutively enrolled patients with pathologically confirmed EC who were undergoing nonsurgical multimodal therapy from February 2024 to February 2026. In vivo T2 mapping MRI was performed before treatment initiation and again prior to the second treatment cycle. Image quality and measurement reproducibility of T2 mapping were evaluated. Quantitative parameters, including pre- and post-treatment T2 values, absolute change (ΔT2), and percentage change (ΔT2%), were calculated. Apparent diffusion coefficient (ADC)-derived parameters were also assessed as established diffusion-based quantitative imaging biomarkers for comparison and multiparametric analysis. Treatment response was assessed using predefined clinical response criteria, and patients were classified into responder and non-responder groups. Between-group differences in T2-derived parameters were analyzed. Receiver operating characteristic (ROC) curve analysis and multivariable logistic regression were performed to evaluate treatment response using clinical-adjusted and imaging-based models. Thirty-four patients were identified as responders (CR/good PR), and twenty-two as non-responders (IR/SD/PD) based on clinical response evaluation. T2 values decreased significantly after early treatment compared with baseline ( p < 0.05). The magnitudes of ΔT2 and ΔT2% were significantly greater in responders than in non-responders. No significant differences were observed in absolute T2 values between the two groups either before or after treatment. The area under the curve (AUC) was 0.738 for ΔT2 and 0.754 for ΔT2% in predicting treatment response. The optimal cutoff values for ΔT2 and ΔT2% were − 5.44 ms and − 6.52%, respectively. In the clinical-adjusted model, ΔT2% remained significantly associated with treatment response after adjustment for age, sex, tumor location, and American Joint Committee on Cancer (AJCC) clinical stage. The imaging-based model showed a higher AUC than ΔADC alone (0.834 vs. 0.823), but the difference was not significant (DeLong test, p = 0.554). In vivo T2 mapping demonstrated good measurement reproducibility in EC. Both ΔT2 and ΔT2% were associated with treatment response, whereas the incremental value of combining ΔT2% with ΔADC requires further investigation.