Meta-Analysis of MRI in Predicting Early Response to Radiotherapy and Chemotherapy in Esophageal Cancer
Xinyu Li,Fang Yuan,Li Ni,Xiaopan Li
DOI: https://doi.org/10.1016/j.acra.2024.08.055
2024-09-11
Abstract:Rationale and objectives: At present, the application of magnetic resonance imaging (MRI) in the prediction of response to neoadjuvant therapy and concurrent chemoradiotherapy for the treatment of esophageal cancer still needs to be further explored, and its early differential value remains controversial, thus we carried out this systematic review with a meta-analysis. In the application, different MRI sequences and corresponding parameters are used for the differential diagnosis of the response to neoadjuvant therapy and concurrent chemoradiotherapy. Methods: All relevant studies evaluated the efficacy and response to MRI in neoadjuvant therapy or concurrent chemoradiotherapy for esophageal cancer on Pubmed, Embase, Cohrane Library, and Web of Science databases published before October 10, 2023 (inclusive) were systematically searched. A revised tool was used to assess the quality of diagnostic accuracy studies (QUADAS-2) to assess the risk of bias in the included original studies. A subgroup analysis of MRI sequences diffusion weighted imaging (DWI), dynamic contrast enhanced (DCE) and their corresponding different parameters, as well as the acquisition timepoints (before and after treatment) for different parameters, was performed during the meta-analysis. The bivariate mixed-effects model was used for meta-analysis. Results: 21 studies were finally included, involving 1128 patients with esophageal cancer. The sensitivity, specificity, and area under receiver operating characteristic curve (ROC curve) of DWI sequence for identifying response to concurrent chemoradiotherapy were 0.82 (95% CI: 0.74-0.87), 0.81 (95% CI: 0.72-0.87) and 0.88 (95% CI: 0.56-0.98), respectively. The sensitivity, specificity, and area under ROC curve of DCE sequence for identifying response to concurrent chemoradiotherapy were 0.78 (95% CI: 0.70-0.84), 0.65 (95% CI: 0.59-0.70) and 0.73 (95% CI: 0.50-0.88), respectively. In patients with esophageal cancer, the sensitivity, specificity, and area under the ROC curve of DWI sequences for identifying response to neoadjuvant therapy were 0.80 (95% CI: 0.69 - 0.88), 0.81 (95% CI: 0.69 - 0.89), and 0.88 (95% CI: 0.34 - 0.99), respectively; the sensitivity, specificity, and area under the ROC curve of DCE sequences for identifying response to neoadjuvant therapy were 0.84 (95% CI: 0.76 - 0.90), 0.61 (95% CI: 0.53 - 0.68), and 0.70 (95% CI: 0.27 - 0.94), respectively. Conclusions: Based on the available evidence, MRI had a very good value in the early identification of response to neoadjuvant therapy and concurrent chemoradiotherapy for esophageal cancer, especially DWI. Apparent diffusion coefficient (ADC) value changes before and after treatment could be used as predictors of pathological response. Also, ADC value changes before and after treatment could be used as a tool to guide clinical decision-making.