用DWMRI预测食管癌原发灶放化疗疗效的临床研究

目的 探讨DWMRI参数早期预测食管癌放化疗反应，寻找优选的预测时间点及预测阈值以指导临床。方法 2010—2011年间连续性入组38例食管鳞癌患者，均行3DRT 60 Gy分30次6周完成，辅以同期或巩固化疗(FP或TP方案)。患者于放疗前、放疗开始后第1—6周末共7个时间点行DWMRI扫描，对扫描所获ADC值及DWMRI序列肿瘤长度进行观察和分析。用成组t检验、重复测量的方差分析和多变量方差分析，用Logistic模型预后分析，用ROC曲线进行预测效能判断和阈值筛选。结果 全组CR 20例(53%)，PR 18例(47%);二者放疗前及第1—6周末的ADC值分别为1.82、1.92、2.06、2.35、2.62、2.71、2.96×10^-3 mm²/s和1.42、1.49、1.67、1.79、2.11、2.18、2.28×10^-3 mm²/s (P=0.006、0.003、0.012、0.001、0.003、0.008、0.002)，但二者△ADC相近。仅第3周末的ADC值为食管癌治疗近期疗效的独立影响因素(OR=0.134，P=0.007)，ROC曲线分析同样提示以第3周末的ADC值曲线下面积最大，诊断效能最优(Az=0.857)。阈值为2.02×10^-3 mm²/s，敏感性为80.0%，特异性92.9%。8例患者在治疗结束1年内发生肿瘤未控或复发，其ADC值曲线在第5周末出现“回落”，且DWMRI序列肿瘤长度从第3~5周缩小幅度明显减慢。结论 DWMRI可作为一种有效功能成像手段用以监测食管癌放化疗反应，放疗第3周末的肿瘤ADC值大小可能为优选疗效预测时间点，患者治疗后程ADC值降低或DWMRI序列肿瘤长度不再缩小提示复发高危倾向。

A clinical study of diffusion-weighted magnetic resonance imaging-based prediction of efficacy of chemoradiotherapy for esophageal carcinoma

Objective To investigate the parameters of diffusion-weighted magnetic resonance imaging (DWMRI) for prediction of the efficacy of chemoradiotherapy (CRT) for esophageal squamous cell carcinoma (ESCC), to determine the optimal time point and threshold for prediction, and to provide a basis for clinical practice. Methods From 2010 to 2011, 38 patients with ESCC were consecutively enrolled as subjects. All patients received three-dimensional conformal radiotherapy with 60 Gy in 30 fractions for 6 weeks. They also received concurrent or consolidation chemotherapy (FP or TP scheme) as adjuvant treatment. Patients received DWMRI scans before radiotherapy and at weeks 1-6 during radiotherapy. The apparent diffusion coefficient (ADC) values and tumor lengths obtained from serial DWMRI scans were recorded and analyzed. Comparison was made by paired t test. Repeated measurements were analyzed by analysis of variance (ANOVA) and multivariate ANOVA. The prognosis was predicted by the Logistic model. The effectiveness analysis and threshold screening were performed using the receiver operating characteristic (ROC) curve. Results In all patients, 20(52.6%) had complete response (CR) and 18(47.4%) had partial response (PR). There were no significant differences in the ADC values before radiotherapy and at weeks 1-6 during radiotherapy between patients with CR and PR (1.82 vs. 1.42;1.92 vs. 1.49;2.06 vs. 1.67;2.35 vs. 1.79;2.62 vs. 2.11;2.71 vs. 2.18;2.96 vs. 2.28×10^-3 mm²/s;P=0.006,0.003,0.012,0.001,0.003,0.008,0.002). The ADC value at third week during radiotherapy was the only independent prognostic factor for short-term treatment outcomes in patients with ESCC (OR=0.134, P=0.007). These results were also supported by the multivariate ANOVA analyses. The analysis of the ROC curve showed that at the third week during radiotherapy, the area under the ADC curve was the largest (Az=0.857) and the diagnostic effectiveness was the best;the threshold value, sensitivity, and specificity were 2.02×10^-3 mm²/s, 80.0%, and 92.9%, respectively. Eight patients who had tumor out of control or recurrence within 1 year after treatment had the ADC curve fall down at the end of the fifth week and the reduction of tumor length substantially slow down from the fifth week. Conclusions DWMRI is an effective imaging approach for monitoring tumor response to CRT in patients with ESCC. The ADC value at the end of the third week during radiotherapy may be the optimal time point for prediction of treatment outcomes. The reduction in the ADC value or non-reduction in tumor length at the end of the treatment indicates a high risk of recurrence.