动态对比增强MRI预测局部进展期直肠癌新辅助放化疗疗效的价值

Value of dynamic contrast-enhanced MRI in predicting response to neoadjuvant chemoradiation in locally advanced rectal cancer

目的：探讨动态增强MRI（DCE-MRI）预测局部进展期直肠癌术前新辅助放化疗疗效的价值。方法前瞻性收集经结肠镜病理证实为直肠癌，行新辅助放化疗后拟行根治性全直肠系膜切除术的局部进展期直肠癌患者纳入研究。患者均在新辅助放化疗前2～5 d、术前1～4 d行直肠DCE-MRI检查，记录DCE-MRI参数，包括转运常数（Ktrans）、细胞外血管外空间的体积分数（Ve）和速率常数（Kep）。术后依据病理结果将患者分为病理完全缓解（pCR）组和非pCR组。采用t检验比较pCR组与非pCR组各DCE-MRI参数值及治疗前后参数变化值间的差异，采用ROC评价治疗前各参数值鉴别pCR和非pCR的效能，并确定Ktrans诊断阈值。结果38例患者纳入研究，pCR组12例，非pCR组26例。新辅助放化疗前，pCR组的Ktrans、Kep和Ve分别为（1.25±0.56）/min、（2.10±1.61）/min和0.73±0.34，非pCR组分别为（0.46±0.39）/min、（1.15±0.77）/min和0.32±0.12，差异均有统计学意义（t值分别为3.45、5.67和6.23，P均〈0.05）。新辅助放化疗后，pCR组的Ktrans、Kep和Ve分别为（0.28±0.13）/min、（0.62±0.27）/min和0.21±0.13，非pCR组分别为（0.32±0.12）/min、（0.83±0.42）/min和0.17±0.10，差异均无统计学意义（P均〉0.05），且2组间治疗前后上述参数的变化值差异也均无统计学意义（P均〉0.05）。治疗前Ktrans预测疗效为pCR的ROC曲线下面积为0.837，以Ktrans值为0.66/min鉴别pCR组和非pCR组，预测疗效为pCR的敏感度和特异度分别为75.0%（9/12）和96.2%（25/26）。Kep和Ve预测pCR的ROC曲线下面积分别为0.655和0.654。结论 DCE-MRI预测局部进展期直肠癌新辅助放化疗疗效具有一定价值，Ktrans值预测价值最大。

Objective To determine the value of dynamic contrast enhanced （DCE-MRI） in predicting treatment response before preoperative chemoradiotherapy in locally advanced rectal cancer. Methods A cohort of consecutive patients with histologically confirmed rectal adenocarcinoma treated with preoperative chemoradiotherapy followed by total mesorectal excision （TME） surgery was enrolled in a prospective, pilot trial. All enrolled patients were examined using DCE-MRI at two time points： 2 to 5 days before neoadjuvant chemoradiation, 1 to 4 days before surgery. The following perfusion parameters （Ktrans, Kep, Ve） were measured for tumor. The patients were classified into pathological complete response （pCR） and non-pCR group according to the pathological results after operation. Those perfusion parameters were compared between the pCR and the non-pCR group and between before and after CRT in pCR and the non-pCR group with the t test. Receiver-operating curves （ROC） were constructed to further investigate the predictive value of Ktrans, Kep, Ve before neoadjuvant chemoradiation and were used to determine a threshold value at which patents with pCR could be distinguished from patients without complete response. Results The final study＆amp;nbsp;population consisted of 38 patients. There were 12 patients with a pCR and 26 patients with non-pCR. Before neoadjuvant chemoradiation, the mean tumor Ktrans, Kep and Ve for pCR group were （1.25 ± 0.56）/min, （2.10 ± 1.61）/min and 0.73 ± 0.34, respectively, for non-pCR group they were （0.46 ± 0.39）/min, （1.15 ± 0.77）/min and 0.32±0.12, respectively. All perfusion parameters showed significant difference between those two groups（t values were 3.45,5.67 and 6.23 respectively, all P〈0.05）. After neoadjuvant chemoradiation, the mean tumor Ktrans, Kep and Ve for pCR group were （0.28 ± 0.13）/min, （0.62 ± 0.27）/min and 0.21 ± 0.13 respectively, for non-pCR group, they were （0.32±0.12）/min, （0.83±0.42）/min and 0.17±0.10, respectively. All perfusion parameters showed no difference between those two groups（P〉0.05）, as well as the changes before and after neoadjuvant chemoradiation in those groups（P〉0.05）. ROC analysis for Ktrans pre-treatment revealed that Ktrans had an AUC of 0.837 in predicting pCR. A Ktrans of 0.66/min was emerged as the optimal cut-off for distinguishing pCR from non-pCR and for Ktrans〉0.66/min, the sensitivity and specificity for predicting pCR were 75.0% （9/12） and 96.2% （25/26）. Kep and Ve showed an AUC of 0.655 and 0.654 in predicting pCR. Conclusions In locally advanced rectal cancer, DCE-MRI can aid in predicting treatment response before preoperative chemoradiotherapy. Ktrans may become a better predictor to classify which patients will benefit from neoadjuvant chemoradiation.