动态增强MRI半定量参数预测乳腺癌新辅助化疗疗效

目的 探讨动态对比增强MRI(DCE-MRI)半定量参数预测乳腺癌新辅助化疗(NAC)疗效的价值。方法 对48例经核芯针穿刺证实为乳腺浸润性癌的患者分别于NAC前和NAC 2周期后行DCE-MRI,采用两种ROI选取方法(ROI_whole法和ROI_hs法)测量最大线性斜率(S_max)、第二期强化程度(SI₂%)和峰值强化程度(SI_peak%)。按照Miller & Payne病理疗效分级系统,将患者分为组织学非显著反应(NMHR)组和组织学显著反应(MHR)组,比较两组 NAC 2周期后各参数的变化值及两种方法测得的参数预测乳腺癌NAC疗效的ROC曲线下面积。结果 48例中,NMHR组35例,MHR组13例。MHR组中应用ROI_whole法测得的S_max、SI₂%和SI_peak%的变化值分别为-54.6%(-86.8%~-19.8%)、-54.5%(-88.5%~-13.3%)和-23.3%(-63.5%~28.1%),NMHR组应用ROI_whole法得到的参数变化值分别为-14.7%(-68.8%~113.8%)、-18.7%(-74.3%~228.6%)和1.0%(-54.5%~29.6%),差异有统计学意义(P均<0.05);MHR组应用ROI_hs法测得的参数变化值分别为-57.3%(-82.2%~15.3%)、-41.1%(-84.3%~-16.8%)和-19.3%(-58.0%~5.8%),NMHR组应用ROI_hs法得到的参数变化值分别为-10.5%(-49.4%~130.8%)、-13.0%(-55.1%~216.7%)和-3.9%(-55.7%~51.2%), 差异亦有统计学意义(P均<0.01)。两种方法测得的S_max、SI₂%和SI_peak%的变化值预测NAC疗效的ROC曲线下面积的差异无统计学意义(P=0.85、0.61和0.84)。结论 DCE-MRI半定量参数可早期预测乳腺癌NAC的疗效;ROI选取方法对参数预测价值的影响不大。

Semi-quantitative dynamic contrast enhanced MRI for prediction of response to neoadjuvant chemotherapy of breast cancer

Objective To explore the value of semi-quantitative parameters in dynamic contrast enhanced MRI (DCE-MRI) for predicting final pathologic response in primary breast cancer patients who underwent neoadjuvant chemotherapy (NAC). Methods Totally 48 patients with pathologically proved infiltrating ductal carcinoma through core needle puncture biopsy underwent DCE-MRI before NAC and 2 cycles after the beginning of treatment. Semi-quantitative parameters (S_max, SI₂% and SI_peak%) were obtained by two ROI methods (ROI_whole and ROI_hs). The patients were categorized as non-major histological response (NMHR) group and major histological response (MHR) group according to histological response evaluated using Miller & Payne system. The changes of semi-quantitative parameter 2 cycles after NAC were compared between MHR group and NMHR group. Areas under ROC for parameters predicting NAC were compared between ROI_whole and ROI_hs. Results Among 48 patients, 35 were grouped into NMHR group and 13 into MHR group. Changes in kinetic parameters (S_max, SI₂% and SI_peak%) obtained by ROI_whole were -54.6% (-86.8% to -19.8%), -54.5% (-88.5% to -13.3%) and -23.3% (-63.5% to 28.1%) in MHR group, while were -14.7% (-68.8% to 113.8%), -18.7% (-74.3% to 228.6%) and 1.0% (-54.5% to 29.6%) in NMHR group, respectively (all P<0.05). Changes of kinetic parameters obtained by ROI_hs were -57.3% (-82.2% to 15.3%), -41.1% (-84.3% to -16.8%) and -19.3% (-58.0% to 5.8%) in MHR group, while were -10.5% (-49.4% to 130.8%), -13.0% (-55.1% to 216.7%) and -3.9% (-55.7% to 51.2%) in NMHR group, respectively (all P<0.01). No significant difference of areas under ROC curve for kinetic parameters predicting NAC between the two ROI methods (ROI_whole and ROI_hs) was found (P=0.85, 0.61 and 0.84). Conclusion Semi-quantitative dynamic parameters of DCE-MRI can predict final pathologic response of primary breast cancer 2 cycles after NAC. ROI methods have no obvious impact on DCE-MRI semi-quantitative parameters.