治疗前18F-FDG PET/CT摄取异质性在HER2阳性转移性乳腺癌疗效预测中的价值

目的探讨治疗前¹⁸F-脱氧葡萄糖(FDG)PET/CT摄取异质性在人表皮生长因子受体2(HER2)阳性转移性乳腺癌靶向治疗效果预测中的作用。方法回顾性分析2012年5月至2018年4月复旦大学附属肿瘤医院的29例HER2阳性转移性乳腺癌患者均为女性,中位年龄52(32~69)岁]临床、病理及治疗前¹⁸F-FDG PET/CT影像资料,患者均接受曲妥珠单克隆抗体(简称单抗)一线治疗。对患者进行中位时间35(6~87)个月的随访。分析临床、病理相关特点及PET/CT代谢参数与患者无进展生存期(PFS)、总生存期(OS)之间的关系。通过Cox单因素分析筛选变量,将P≤0.01的变量纳入多因素Cox比例风险回归模型进行分析。采用时间依赖性受试者工作特征(ROC)曲线评价Cox回归模型的预测效能。生存曲线依据Kaplan-Meier法绘制,采用log-rank检验进行比较。结果29例患者中位OS为30(6~83)个月,中位PFS为10(2~29)个月。单因素分析结果示PET/CT瘤间摄取异质性指数,包括最大标准摄取值(SUVmax)最高病灶与最低病灶的SUVmax比值SUVmax-R;风险比(HR)=8.6(95%CI:2.7~27.8),P<0.001]、平均标准摄取值(SUVmean)-2.5标准摄取值(SUV)阈值为2.5]比值SUVmean-2.5-R;HR=2.6(95%CI:1.2~5.9),P=0.020]、肿瘤代谢体积(MTV)-2.5比值MTV-2.5-R;HR=2.4(95%CI:1.1~5.2),P=0.030]和总病灶糖酵解量(TLG)-2.5比值TLG-2.5-R;HR=3.2(95%CI:1.4~7.4),P=0.008]均与PFS有关,但与OS均无明显关系(均P>0.05)。多因素分析结果显示SUVmax-R为影响PFS的独立危险因素HR=6.8(95%CI:1.8~26.1),P<0.01]。SUVmax-R的ROC曲线下面积为0.747,以1.8作为界值,SUVmax-R可以有效地区分曲妥珠单抗受益与非受益人群(PFS:15.0与7.0个月;χ²=18.68,P<0.01)。结论PET/CT瘤间摄取异质性指数SUVmax-R是HER2阳性转移性乳腺癌患者靶向治疗的独立预后因素,能够早期预测靶向治疗效果。

Value of pretreatment18F-FDG PET/CT uptake heterogeneity for early prediction of response to targeted therapy in patients with HER2 positive metastatic breast cancer

Objective To evaluate the value of pretreatment¹⁸F-fluorodeoxyglucose(FDG)PET/CT-based heterogeneity for early prediction of targeted therapy outcome in patients with human epidermal growth factor receptor 2(HER2)positive metastatic breast cancer.Methods From May 2012 to April 2018,29 patients(all females,median age:52(32-69)years)who had HER2 positive metastatic breast cancer and underwent pretreatment¹⁸F-FDG PET/CT in Fudan University Shanghai Cancer Center were retrospectively enrolled.All patients received trastuzumab as first-line treatment and were followed up for 6-87(median time:35)months.The relations between clinicopathologic parameters or PET/CT-based parameters and progression-free survival(PFS)/overall survival(OS)were analyzed with Cox univariate analysis.The parameters with P≤0.01 were further analyzed with Cox multivariate analysis.Optimal cut-off values were determined by time-dependent receiver operating characteristic(ROC)curve analysis.The survival analyses were estimated by Kaplan-Meier method and log-rank test.Results The median OS of the 29 patients was 30(6-83)months,and the median PFS was 10(2-29)months.The PET/CT-based heterogeneity index(HI),including the maximum standardized uptake value(SUVmax)ratio(SUVmax-R;hazard ratio(HR)=8.6,95%CI:2.7-27.8,P<0.001),the mean standardized uptake value(SUVmean)-2.5(the cut-off value of standardized uptake value(SUV)=2.5)ratio(SUVmean-2.5-R;HR=2.6,95%CI:1.2-5.9,P=0.020),the metabolic tumor volume(MTV)-2.5 ratio(MTV-2.5-R;HR=2.4,95%CI:1.1-5.2,P=0.030),and the total lesion glycolysis(TLG)-2.5 ratio(TLG-2.5-R;HR=3.2,95%CI:1.4-7.4,P=0.008)of the lesion with the highest SUVmax to that with the lowest SUVmax,were significantly associated with PFS.None of the parameters was significantly associated with OS(all P>0.05).Multivariate analysis showed that the SUVmax-R was the only independent predictor for PFS(HR=6.8,95%CI:1.8-26.1,P<0.01).Area under the ROC curve for SUVmax-R was 0.747.With a cut-off value of 1.8,SUVmax-R could effectively distinguish the benefit from non-benefit population treated with trastuzumab(15.0 vs 7.0 months;χ²=18.68,P<0.01).Conclusion Pretreatment¹⁸F-FDG PET/CT-based HI has potential value for early prediction of first-line trastuzumab treatment outcome in patients with HER2 positive metastatic breast cancer.