集成磁共振成像在鉴别乳腺良、恶性病灶中的临床应用价值

Clinical value of integrated magnetic resonance imaging in identifying benign and malignant breast lesions

目的 探讨集成磁共振成像(SyMRI)对乳腺良、恶性病灶的鉴别诊断价值。方法 回顾性收集108例经术后病理证实的乳腺病灶患者,术前均进行常规MRI和集成磁共振成像,其中良性病灶30例,恶性病灶78例,对每个病灶生成SyMRI生成的定量参数图,包括T1加权(T1WI)、T2加权(T2WI)和质子密度加权(PDWI)图像,增强前和增强后测量并分别记录为T1、T2、PD、R1、R2弛豫值及标准差(SD),计算其相对变化率(ΔT1%、ΔT2%、ΔPD%);比较良、恶性病灶间定量参数之间的差异;采用受试者工作特征曲线(ROC)评价各定量参数值及诊断模型对乳腺良、恶性病灶的鉴别诊断价值。结果 良性病灶定量参数均高于恶性病灶组(P<0.05);乳腺良、恶性病灶组T1e值、T1p值、T2e值、T2p值、ΔT1%、ΔT2%、ΔPD%比较差异均有统计学意义(P<0.05)。多因素logistic回归分析显示T1e值、T1p值、T2e值是预测乳腺癌的独立影响因素,T1e+T1p+T2e诊断模型鉴别诊断乳腺良、恶性效果最佳(AUC=0.902)。结论 SyMRI可进一步提升乳腺良、恶性病灶的鉴别效能,乳腺良性病灶的定量参数高于恶性病灶,T1e+T1p+T2e诊断模型效能最佳。

Objective To investigate the differential diagnostic value of integrated magnetic resonance imaging(SyMRI)for benign and malignant breast lesions.Methods 108 patients with breast lesions were retrospectively collected and confirmed by postoperative pathology,routine MRI and integrated magnetic resonance imaging were performed before surgery.Among them,30 patients had benign lesions,78 malignant lesions.Quantitative parametric plots generated by SyMRI were generated for each lesion,including T1-weighted(T1WI),T2-weighted(T2WI),and proton density-weighted(PDWI)images,pre-and post-enhancement were measured and recorded as T1,T2,PD,R1,R2 relaxation values and standard deviation(SD),and calculate its relative rate of change(ΔT1%,ΔT2%,ΔPD%).Differences between quantitative parameters between benign and malignant lesions were compared.The receiver operating characteristic curve(ROC)was used to evaluate the differential diagnostic value of each quantitative parameter value and the diagnostic model for benign and malignant breast lesions.Results The quantitative parameters of benign lesions were higher than those of malignant lesions group(P<0.05);the differences in T1e value,T1p,T2e value,T2p value,PDp value,ΔT1%andΔT2%between benign and malignant breast lesion groups were statistically significant(P<0.05).Multivariate logistic regression analysis showed that T1e value,T1p value,and T2e value are independent influencing factors in predicting breast cancer,and the T1e+T1p+T2e diagnostic model has the best effect in differential diagnosis of benign and malignant breasts(AUC=0.902).Conclusion Integrated magnetic resonance imaging(syMRI)can improve the discrimination of benign and malignant breast lesions,and the quantitative parameters of benign breast lesions are higher than those of malignant lesions,and the T1e+T1p+T2e diagnostic model has the best efficacy.