乳腺核磁共振简易序列在乳腺肿瘤筛查中的应用价值

A study of a breast MRI abbreviated protocol for breast neoplasm screening

目的评价乳腺核磁共振(MRI)简易序列(AP)对乳腺肿瘤的诊断准确性。方法回顾性分析86例乳腺MRI检查,采用AP和常规序列(FDP)两种检查方法对图像进行对比分析,AP包括轴位T2加权成像(T2-weighted imaging,T2WI)、扩散加权成像(diffuion-weighted imaging,DWI)和动态增强4期图像,FDP包括矢状位T2WI、轴位T1加权成像(T1-weighted imaging,T1WI)、T2WI、DWI和动态增强7期图像。根据乳腺影像报告和数据系统标准(breast imaging reporting and data system,BI-RADS)各自诊断并比较两者的一致性,再以病理结果为标准比较两种序列对乳腺肿瘤的诊断效能。结果2位医师独立阅片时,相同医师在不同序列间通过BI-RADS分类的诊断一致性为0.856、0.900,通过时间信号强度曲线(time-signal intensity curve,TIC)类型的诊断一致性为0.822、0.922,差异均无统计学意义(P>0.05);相同序列在不同医师间通过BI-RADS分类的诊断一致性为0.744、0.822,通过TIC类型的诊断一致性为0.889、0.878,差异均无统计学意义(P>0.05)。2位医师共同阅片,AP和FDP诊断恶性肿瘤在BI-RADS分类的敏感度分别为89.8%(95%CI:0.785~0.958)、91.5%(95%CI:0.806~0.968),特异度为71.0%(95%CI:0.518~0.851)、77.4%(95%CI:0.585~0.897),曲线下面积为0.804(95%CI:0.698~0.910)、0.845(95%CI:0.748~0.941);两种序列诊断恶性肿瘤在TIC类型的敏感度分别为86.4%(95%CI:0.745~0.936)、89.8%(95%CI:0.785~0.958),特异度为61.3%(95%CI:0.423~0.776)、67.7%(95%CI:0.485~0.827),曲线下面积为0.739(95%CI:0.623~0.855)、0.788(95%CI:0.679~0.897),以上差异均无统计学意义(P>0.05)。AP和FDP的扫描时间分别为(11.97±0.94)min、(21.25±1.12)min,差异有统计学意义(P<0.001)。AP的平均阅片时间较FDP减少13.5%。结论与FDP相比,采用AP可缩短乳腺MRI扫描时间并能保持诊断效能,可作为高危人群乳腺肿瘤MRI筛查的改进模式。

Objective This study aimed to evaluate the performance of breast magnetic resonance imaging(MRI)abbreviated protocol(AP)in diagnosing breast neoplasms.Methods We retrospectively analyzed the data of 86 patients who had undergone breast MRI examinations and compared the images using an AP and full diagnostic protocol(FDP).The AP consisted of axial T2-weighted imaging(T2WI),diffusion-weighted imaging(DWI),and four-phase dynamic enhancement sequences.The FDP consisted of sagittal T2WI,axial T2WI,T1-weighted imaging,DWI,and seven-phase dynamic enhancement sequences.All the images were analyzed using the Breast Imaging Reporting and Data System(BI-RADS).The consistencies between the different protocols were then calculated.With the pathological diagnosis as the gold standard,the diagnostic capabilities of the two protocols were compared.Result Two radiologists analyzed the AP and FDP images.The consistencies in the BI-RADS between the different protocols were 0.856 and 0.900,and those in time-signal intensity curves(TICs)were 0.822 and 0.922.Within the same protocol,the consistencies in the BI-RADS between different radiologists were 0.744 and 0.822,and those in TICs were 0.889 and 0.878.No significant differences were found(P>0.05).In terms of diagnosing malignant neoplasms using the BI-RADS,the sensitivities of the AP and FDP were 89.8%(95%CI:0.785-0.958)and 91.5%(95%CI:0.806-0.968),respectively;their specificities were 71.0%(95%CI:0.518-0.851)and 77.4%(95%CI:0.585-0.897),respectively;and the areas under the curves(AUCs)were 0.804(95%CI:0.698-0.910)and 0.845(95%CI:0.748-0.941),respectively.Diagnosing malignant neoplasms using TICs,the sensitivities of the AP and FDP were 86.4%(95%CI:0.745-0.936)and 89.8%(95%CI:0.785-0.958),respectively;their specificities were 61.3%(95%CI:0.423-0.776)and 67.7%(95%CI:0.485-0.827),respectively,and the AUCs were 0.739(95%CI:0.623-0.855)and 0.788(95%CI:0.679-0.897),respectively.There was no significant difference between the AP and FDP(P>0.05).The MRI acquisition times of the AP and FDP were 11.97±0.94 min and 21.25±1.12 min,respectively,with a significant difference(P<0.001).The average reading time was reduced by 13.5%using the AP compared with that using the FDP.Conclusion Compared with the FDP,the AP reduced the acquisition time and maintained the diagnostic accuracy,which can be used as an improved pattern for MRI screening in high-risk populations of breast neoplasms.