多种MRI参数联合诊断肿块型肉芽肿性乳腺炎的Logistic分析

Logistic regression analysis in diagnosis of mass type granulomatous mastitis by combining multiple MRI parameters

目的:探讨多种MRI参数联合对肿块型肉芽肿性乳腺炎(MTGM)与浸润性导管癌(IDC)鉴别诊断的意义。方法:回顾性分析经手术病理证实的17例肿块型MTGM和51例IDC患者的临床及MRI资料,比较两组在年龄、发病部位、大小、形状、边界、临近皮肤增厚、同侧腋窝淋巴结肿大、增强曲线类型、内部环形强化、DWI信号、ADC值的差异,并采用单因素及多因素分析法进行统计学分析。结果:与IDC组比较,MTGM组病灶多发生于乳晕后区(10/17vs15/51,P=0.029),多引起周围皮肤水肿增厚(13/17vs24/51,P=0.035),增强扫描多见环形强化灶(11/17vs13/51,P=0.003),病灶内部液性区(T2WI高信号)的ADC值明显减低[(0.994±0.249)vs(1.762±0.545)×10^-3mm^2/s,P=0.001]。另外以上述征象为变量建立Logistic回归模型,筛选出肿块内部环形强化、液性区域的ADC值为预测MTGM的危险因素(P<0.05);2个因素联合预测MTGM的ROC曲线下面积为0.913。结论:肿块内部环形强化、液性区域的ADC值是预测MTGM的独立危险因素,联合应用可以提高MTGM的诊断符合率。

Objective: To explore the significance of combining multiple MRI parameters in differentiating mass type granulomatous mastitis (MTGM) from infiltrating ductal carcinoma (IDC). Methods: Clinical and MRI data of 17 patients with MTGM and 23 patients with IDC confirmed by operation and pathology were retrospectively analyzed.Differences of age,location,size,shape,border,adjacent skin thickening,ipsilateral axillary lymph node enlargement,enhancement curve type,internal ring-like enhancement,DWI signal and ADC value between the two groups were compared.Univariate and multivariate analysis was used for statistical analysis. Results: Compared with IDC group,the posterior areola area (10/17 vs 15/51, P =0.029),edema and thickening of the surrounding skin (13/17 vs 24/51, P =0.035),intra-tumoral ring-like enhancement (11/17 vs 13/51, P =0.003) and low ADC value of the high T 2WI signal intensity region [(0.994±0.249) vs (1.762±0.545)×10^-3 mm^2/s, P =0.001] were significant different in the MTGM group.In addition,further multivariate logistic regression model was established to identify two independant risk factors of MTGM ( P <0.05),including ring-like enhancement and ADC value of the high T2WI signal intensity region.The area under ROC curve of the combination of the two factors was 0.913. Conclusion: Intra-tumoral ring-like enhancement and low ADC value of the high T2WI signal intensity region are independent risk factors in predicting TMGM,and the combination of two parameters can improve the diagnostic accuracy of MTGM.