临床研究DWI联合DKI、SWI在腮腺肿瘤诊断中的临床应用初探

Clinical application of DWI combined with DKI and SWI in the diagnosis of parotid tumors

目的 初步探讨扩散加权成像(diffusion weighted imaging, DWI)联合扩散峰度成像(diffusion kurtosis imaging, DKI)、磁敏感加权成像(susceptibility weighted imaging, SWI)在腮腺良恶性肿瘤的诊断及鉴别诊断中的临床应用价值。材料与方法 前瞻性纳入70例经术后病理证实为腮腺肿瘤的患者,其中良性肿瘤患者48例、恶性肿瘤患者22例,术前均行常规MRI检查、DWI、DKI、SWI序列扫描,通过工作站分析表观扩散系数(apparent diffusion coefficient, ADC)值、平均峰度(mean kurtosis, MK)值、径向峰度(radial kurtosis, Kr)值、轴向峰度(axial kurtosis, Ka)值、肿瘤周围静脉分布、磁敏感信号强度(susceptibility signal intensity, ITSS)等级、最大静脉直径Dv-max等功能成像指标。对差异具有统计学意义的结果绘制受试者工作特征(receiver operating characteristic, ROC)曲线,分析其诊断效能并获得诊断阈值。结果 腮腺良性肿瘤的ADC值高于恶性肿瘤,恶性肿瘤及Warthin瘤的ADC平均值[分别为(0.912±0.337)×10^(-3) mm^(2)/s、(0.867±0.181)×10^(-3) mm^(2)/s]明显低于多形性腺瘤ADC平均值[(1.837±0.318)×10^(-3) mm^(2)/s],各组间差异均有统计学意义(P<0.05),且具有较好的诊断效能。Warthin瘤的ADC平均值[(0.867±0.181)×10-3mm2/s]与恶性肿瘤ADC平均值[(0.912±0.337)×10^(-3) mm^(2)/s]接近,组间差异无统计学意义(P<0.05)。腮腺恶性肿瘤的MK、Kr、Ka平均值(分别为1.6772±0.5639、1.7327±0.5135、1.3325±0.4350)高于良性肿瘤MK、Kr、Ka平均值(分别为0.8083±0.2732、0.8821±0.3000、0.6776±0.2221),各组间差异均有统计学意义(P<0.05),且具有较好的敏感性和特异性。腮腺良性肿瘤静脉分布以周边分布为主,恶性肿瘤以中央分布为主,良性肿瘤ITSS等级以1级为主,恶性肿瘤ITSS等级以2~3级为主,恶性肿瘤的Dv-max高于良性肿瘤,各组间差异具有统计学意义(P<0.05)。DKI、DWI、SWI联合诊断模型中,以MK>1.0400、Kr>1.1500、Ka>0.8670、ADC值<1.140×10-3mm2/s、最大静脉直径Dv-max>1.500 mm诊断为恶性肿瘤,曲线下面积为0.995,敏感度为93.9%,特异度为100.0%。结论 DWI、DKI、SWI联合诊断模型可提高腮腺良恶性肿瘤的鉴别诊断效能。

Objective:To investigate the clinical application value of diffusion weighted imaging(DWI)combined with diffusion kurtosis imaging(DKI)and susceptibility weighted imaging(SWI)in the diagnosis and differential diagnosis of benign and malignant tumors of parotid gland.Materials and Methods:A prospective study was conducted on 70 patients with parotid tumors confirmed by postoperative pathology in our hospital,including 48 benign tumors and 22 malignant tumors.All patients underwent conventional MRI examination,DWI,DKI,SWI sequence scan before surgery.Apparent diffusion coefficient(ADC)value,mean kurtosis(MK)value,radial kurtosis(Kr)value,axial kurtosis(Ka)value,venous distribution around tumor,susceptibility signal intensity(ITSS)grade,maximum venous diameter Dv-max and other technical indicators of functional imaging were analyzed by workstation.Receiver operating characteristic(ROC)curves were drawn for statistically significant results to analyze their diagnostic efficiency and obtain diagnostic thresholds.Results:The ADC value of benign parotid tumor was higher than that of malignant tumor,and that of malignant tumor and Warthin tumor was lower than that of pleomorphic adenoma.The differences among all groups were statistically significant,with good diagnostic efficacy.The ADC value of Warthin tumor was close to that of malignant tumor,and there was no statistical significance between groups(P<0.05).The mean values of MK,Kr and Ka in malignant parotid tumors were higher than those in benign tumors(P<0.05),with statistical significance among all groups and good sensitivity and specificity.The venous distribution of benign parotid tumors was mainly peripheral,and that of malignant tumors was mainly central.The ITSS grade of benign tumors was mainly grade 1,and that of malignant tumors was mainly grade 2-3.The Dv-max of malignant tumors was higher than that of benign tumors,and the difference between each group was statistically significant(P<0.05).In the joint diagnostic model of DKI,DWI and SWI,MK>1.0400,Kr>1.1500,Ka>0.8670 and ADC<1.140×10^(-3) mm^(2)/s,the maximum vein diameter(Dv-max)>1.500 mm,the area under the curve is 0.995,the sensitivity is 93.9%,and the specificity is 100.0%.Conclusions:DWI,DKI,SWI combined diagnosis model can improve the differential diagnosis efficiency of benign and malignant parotid tumors.