The Value Of3.0T MR Dynamic Contrast-enhanced And Diffusion-weighted Imaging In The Diagnosis And Staging Of Endometrial Carcinoma

Part I:The value of3.0T MR Diffusion-weighted Imaging in the diagnosis of endometrial carcinomaObjective:To evaluate the diagnostic value of3.0T MR diffusion weighted imaging (DWI) and apparent diffusion coefficient (ADC) measurements in endometrial carcinoma.Methods:The study includes histopathologically proved50cases of endometrial carcinomas,62cases of various benign lesions (38polyps and12hyperplasia and12submucosal leiomyomas), and20cases with normal endometrium. All patients underwent conventional MRI serial sequences and diffusion-weighted imaging (DWI) with b values of0,500,1000s/mm2on3.0T MRI. Apparent diffusion coefficients (ADC value) were measured in cancers, benign lesions and normal endometriums. Statistical difference was compared among three groups using one-way ANOVA. ROC analysis was used to investigate the proper ADC value threshold to differentiate cancers and benign lesions. The ADC values of endometrioid adenocarcinomas among groups at different pathological grades were also compared using one-way ANOVA. The correlation of ADC of the tumors and pathological grades was calculated with Spearman correlation analysis.Result:Within all of the50endometrial carcinomas,41cases were detected on T2-weighted imaging, with a detection rate of82%, while45cases were detected on DW MRI combined with T2-weighted imaging, with the detection rate of90%. The ADC value of endometrial cancer [(0.81±0.21) X10-3mm2/s] was significantly lower than those of normal endometrium [(1.65+0.46)×10-3mm/s]and benign lesion [(1.54+0.29)×10-3mm2/s](P<0.05), despite an inter-patient overlap between cancers and benign lesions. ROC analysis revealed a discriminatory performance in differentiating cancer and benign lesions of AUC=0.986for ADC, with a best threshold of low-ADC for the diagnosis of endometrial cancer was1.18×10-3mm2/s, the diagnostic sensitivity and specificity were of94.0%and91.9%, respectively.45endometrioid adenocarcinomas of well-differentiated (G1) in26cases, moderately-differentiated (G2) in13cases, poorly-differentiated (G3) in6cases. The mean ADC values of G1、G2、G3endometrial carcinoma were (0.87±0.22)×10-3mm/s、(0.78±0.16)×10-3mm2/s (0.59+0.21×10-3mm2/s, respectively. The ADC value of G3tumors was significantly lower than that of G1tumors (P<0.05). And there was a weak correlation between ADC and pathological grades (r=-0.229, P=0.009).Conclusion:DWI combined with conventional MR series can provide more information for the differential diagnosis of endometrial carcinoma and DWI is a helpful complement for conventional MR examination. ADC measurement has a potential ability to differentiate among cancerous and benign and normal tissue of the endometrium. The ADC values of endometrial cancers of higher grade show tendency to decrease compared to those of lower grade, although estimation of histologic grade based on ADC values seems difficult because of considerable overlap. Part II:The value of3.0T MR dynamic contrast-enhanced imaging in the diagnosis of endometrial carcinomaObjective:To investigate the imaging characteristics of dynamic contrast-enhanced MR imaging(DCE-MRI)in endometrial carcinoma, and to evaluate the efficacy of DCE-MRI in the diagnosis of endometrial carcinomaMethods:The study includes histopathologically proved32endometrial carcinomas and30benign lesions (16polyps,10hyperplasia and4submucosal leiomyomas). All patients underwent conventional MRI series and dynamic contrast-enhanced MR imaging on3.0T MR. Morphological feature between the benign and malignant lesions was evaluated and statistical difference was compared. Morphological feature included contour, signal intensity (SI), SI homogeneity, intratumoral cyst, fibrous core, junctional zone. Time-signal intensity curve between cancers and benign lesions were obtained and statistical difference was compared. Several semi-quantitative parameters on DCE-MRI, including maximum signal intensity (SImax), time to peak (TTP), the maximum enhancement ratio (SIpeak%) and washout were calculated and compared with independent sampled t test. ROC analysis was used to investigate the best DCE parameters and proper threshold to differentiate cancers and benign lesions. The DCE parameters of EC among groups at different pathological grades were also compared using one-way ANOVA. And the correlation of DCE parameters in the tumors with pathological grades was calculated by Spearman correlation analysis.Result:There was no statistical difference between endometrial carcinoma and benign lesions in morphological feature except junctional zone. TIC had statistical significance in differentiation between benign and malignant lesions. The TIC of endometrial cancer was mainly type III while Type I most appeared in benign lesions. All DCE parameters between endometrial carcinoma and benign lesions had significant differences,(P <0.05), but there were overlapped in some degree. ROC analysis revealed a discriminatory performance in differentiating cancer and benign lesions of AUC=0.827for TTP (with a threshold of154, the diagnostic sensitivity was81.3%, and specificity was76.7%) and AUC=0.808for washout (with a threshold of0.0853, the diagnostic sensitivity was75.0%, and specificity was80%). The DCE parameters of EC among groups at different pathological grades had no significal difference (P>0.05). And there was no significant correlation between DCE parameters and pathological grades.Conclusion:Dynamic contrast-enhanced MRI is useful in the diagnosis of endometrial cancer. The reference parameters and time-signal intensity curve is valuable in differentiating the endometrial cancer from benign lesions. It is difficult for all DCE parameters to differentiate the grade of endometrial cancer. Part III:Staging of endometrial carcinoma:role of diffusion-weighted and dynamic contrast-enhanced MR ImagingObjective:To evaluate the diagnostic performance of dynamic contrast material-enhanced (DCE) MR imaging and diffusion weighted (DW) magnetic resonance (MR) imaging in predicting the depth of myometrial invasion, cervical invasion and overall stage in patients with endometrial cancer.Methods:Fifty patients with histopathologically proven endometrial carcinoma were underwent preoperative3-T MRI imaging before total hysterectomy. The MR protocol included T2-weighted imaging, DCE MR imaging and DW images acquired with b-values of0,500and1000sec/mm2. MRI findings were compared with postoperative histopathological findings; the latter was taken as the gold standard. Software (SPSS17.0for Windows) was used for analysis. Sensitivity, specificity, positive and negative predictive values, diagnostic accuracy of each sequence in depicting myometrial invasion and cervical invasion were calculated. The accuracy of the evaluation of the myometrium invasion, cervical invasion and stage was compared among each sequences using Fisher’s exact test. The presence of pitfalls associated with inaccurate assessment of myometrial invasion at each sequence was examined with a Fisher exact test.Results:The diagnostic accuracy for the identification of myometrial invasion were76.0%on T2-weighted imaging alone,84.0%on DCE MR combined with T2-weighted imaging and92.0%on DW MR combined with T2-weighted imaging, there was statistical difference of accuracy between T2-weighted imaging and DW MR combined with T2-weighted imaging (P=0.002<0.05) There was no association between inaccurate assessment of myometrial invasion and standard pitfalls with DW MR imaging (P>0.05), Tumor extension to uterine cornu was the only variable significantly associated with incorrect estimation of depth of myometrial invasion on DCE-MR (P=0.007<0.05). In assessing cervical involvement, the accuracy of T2-weighted imaging alone, DCE MR combined with T2-weighted imaging, and DW MR combined with T2-weighted imaging was88%,94%,96%, respectively; no statistically significant difference was observed (P>0.05). For overall stage, the accuracy of T2-weighted imaging alone, DCE MR combined with T2-weighted imaging, and DW MR combined with T2-weighted imaging was66%,78%,92%, respectively; DW MR combined with T2-weighted imaging has significantly higher staging accuracy than that of DCE MR combined with T2-weighted imaging (P=0.0391<0.05) and T2-weighted imaging (P=0.0002<0.05), and DCE MR combined with T2-weighted imaging has statistical higher staging accuracy than that of T2-weighted imaging alone (P=0.0313<0.05).Conclusion:DW MR imaging and DCE-MR imaging showed high diagnostic accuracy in assessment of the depth of myometrial invasion, cervical invasion and overall stage in patients with endometrial cancer, but DWI has significantly higher staging accuracy compared with DCE MR imaging, and we propose adding DW imaging and DCE-MR imaging as adjunct sequence in the routine MR protocol for the assessment of uterine lesions.