Dignostic Value Of Multimodal MRI In Ovarian Endometriosis

Purpose: To evaluate the feasibility of multimodal MRI in differentiating endometrial cysts from non-endometrial ones. The most excellent combination of the above sequences was made to improve the diagnosis rate of endometriosis.Methods: Sixty-eight patients with 80 histopathologically-proven ovarian cystic lesions underwent pelvic MRI including T1 WI, T2 WI, LAVA, DWI and ESWAN(axial3D,first echo:TR/TE:16.3/2.1ms,5 echo). Ovarian cystic lesions were divided into endometrial cysts(group A; n = 42), ovary pyosalpinx and hydrosalpinx(group B; n =14), and other ovarian cystic lesionss(group C; n = 24). The morphological characteristics were observed includes: whether it was sausage-like, with or without separation and the visible non-liquid components, thickness of the wall, and the wall of lesions with or without enhancement in the LAVA sequence. In the phase image whether there is ring that hyperintense outside and hypointense inside and hemorrhagic introtumoral susceptibility signal intensity(ITSS) in the wall. The above characteristics were compared by the Fisher exact probabilities or Chi-square test. The Spearman correction was used to evaluate the agreement between different measuremant of the T1R( T1 WI signal intensity ratio), T2R( T2 WI signal intensity ratio), ADC, phase and R2*(effective transverse relaxation rate) values, and the above values were measured and pairwise comparison of the R2* values among the three groups was made using Mann-Whitney U test. Then the comparision of the above quantitative values between endometriosis and non-endometriosis was made using Mann-Whitney U test, too.Receiver Operator Characteristic curve was used to calculate cut-off values and performance of T1 R, T2 R, ADC, phase and R2* values for distinguishing among groups.T1WI signal intensity and R2* value were also compared using AUC. The logistic regression was used to analyze the above morphological characteristics and quantitative values. The best combination of morphological characteristics and quantitative values was made by comparing the Youden index. Then morphological characteristics and quantitative values combined and with the sensitivity and specificity.Results: 1)The sausage-like lesions and thick wall lesions were more in group B than those of in group A and group B(P=0.000).There is no difference among the three groups of the lesions with the separation or with the visible non-liquid components(P>0.05).The enhanced lesions were more in group B and C than those of in group A(P=0.000). The lesions with the ring that hyperintense outside and hypointense inside or with hemorrhagic ITSS along the wall in phase image of group A were more than those in group B and C(P=0.000), and there was no difference between group B and C(P value were 0.319 and 0.240). 2)T1R and R2* values for group A were statistically higher than those for groups B and C( P =0.000). T2 R and ADC values for group A were statistically lower than those for groups B and C(the P values were 0.001 and0.000).There was no difference of the T1 R, T2 R, ADC and R2* values between group B and C(the P value were 0.643, 0.235, 0.210 and 0.150). The phase value of group A is lower than that of group C(P=0.008). The T2 R, ADC value, phase value of endometriosis group were higher than those of non-endometriosis group( the P value both are 0.000).The T1 R, R2*value, phase value of endometriosis were lower than those of non-endometriosis group( the P value were 0.000、0.000、0.004).Then the best combination of morphological characteristics were sausage-like lesions, lesion with the ring that hyperintense outside and hypointense and hemorrhagic ITSS inside wall in the phase image. The best quantitative values combination were T1 R, ADC and R2* value Combined the best combination of morphological characteristics and the ring that hyperintense outside and hypointense inside or with hemorrhagic ITSS along the wall in phase image with a sensitivity, specificity of 95.2% and 92.1%, respectively.Conclusion: The T1 R value and R2* value provides an effective way to discriminate endometrial cysts from other ovarian cystic lesions. Combine T1 R, ADC and R2*value,and non-sausage-like shape, observe the hemorrhagic ITSS and the hyperintense outside and hypointense inside ring in phase map provide the most excellent combination to discriminate endometrial cysts from other ovarian cystic lesions, and can improved the diagnosis rate of endometriosis.