| Objective To investigate differential diagnostic value of enhanced T2 star weighted angiography (ESWAN) sequence between ovarian endometriomas and other pelvic cystic lesions by means of measuring the corrected phase value in corrected phase image and detecting signal voids within the wall of cysts.Materials and Methods From June 2010 to January 2011,41 patients who were first diagnosed as ovarian or pelvic cystic lesions by US were recruited in this prospective study. All of them were informed consent. Conventional MRI and ESWAN sequence were performed before treatment, and operation were carried out within two days to one week. All of them were pathologically confirmed. According to the appearance of conventional MRI,5 ovarian cystic mature teratomas were excluded. The remains were divided into the two groups:the 1st group was ovarian endometiomas (18 patients,30 cysts,32 loculi which were more than 2 cm in diameter), aged from 24～48 years old (mean 31.33±6.07); the 2nd group was other pelvic cystic lesions (18 patients,25 cysts,28 loculi which were more than 2 cm in diameter), aged from 19～75 years old (mean 51.17±14.17).T1SIR, T2SIR, CPV of the loculi which were more than 2 cm in diameter of the the two groups were measured, and then the data.was analysed as follows:①T1SIR, T2SIR, CPV of ovarian endometriomas and other pelvic cystic lesions were compared by independent samples t test.②Differential diagnostic efficacy, the optimal cut-off value, sensitivity, and specificity were obtained by means of analysing T1SIR, T2SIR, CPV of ovarian endometriomas and other pelvic cystic lesions with ROC curve.③Discriminant rate of using one of or combination of T1SIR, T2SIR, and CPV to differentiate ovarian endometriomas and other pelvic cystic lesions were acquired by way of discriminant analysis.Two experienced radiologists evaluated the MR images of the the two groups individually, which emphasized on the presence or absence of signal voids within the wall of cysts on the mIP magnitude image of ESWAN sequence. They were blinded to the histopathologic and clinical diagnoses of the lesions. Reader agreement was assessed in terms of the percentage of times the readers were concordant by using Kappa coefficients. The signal intensity of the lesions on T1-and T2-weighted image and the presence of shading on T2-weighted image were also assessed. Sensitivity, specificity, positive and negative predictive value of perceived signal voids within the wall of cysts for differentiating ovarian endometriomas from other pelvic cystic lesions were evaluated.The results were presented as mean±tandard deviation (x±s) that the test was significant level ofα=0.05, p<0.05 indicated significance for all tests.Results1. T1SIR of ovarian endometriomas was higher than that of other pelvic cystic lesions, while T2SIR and CPV of ovarian endometriomas were lower than that of other pelvic cystic lesions. Differences of T1SIR, T2SIR and CPV between the the two groups were statistically significant (p=0.000<0.05).2. By means of ROC curve analysis, T1SIR, T2SIR and CPV of the two groups were reliable indicators of differentiation. The area under ROC curve of CPV was the largest (0.965), so that the efficacy of CPV was the best. The cut-off values of T1SIR, T2SIR and CPV of the two groups were 0.90,3.46, and 2048.37, respectively. Sensitivity was the highest (89.3%) when using 2048.37 as cut-off value of CPV, and specificity was 100%.3. By way of discriminant analysis, the total discriminant rate of CPV (86.7%) was higher than that of T1SIR and T2SIR when using one of them only. The ovarian endometriomas discriminant rate of CPV (84.4%) was higher than that of T1SIR and T2SIR, and the other pelvic cystic lesions discriminant rate of CPV (89.3%) was equal to that of T1SIR, higher than that of T2SIR. CPV in combination with T1SIR (and/or T2SIR) showed the the highest total discriminant rate (91.7%).4. The readers were concordant in their evaluation of the presence or absence of signal voids within the wall of cysts in 52 (96.3%) of the 54 cystic lesions, with a Kappa coefficient of 0.925, which indicated nearly perfect agreement. Punctate signal voids within the wall of cysts were visualized on mIP magnitude image of ESWAN sequence in 27 of 30 endometriomas (90.0%), and in 2 of 24 other pelvic cystic lesions (1 of the 25 other pelvic cystic lesions was excluded because of that prominent susceptibility artifacts caused by colonic gas extended to the lesion). Sensitivity, specificity, positive and negative predictive value of perceived signal voids within the wall of cysts for differentiating ovarian endometriomas from other pelvic cystic lesions were 90.0%,91.7%,93.1%,88.0%, respectively.17 of the 30 endometriomas (56.7%) corresponded the conventional MRI criteria, that is, a cyst hyperintense on T1-weighted image and hypointense (shading) on T2-weighted image. The other 13 lesions did not correspond the criteria on conventional MR images. Signal voids within the wall of cysts on mIP magnitude image of ESWAN sequence were visualized in 12 of the 13 endometriomas that did not correspond the criteria on conventional MR images. With the addition of mIP magnitude image of ESWAN sequence to conventional MRI,29 of 30 endometriomas (96.7%) were diagnosed correctly.5.4 of the 6 cysts (8 loculi) which were discriminated incorretly when using CPV only were diagnosed correctly when detecting signal voids within the wall of cysts on mIP magnitude image of ESWAN sequence, the remain incorrect were 2 cysts of ovarian endometrioid carcinoma. In addition, the CPV of 1 cyst, low intensity on T1-weighted image and high intensity on T2-weighted image lacked signal voids within the wall of cysts on mIP magnitude image of ESWAN sequence, was 2016.12. Owing to being lower than 2048.37, correct diagnosis could be acquired. All in all, apart from 2 cysts of ovarian endometrioid carcinoma, the pelvic cystic lesions were diagnosed correctly by ESWAN sequence.Conclusion CPV was the most valuable to differentiate ovarian endometriomas and other pelvic cystic lesions when using T1SIR, T2SIR, and CPV only. CPV in combination with T1SIR (and/or T2SIR) was more helpful to diagnosis and differential diagnosis. Perceived signal voids within the wall of cysts were quite valuable for diagnosing ovarian endometriomas. Combination of corrected phase image and mIP magnitude image of ESWAN sequence was more valuable to differential diagnosis. Thus, ESWAN sequence could be carried out when the signal intensity of ovarian endometriomas are not typical. Measuring corrected phase value and detecting the presence or absence of signal voids within the wall of cysts on the mIP magnitude image are helpful to differential diagnosis between ovarian endometriomas and other pelvic cystic lesions. ESWAN could provide new information for diagnosis, is a new technique to evaluate ovarian endometriomas effectively and reliablely, and is complementary to traditional MRI. |
PDF Full Text Request