Histographic Analysis of Magnetic Resonance Imaging for the Evaluation of Ovarian Endometrial Invasion

Endometriosis occurs in 6%–10% of reproductive-age females [1,2]. Approximately 30%–50% of females with endometriosis are infertile and 25%–50% of infertile females Received: November 11, 2019 Revised: January 17, 2020 Accepted: January 21, 2020 Correspondence Hyun Jung Lee, MD, PhD Department of Obstetrics and Gynecology, School of Medicine Kyungpook National University, 130 Dongdeok-ro, Jung-gu, Daegu 41944, Korea. E-mail: [email protected] This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. Original Article eISSN 2384-1109 iMRI 2024;28(1);1-7 https://doi.org/10.13104/imri.2019.1030 Histographic Analysis of Magnetic Resonance Imaging for the Evaluation of Ovarian Endometrial Invasion Hyun Jung Lee Department of Obstetrics and Gynecology, School of Medicine, Kyungpook National University, Daegu, Korea Purpose: The purpose of this study was to evaluate the effectiveness of histographic analysis for the perfusion map of pelvic magnetic resonance imaging (MRI) in predicting remnant ovarian tissue in patients with ovarian endometriosis.

To generate the perfusion map, subtracted T1-weighted image (T1-WI) was divided by contrast enhanced T1-WI with using image analysis software Im- ageJ. Each region of interest (ROI) was quantified by outlining of the affected ovaries with endometrioma at the level with the largest area of normal ovary tissue and normal contralateral ovaries using the measurement tool on the software. Consequently, the number of ratios per each pixel comprising the perfusion map was scored from 0 (not perfused) to 1 (totally perfused). The pixel information, including area within ROI, mean with standard deviation of signal intensity, as well as integrated density of affected ovary with endometrioma, were compared with that of the normal ovary. Additionally, we com- pared the histogram according to the severity of the ovarian invasion.

In comparison between the affected ovary with endometrioma and the normal ovary, the perfusion ratio of the normal ovary was higher than that of the affected ovary (0.48 ± 0.07 vs. 0.20 ± 0.12, p < 0.001), whereas the area within the ROI and the perfusion ratio was higher in the affected ovary. According to the severity of the endometrial inva- sion of the ovary based on the surgical findings, the area with the perfusion ratio between 0.4 and 0.8 (199.17 ± 163.15 vs. 528.00 ± 154.43, p = 0.003), perfusion ratio (0.11 ± 0.07 vs. 0.27 ± 0.11, p = 0.012), and, and integrated density (187.33 ± 106.32 vs. 427.125 ± 132.24, p = 0.003) was lower in the group of severe invasion than those of the mild group and moderate invasion group.

The histographic analysis for the perfusion map of the pelvic MRI could be valuable in revealing the extent of the endometrial invasion and viable remnant ovarian tissue.

Endometriosis; Ovary; Histogram; Magnetic resonance imaging 2 www.i-mri.org MR Histogram for Ovarian Endometriosis | Hyun Jung Lee have endometriosis [3]. Recently, patients with the ovarian endometrioma want to preserve the ovarian function by min- imizing damage to the remnant healthy ovarian tissue during ovarian cystectomy for endometriosis. Thus, the surgeon must consider the remnant ovarian tissue from invasion of endo - metriosis to avoid its destruction with coagulation or disrup- tion of the ovarian blood supply during the appropriate surgi- cal approach [4]. However, it is difficult to determine the degree of invasion of endometriosis from stretching of the sur- rounding tissues, because of the presence of cysts with inflam- matory changes of endometriosis [5]. Recently, clinical data have been used to develop a clinical tool to predict an infertile probability of pregnancy after sur- gery. The revised American Fertility Society classification (rAFS classification), the most widely used staging system of endo- metriosis, is used to predict the recurrence potential of endo- metriosis after surgery [6,7]. As the rAFS classification has limited predictive ability for pregnancy after surgery, the en- dometriosis fertility index (EFI) was proposed to predict fecun- dity after endometriosis surgery [8]. In addition to providing a detailed score including fallopian tubes, fimbriae of fallopian tubes, and ovaries by calculating the least function scores, the EFI also combines conception related factors such as age, du- ration of infertility, and gravidity history. However, it could only be performed during surgery. Also, the reliability of the

is presented in question as a subjective evaluation

according to the operator. Thus, the need for preoper- ative imaging to assess ovarian tissue around the endometrio- sis is raised to minimize damage to the ovarian reserve. Typical magnetic resonance imaging (MRI) features of ovari- an endometriosis include a high signal intensity on T1-weight- ed images (T1-WIs) and T2-WIs [9,10]. However, chronic bleeding has led to high concentration of iron and protein in the endometrial cysts, resulting in gradual changes in signal intensi- ty, making it difficult to identify ovarian tissue in hemorrhagic endometrial lesions [9,11,12]. However, it is mandatory for the surgeon, who wants to perform ovary preserving surgery, to con- sider the remnant ovarian tissue and endometriosis involving that area, to avoid its destruction with blood coagulation or disruption of the ovarian vessel while using the appropriate sur- gical approach [4]. In this situation, the subtraction imaging may be valuable in determination of displacing the ovarian tissue from the endometriosis. After subtraction, native T1 signals disappear, and the remaining signals are from the en- hancement associated with the intravenous contrast adminis- tration mainly [13-15]. Also, analysis for the contrast enhance- ment effect after the normalization of each image could characterize the different perfusion pattern between the endo- metrial tissue and remnant ovarian tissue [16]. In this study, a perfusion map was obtained by dividing the subtracted T1-WI with contrast enhanced T1-WI to evaluate the effect of contrast enhancement. In contrast to the poorly perfused endometrial cyst, regardless T1 high signal intensity, remnant ovarian tissue could show well perfused. Additionally, the histographic analysis could estimate the extent of viable remnant tissue quantifiably. Consequently, the purpose of this study was to evaluate the effectiveness of the histographic analysis for the perfusion map of the pelvic MRI in predicting remnant ovarian tissue in patients with ovarian endometriosis.

Retrospective data collection and analysis were approved by the Institutional Review Board of Kyungpook National University Hospital (KNUH 2017-06-012). The need for in - formed consent was waived due to the retrospective design of this study. Eight patients had stage III endometriosis, and 6 patients had stage IV endometriosis. All of patients under- went laparoscopic ovarian cystectomy. Endometriosis was classified based on the revised American Society for Repro - ductive Medicine classification [7]. The final diagnosis was obtained based on the pathological examination of the surgi- cally excised specimen of the lesion. Before MRI examination, the patients underwent 6 hours of fasting, followed by an intramuscular administration of a peri- staltic inhibitor. MRI examinations were performed on a 3.0-T machine (Skyra; Siemens Health Care, Erlangen, Germany). We used a set of 8-channel phased array coils dedicated to different body parts included in this study. MRI scans were interpreted on a picture archiving and communications system workstation (PiViewStar; INFINITT, Seoul, Korea) to identify endometriosis. An author (HJL) qualitatively analyzed all MRI scans for ovarian en- dometriosis blinded to any information on clinical details. To generate the subtracted T1-WI, T1-WI and contrast en- hanced T1-WI were performed with similar parameters, such as field of view, slice thickness, repetition time, echo time, and fat suppression. Subtraction images were obtained after intrave- nous contrast injection to assess the presence of enhancement for adequate depiction of these lesions. To evaluate the effect of contrast enhancement, a perfusion map was obtained by di- viding subtracted T1-WI with contrast enhanced T1-WI using the digital image analysis software ImageJ (version 1.47q; Na- tional Institutes of Health, Bethesda, MD, USA). Consequently, the number of ratios per each pixel composing perfusion map was scored from 0 (not perfused) to 1 (totally perfused) (Fig. 1). Each region of interest (ROI) was quantified by outlining of affected ovaries with endometrioma at the level with the larg- est area of normal ovary tissue and normal contralateral ova- ries using the measurement tool on the software. Finally, the 3 www.i-mri.org https://doi.org/10.13104/imri.2019.1030 histogram for the perfusion map was produced for the evalu- ation of the enhancement effect of endometriosis and normal ovarian tissue (Bins = 100). The enhancement degree was classified into poor (perfusion ratio lesser than 0.2), mild (per- fusion ratio between 0.2 and 0.4), moderate (perfusion ratio between 0.4 and 0.8), and high enhancement (perfusion ratio greater than 0.8). The pixel information, including area, mean with standard deviation of signal intensity, as well as integrated density of affected ovary with endometrioma, was compared with that of the normal ovary. The integrated density, the sum of all the pixel intensities in the ROI, expect total amount of contrast enhancement effect in this study. Additionally, we compared histogram according to severity of ovarian invasion based on EFI [8]. Statistical analyses were performed using SPSS version 13.0 for Windows (SPSS Inc., Chicago, IL, USA). The Student’s t-test was used to examine the difference in numerical variables. Sta- tistical significance was set at a p-value < 0.05.

Demographic, clinical, and laboratory characteristics of 14 patients with surgically confirmed ovarian endometriosis are summarized in Table 1. The mean age was 28.21 ± 5.71 years. There were no statistically significant differences between the rAFS classification III and IV with regards to age, birth history, body mass index (BMI), C reactive protein, anti-Müllerian hor- mone, or least function score of ovary in EFT. EFI score was higher in rAFS classification III than IV (p = 0.008). The normal ovary was defined in all patients. On T2-WI, the normal ovary showed cysts of varying sizes surrounded by the darker solid ovarian stromal tissue at the level of the maxi - mum ovarian diameter and bright cysts surrounded by the Fig. 1. Analysis of the histogram of the normal ovary. A: The normal ovary shows bright cysts surrounded by the darker solid ovarian stroma on T2-WI. B: Contrast enhanced T1-WI shows intense enhancement of stroma, which is a more prominent contrast to the subtracted follicular cyst. C: The contrast enhanced TI-WI is subtracted by the unenhanced T1-WI to yield the subtracted T1-WI. D: The perfusion map is obtained by dividing the subtracted T1-WI (C) with the contrast enhanced T1-WI (B). E: The histogram for the perfusion map for the normal ovary shows the perfusion ratio of 0.40 to 0.80 of 70.3% ± 20% of included pixels. T1-WI, T1-weighted image; T2-WI, T2 weighted image. 35 30 25 20 15 10 5 0 No of pixels 0 0.2 0.4 0.6 0.8 1.0 Perfusion ratio A C B D E Table 1. Demographic, clinical, and laboratory characteristics of study subjects rAFS classification pIII (n = 8) IV (n = 6) Age (yr) 26.88 ± 4.49 30.00 ± 7.07 0.331 G + P + A ≥ 1, n (%) 1 (12.5) 1 (16.7) 0.692 BMI 20.00 ± 2.16 18.85 ± 2.32 0.357 CRP 0.47 ± 0.81 0.02 ± 0.02 0.159 AMH 11.12 ± 7.12 5.65 ± 2.39 0.074 Diameter of endometrioma 53.75 ± 33.83 48.33 ± 24.91 0.302 EFI 8.00 ± 0.76 6.33 ± 1.21 0.008 Least function score for ovary 1.75 ± 1.04 1.50 ± 0.55 0.603 Values are presented as mean ± standard deviation unless otherwise indicat- ed. G + P + A, gravidity, parity and abortion. rAFS, revised American Fertility Society; III, stage III endometriosis; IV, stage IV endometriosis; BMI, body mass index; CRP, C-reactive protein; AMH, anti- müllerian hormone; EFI, endometriosis fertility index. 4 www.i-mri.org MR Histogram for Ovarian Endometriosis | Hyun Jung Lee darker solid ovarian stroma. The histogram for the normal ovary was measured in all patients. At the level of largest area, the normal ovary showed 350 mm2 of area and 0.48 ± 0.07 of perfusion rate. Perfusion rates in 70.3% ± 20% of pixels from normal ovaries ranged from 0.40 to 0.80 (Fig. 1). According to the EFI classification, 1 case was classified into mild, 7 cases were into moderate, and 6 cases were classified into severe, which was correlated with the MRI. Although the MRI of ovarian endometriosis showed variable signal intensi- ties on T1-WI and T2-WI, the cystic lesion of endometriosis showed lesser than 0.1 perfusion ratio (Fig. 2). In comparison between the affected ovary by endometriosis and the contra- lateral normal ovary, perfusion ratio of normal ovary was higher than that of endometriosis (0.48 ± 0.07 vs. 0.20 ± 0.12, p < 0.001), whereas area within the ROI and perfusion ratio was higher in the affected ovary (Table 2). According to the severity of the endometrial invasion of the ovary based on the surgical findings, the area with perfusion ratio between 0.4 and 0.8 (199.17 ± 163.15 vs. 528.00 ± 154.43, p = 0.003), per- fusion ratio (0.11 ± 0.07 vs. 0.27 ± 0.11, p = 0.012), and integrat- ed density (187.33 ± 106.32 vs. 427.125 ± 132.24, p = 0.003) was lower in the endometriosis in group of severe invasion than those of mild and moderate invasion group (Fig. 3 and Table 3).

Although ultrasonography (US) is the main modality of choice for identifying and characterizing adnexal cystic le - sions, MRI is performed in selected patients according to the

of US and the severity of symptoms [9,17,18]. MRI is generally performed to exclude malignancies in cases of in - termediate US features of ovarian masses [10,19,20]. Typical MRI findings of endometriosis include a high signal intensity on T1-WI and T2-WI [9,10]. As data are limited on gadolinium enhancement in the evaluation of endometriosis, contrast en- hancement could be recommended in the evaluation of inde- terminate adnexal endometriosis, such as for distinction from A B C Table 2. Comparison of histogram analysis between affected ovary by endometriosis and contralateral normal ovary Affected ovary by endometriosis Contralateral normal ovary p* Area (mm2) 2100.86 ± 1315.10 350.07 ± 138.08 < 0.001 0 < perfusion ratio < 0.2 1402.57 ± 1288.86 30.93 ± 35.56 0.001 0.2 < perfusion ratio < 0.4 141.14 ± 77.82 50.86 ± 30.58 0.001 0.4 < perfusion ratio < 0.8 387.07 ± 227.15 246.79 ± 124.10 0.026 Perfusion ratio 0.20 ± 0.12 0.48 ± 0.07 < 0.001 Integrated density 324.36 ± 170.08 167.21 ± 67.89 0.004 *Paired t-test. Fig. 2. The ROI (dashed line) for each stage of endometrial invasion of the ovary. A: Poorly perfused cystic endometriosis (e) is detected within the ovary, outlined by the ROI (dashed line), preserving the ovarian stroma regarded as minimal invasion. B: The enlarged endometrial cyst (e) displaced normal stroma (arrow) with structural distortion. C: Severe endometrial invasion stretches the normal ovarian stroma (arrow) with volume loss. ROI, region of interest. 5 www.i-mri.org https://doi.org/10.13104/imri.2019.1030 other hemorrhagic adnexal lesions, luteal ovarian cysts, or tubo- ovarian abscesses. Additionally, gadolinium enhancement is cru- cial for depicting strongly enhanced mural nodules if atypical features suggest potential malignancy on US or T2-WI [21,22]. However, this study highlighted the role of contrast enhance- ment that revealed endometrial invasion of the adjacent healthy ovarian tissue. The ovaries are readily identified on MRI because they con- tain multiple various stages of ovarian follicles in the majority of females in their reproductive period [23,24]. T2-WI are the most useful sequences in the diagnosis of ovaries and ovarian follicles in females of reproductive age. In postmenopausal females, ovaries show more homogeneous low signal intensity in T2-WI, and are more difficult to identify because of atro - phic changes [25]. Ovarian stroma shows contrast enhancement similar to myometrium and contrast enhancement pattern also correlated with age and menopausal status on contrast en - hanced T1-WI [25]. Cystic follicles and functional ovarian cyst were found frequently and had variable appearance. Most cysts show discrete enhancement of the wall. In this study, the histo- gram for normal ovary shows narrow peak between 0 and 0.03 perfusion ratio, representative ovarian follicular cyst. Overall perfusion ratio of normal ovary was 0.48 ± 0.07, which was significantly different for endometrial invasion of the ovary. In this study, the perfusion map showed the absence of en- hancement within the cystic lesions, regardless of the vari - able signal intensity of T1 or T2-WI. The perfusion map could be valuable in differentiating complex cystic lesions with T1 high signal intensity such as ectopic pregnancy, borderline en- dometriosis or small malignant lesion within the cystic lesion [26]. The perfusion map can adequately determine the degree of contrast enhancement in the background of tissues that 3000 2500 1000 500 0 60 50 40 30 20 10 0 No of pixels 0.2 0.4 0.6 0.8 Severe Moderate Normal ovary 0.4 0.6 0.8 Perfusion ratio Fig. 3. The histogram of the perfusion map for ovarian endometriosis with the normal ovary. The histogram showed increased peak (arrow) due to the hemorrhagic cyst of severe endometriosis around zero perfusion. The area under the curve of the perfusion ratio between 0.4 and 0.8 highest in the group with moderate endometriosis followed by the normal ovary and severe endometriosis groups. Table 3. Comparison of histogram analysis according to severity of endometrial invasion to ovary Mild and Moderate (LF score = 2 or 3) (n = 8) Severe (LF score = 1 or 0) (n =6) p* Area (mm2) 2102.88 ± 1499.05 2098.17 ± 1162.15 0.995 0 < perfusion ratio < 0.2† 1389.13 ± 1453.11 1420.50 ± 1167.15 0.996 0.2 < perfusion ratio < 0.4 173.00 ± 74.74 98.67 ± 64.31 0.070 0.4 < perfusion ratio < 0.8 528.00 ± 154.43 199.17 ± 163.15 0.003 Perfusion ratio 0.27 ± 0.11 0.11 ± 0.07 0.012 Integrated density 427.125 ± 132.24 187.33 ± 106.32 0.003 LF score (least function score): 3 = mild dysfunction, 2 = moderate dysfunction, 1 = severe dysfunction, 0 = absent or nonfunctional. *Student t-test. 6 www.i-mri.org MR Histogram for Ovarian Endometriosis | Hyun Jung Lee showed a high signal intensity on contrast enhanced T1-WI such as adjacent uterus or inflammatory change related with endometriosis. Our results showed increased number of pixels between 0.4 to 0.8 perfusion ratios in moderate advanced en- dometriosis, suggesting increased volume of the inflammatory change of endometriosis as well as cystic lesions. However, advanced status endometriosis showed decreased or similar volume of endometrial tissue between 0.2 to 0.8 perfusion ratio suggestive ovarian invasion of endometriosis. The perfusion map can aid in evaluation of the remnant ovarian tissue. Endometrial invasion of the ovarian tissue is as- sociated with the ovarian function. Also, bilateral invasion of the ovarian tissue is associated with infertility. In this study, severe invasion of endometriosis showed the lower area with moderate perfusion ratio similar to that of the ovary, suggest- ing the perfusion map could predict remnant ovarian preserve. Such results support the assessment of the extent of ovarian involvement could be mandatory to determine the preopera- tive grade [6,27]. The rAFS classification, the most widely used staging sys - tem of endometriosis, depends on the results of laparoscopic examination and laparotomy. The staging of endometriosis requires the detailed observation and recording of the site, number, size, and depth of the endometriosis lesions, as well as the degree of adhesions, to define the final score. However, several studies observed no association between the endome- triosis stage or lesion type and lesion site and the cumulative probability of pregnancy [28]. Fujishita et al. [29] modified the AFS classification of endometriosis by adding the TOP score (fallopian tubes, ovaries, peritoneum, and other factors), fo - cused on the ovaries. However, there are no definite MRI find- ings suggestive of deep invasion of the ovary. In this study, we suggested histogram analysis focused on the viable ovary. Al- though further clinical studies are required, the value could be a marker for remnant viable ovarian tissue. The presented tech- nique could be valuable in determining the surgical approach for ovarian endometriosis. However, this study has several limitations. First, one en - rolled patient was too limiting to comprehensively understand the relationship between the imaging result and clinical out- come. Additionally, the case with unilateral ovarian involve- ment was included. For the validation of the effectiveness of the histographic analysis for prediction ability for pregnancy, a larger population is mandatory. Second, the perfusion map using subtraction technique requires strict adherence to a specific protocol with a radiology staff, trained to ensure that correct images are obtained, and the post-acquisition analysis is appropriate, including matching the exact pre- and post- contrast images to generate the subtracted image, which can be hampered by patient movement [30,31]. In summary, the parameters of the perfusion map including the area with perfusion ratio between 0.4 to 0.8, perfusion ratio, and integral density were related with discriminating severe endometrial invasion into the ovary from mild or mod- erate invasion, in addition to confirming the diagnosis of en- dometriosis showing the absence of enhancement. The appli- cation of the histographic analysis for the perfusion map of the pelvic MRI could be promising for the preoperative evalu- ation of remnant ovarian tissue in patients who desire ovarian preservation during the surgical treatment. In conclusion, the histographic analysis for the perfusion map of the pelvic MRI could be valuable in revealing the extent of the endometrial invasion and viable remnant ovarian tissue. Availability of Data and Material The datasets generated or analyzed during the study are available from the corresponding author on reasonable request. Conflicts of Interest The author has no potential conflicts of interest to disclose. ORCID iD Hyun Jung Lee https://orcid.org/0000-0002-3942-405X Funding Statement None Acknowledgments None

1. Yu JS, Rofsky NM. Dynamic subtraction MR imaging of the liver: advantages and pitfalls. AJR Am J Roentgenol 2003;180:1351- 1357. 2. Hummelshoj L, Prentice A, Groothuis P. Update on endometriosis. Womens Health (Lond) 2006;2:53-56. 3. Counseller VS, Crenshaw JL Jr. A clinical and surgical review of endometriosis. Am J Obstet Gynecol 1951;62:930-942. 4. Working group of ESGE, ESHRE and WES; Saridogan E, Becker CM, et al. Recommendations for the surgical treatment of en - dometriosis. Part 1: ovarian endometrioma. Hum Reprod Open 2017;2017:hox016. 5. Grammatikakis I, Evangelinakis N, Salamalekis G, et al. Prevalence of severe pelvic inflammatory disease and endometriotic ovarian cysts: a 7-year retrospective study. Clin Exp Obstet Gynecol 2009; 36:235-236. 6. Hornstein MD, Gleason RE, Orav J, et al. The reproducibility of the revised American Fertility Society classification of endometriosis. Fertil Steril 1993;59:1015-1021. 7. Society TAF. Revised American Fertility Society classification of en- 7 www.i-mri.org https://doi.org/10.13104/imri.2019.1030 dometriosis: 1985. Fertil Steril 1985;43:351-352. 8. Adamson GD, Pasta DJ. Endometriosis fertility index: the new, vali- dated endometriosis staging system. Fertil Steril 2010;94:1609- 1615. 9. Woodward PJ, Sohaey R, Mezzetti TP Jr. Endometriosis: radiologic- pathologic correlation. Radiographics 2001;21:193-216. 10. Sugimura K, Okizuka H, Imaoka I, et al. Pelvic endometriosis: de- tection and diagnosis with chemical shift MR imaging. Radiology 1993;188:435-438. 11. Sanchez AM, Viganò P, Somigliana E, Panina-Bordignon P, Vercelli- ni P, Candiani M. The distinguishing cellular and molecular fea- tures of the endometriotic ovarian cyst: from pathophysiology to the potential endometrioma-mediated damage to the ovary. Hum Reprod Update 2014;20:217-230. 12. Togashi K, Nishimura K, Kimura I, et al. Endometrial cysts: diagno- sis with MR imaging. Radiology 1991;180:73-78. 13. Guan X, Yu X, Liu X, Long J, Dai J. CT perfusion imaging and CT sub- traction angiography in the diagnosis of ischemic cerebrovascular disease within 24 hours. Chin Med J (Engl) 2003;116:368-372. 14. Lee VS, Flyer MA, Weinreb JC, Krinsky GA, Rofsky NM. Image sub- traction in gadolinium-enhanced MR imaging. AJR Am J Roent- genol 1996;167:1427-1432. 15. Eid M, Abougabal A. Subtraction images: a really helpful tool in non-vascular MRI. Egypt J Radiol Nucl Med 2014;45:909-919. 16. Sun X, Shi L, Luo Y, et al. Histogram-based normalization tech - nique on human brain magnetic resonance images from different acquisitions. Biomed Eng Online 2015;14:73. 17. Hauth EAM, Stattaus J, Kimmig R, Schmidt M, Forsting M. Mag- netic resonance imaging (MRI) of the pelvis in diagnosing endome- triosis. Zentralbl Gynakol 2005;127:76-82. 18. Kinkel K, Frei KA, Balleyguier C, Chapron C. Diagnosis of endome- triosis with imaging: a review. Eur Radiol 2006;16:285-298. 19. Nishio N, Kido A, Kataoka M, et al. Longitudinal changes in magnetic resonance imaging of malignant and borderline tumors associated with ovarian endometriotic cyst comparing with endometriotic cysts without arising malignancy. Eur J Radiol 2018;105:175-181. 20. Tanaka YO, Okada S, Yagi T, et al. MRI of endometriotic cysts in as- sociation with ovarian carcinoma. AJR Am J Roentgenol 2010;194: 355-361. 21. Bazot M, Bharwani N, Huchon C, et al. European Society of Uro- genital Radiology (ESUR) guidelines: MR imaging of pelvic endo- metriosis. Eur Radiol 2017;27:2765-2775. 22. Forstner R, Meissnitzer MW, Schlattau A, Spencer JA. MRI in ovar- ian cancer. Imaging Med 2012;4:59-75. 23. Messinger Y, Yanishevski Y, Avramis VI, et al. Treatment of human B-cell precursor leukemia in scid mice using a combination of the investigational biotherapeutic agent B43-PAP with cytosine ara- binoside. Clin Cancer Res 1996;2:1533-1542. 24. Togashi K, Nakai A, Sugimura K. Anatomy and physiology of the female pelvis: MR imaging revisited. J Magn Reson Imaging 2001; 13:842-849. 25. Outwater EK, Mitchell DG. Normal ovaries and functional cysts: MR appearance. Radiology 1996;198:397-402. 26. Chanana C, Gupta N, Bansal I, et al. Different sonographic faces of ectopic pregnancy. J Clin Imaging Sci 2017;7:6. 27. Schultes G. [Classification of endometriosis]. Wien Med Wochen- schr 1999;149:361-365. German. 28. Vercellini P, Fedele L, Aimi G, De Giorgi O, Consonni D, Crosignani PG. Reproductive performance, pain recurrence and disease relapse after conservative surgical treatment for endometriosis: the pre- dictive value of the current classification system. Hum Reprod 2006;21:2679-2685. 29. Fujishita A, Khan KN, Masuzaki H, Ishimaru T. Influence of pelvic endometriosis and ovarian endometrioma on fertility. Gynecol Ob- stet Invest 2002;53 Suppl 1:40-45. 30. Chan JH, Peh WC, Tsui EY, Wong KP, Yuen MK. Three-dimensional time-of-flight subtraction angiography of subacute cerebral hem- orrhage. AJR Am J Roentgenol 2003;181:242-244. 31. Cheng B, Cai W, Sun C, Kang Y, Gong J. 3D bone subtraction CT an- giography for the evaluation of intracranial aneurysms: a compari- son study with 2D bone subtraction CT angiography and conven- tional non-subtracted ct angiography. Acta Radiol 2015;56:1127- 1134.