Middle East Fertility Society Journal (2015) 20, 165–175

Middle East Fertility Society Middle East Fertility Society Journal

www.mefsjournal.org www.sciencedirect.com

ORIGINAL ARTICLE Magnetic resonance imaging for diagnosis of pelvic lesions associated with female

Manal Ezzat Badawy a, Dina Gamal Eldeen Elkholi b,*, Mohamed Fouad Sherif a, Mohamed Abd Elgeleel Hefedah a a Department of Radio-diagnosis, Faculty of Medicine, Tanta University, Tanta, Egypt b Department of Obstetrics and Gynecology, Faculty of Medicine, Tanta University, Tanta, Egypt

Received 18 December 2014; accepted 18 December 2014 Available online 4 March 2015

KEYWORDS Abstract Objective: To study the mandatory indications and accuracy of magnetic resonance Magnetic resonance imaging; (MR) imaging for diagnosis of pelvic lesions associated with . Tubal lesions; Design: Prospective cross-sectional study. Pelvic ; Main outcome measures: Indications and accuracy of MR imaging for diagnosis of pelvic lesions. Ovarian neoplasms Uterovaginal lesions were excluded as they were studied in a previous article (under publication). Materials and methods: 423 infertile women were investigated by hysterosalpingography (HSG), transabdominal and transvaginal ultrasonography after thorough clinical examination. Hundred and thirty (30.95%) patients were associated with pelvic lesions not conclusively diagnosed by HSG or/and ultrasonography and were examined by MR imaging to confirm the diagnosis. Fifty-four patients (41.53%) had uterovaginal lesions. They were discussed in a previous article. Seventy-six (58.46%) patients had other pelvic lesions. The present study concerned with these lesions. Results: The diagnosis by MR imaging was conclusive for 11 cases of pyosalpinx, 3 cases of , 25 cases of , 24 cases of ovarian , deep endometriosis of the of rectosigmoid (3 cases), urinary bladder (3 cases), one case of endometriosis of the abdom- inal wall after repeated cesarean sections and six ovarian tumors, 5 cases of benign cystic teratoma and 1 case of serous cystadenocarcinoma. The diagnosis of these lesions was confirmed by laparo- scopy or laparotomy and histopathological examination. Magnetic resonance imaging failed to diagnose peritubal adhesions in 22 out of 39 cases (56.41%) of tubal lesions and peritoneal implants of endometriosis in 12 out of 31cases (38.70%) of pelvic endometriotic lesions. They were discov- ered during the surgical treatment of the tubal and ovarian lesions through laparoscopy or laparo- tomy.

* Corresponding author at: 28nElnaser Street, Tanta, Egypt. Tel.: +20 1223661218, +20 040 3332445, +20 040 2250221. E-mail address: [email protected] (D.G.E. Elkholi). Peer review under responsibility of Middle East Fertility Society.

Production and hosting by Elsevier http://dx.doi.org/10.1016/j.mefs.2014.12.003 1110-5690 2015 Production and hosting by Elsevier B.V. on behalf of Middle East Fertility Society. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 166 M.E. Badawy et al.

Conclusion: The following pelvic lesions associated with female infertility were not conclusively diagnosed by HSG or/ and US but were precisely diagnosed by MR imaging with 100% accuracy: Pyosalpinx, hydrosalpinx and hematosalpinx, ovarian and deep infiltrating endometriosis and benign and malignant ovarian tumors as benign cystic teratoma and serous cystadenocarcinoma. 2015 Production and hosting by Elsevier B.V. on behalf of Middle East Fertility Society. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction The following radiological and imaging procedures were performed to diagnose pelvic lesions associated with infertility: Tubal and ovarian disorders are responsible for more than 70% of causes of infertility (1,2). Pelvic lesions are investigated (a) Transvaginal ultrasonography by 5–7 MHz sector trans- mainly by hysterosalpingography (HSG) and ultrasonography ducer and transabdominal ultrasonography by 3.5 MHz (US). However, with these methods the diagnosis may not be sector transducer for large pelvic-abdominal masses conclusive. These methods are supplemented with laparoscopy (Seimens Omnia, Medison 9900, Germany). and occasionally magnetic resonance (MR) imaging (2). Being (b) Hysterosalpingography to assess tubal and uterine an expensive procedure many investigators are reluctant to use conditions. MR imaging in this field. (c) Magnetic resonance imaging was performed when the Magnetic resonance imaging has the advantage of nonuse previous procedures did not give conclusive diagnosis. of ionizing radiation which is an important consideration in women of reproductive age. Another advantage is that MR imaging is less invasive than laparoscopy and less observer 2.1. Inclusion criteria dependent than the classic imaging techniques as US. Furthermore, recent advances in MR imaging with phased- All infertile patients with pelvic lesions not conclusively diag- array coil have created further imaging possibilities resulting nosed by HSG and US were included in the study. Magnetic in excellent spatial and tissue contrast resolution, multiplanar resonance imaging was mandatory for diagnosis of the follow- capability, and relatively fast techniques. The disadvantages ing pelvic lesions: of MR imaging is the high cost and the long time of examina- tion (2). Differentiation between pyosalpinx, hydrosalpinx and Magnetic resonance imaging is contraindicated in patients hematosalpinx. with pacemakers, cochlear implants, certain metallic objects Sure diagnosis and extent of pelvic endometriosis spe- and impaired renal function (2). cially ovarian and deep infiltrating varieties. The aim of the present study was to identify the obligatory Definite diagnosis of benign cystic teratoma and stage indications and assess the accuracy of MR imaging for the pre- 1a cystadenocarcinoma. operative diagnosis of pelvic lesions (after exclusion of uterovaginal lesions) associated with female infertility. As a rule we resorted to MR imaging if the other imaging modalities 2.2. Exclusion criteria did not provide conclusive diagnosis. Patients with chronic renal impairment (high serum creati- 2. Materials and methods nine), patients with previous allergy to the contrast medium and patients with pacemakers, cochlear implants, and certain metallic objects. Between June 2008 and April 2013, 423 infertile women were The diagnosis of the associated lesions was confirmed after investigated by HSG and US in the Department of operative interference by laparoscopy, and conventional Obstetrics and Gynecology and Department of abdominal operations and histopathological study. Radiodiagnosis, Tanta University Hospitals. Hundred and Operative laparoscopy was resorted for tubal lesions, ovar- thirty (30.73%) patients were associated with pelvic lesions ian and one case endometriosis of the urinary not conclusively diagnosed by HSG or/and US and were bladder all other lesions were managed by traditional surgery. examined by magnetic resonance (MR) imaging. Fifty-four patients (41.53%) had uterovaginal lesions. They were present- 3. Technique of hysterosalpingography (HSG) ed in a previous article (under publication). Seventy-six (58.46%) patients, discussed in the present study, had pelvic lesions other than uterovaginal morbidities. Before examined Hysterosalpingography was performed during the early fol- by HSG and US each patient was asked about the duration licular phase of the menstrual cycle, 3–4 days after cessation of infertility, whether primary or secondary. A detailed men- of menstrual flow. strual and obstetric history was taken with emphasis on previ- Fractionated technique under the guidance of an image ous pregnancies or history of any abortion and gynecological intensifier screen using iodinated water-soluble contrast medi- history with emphasis on genital tract infections. um (Telebrix Hystero) was the method of choice. The initial X- Detailed general, abdominal and pelvic examinations were ray was taken after the injection of 3 ml of contrast material to performed. Husband’s semen analysis was a pre-requisite visualize the cavity of the to exclude intrauterine adhe- before any investigation for the female partner. sions or a mass forming a filling defect. Further injection of MRI for diagnosis of pelvic lesions 167 contrast medium till the uterus and tubes were filled and dye The usual imaging time per patient was approximately spilled from the fimbrial ends of the tubes. During this frac- 24 min. The added examination time of diffusion-weighted tionated injection of the contrast medium, 2–3 serial films were imaging was approximately 4 min. taken. Another film was taken 20 min later after removal of Short time inversion-recovery (STIR) sequences: is based the speculum and cannula (1). on the rapid T1 recovery of fat and is, therefore, not substan- tially affected by field inhomogeneities. Unfortunately, STIR 4. Technique of pelvic MR imaging (3) suppresses all short T1 species, including tissues that have absorbed gadolinium. Hence, it is not possible to use STIR MR imaging was obtained on the available 1.5 T supercon- fat to improve the detection of contrast medium enhancement ducting MRI (Signa, GE Medical System, Milwaukee, Wis) (3). with a surface phased-array coil installed at Department of Radio-diagnosis, Tanta University Hospitals. 4.3. Interpretation of MR images: (3) It is recommended that an antiperistaltic (e.g. 1 US Pharmacopia unit of glucagon or 20 mg of scopolamine butyl- Image interpretation was done according to signal intensity, bromide) be administered intramuscularly or intravenously shape, site, size, pattern of enhancement and their relations before examination to minimize intestinal peristalsis. MR to surrounding structures at unenhanced and enhanced T1 imaging is contraindicated in patients with pacemakers, and T2 weighted images. cochlear implants, jewelry of any kind and certain metallic T1-weighted image was used as a localizer of the lesions. It objects. helped to characterize hemorrhage as in hematosalpinx and ovarian endometriomas which were both best demonstrated 4.1. 1-T1-weighted imaging on T1-weighted images to demonstrate high-signal intensity blood products. Repetition time (TR)/echo time msec (TE), 450–650/6.5– 17 ms; section thickness/intersection gap, 5/1.5 mm; acquisi- Contrast-enhanced T1-weighted images were used to tion time, 2.5 min) and field of view (FOV) ranged from 350 measure tumor size and detection of malignancy in to 400 (depending on the size of the patient). Postcontrast an . T1-weighted images (0.1 mmol/kg of gadolinium chelate) are Fat suppressed T1-weighted images were used in suspi- useful in selected cases, such as pelvic inflammatory disease, cious adnexal masses to detect its fatty or hemorrhagic evaluation of vascularity in uterine leiomyoma, and detection content (3). of malignancy in an adnexal mass. DWI and ADC map were used in cases of endometriosis, in 4.2. 2-Fat suppressed spin-echo T1-weighted imaging which there is no restricted diffusion with higher ADC value and in cases of neoplasm whether benign or malignant to dif- T1-weighted images obtained with a selective chemical fat- ferentiate between them as benign mass shows no restricted suppression technique are especially useful for detecting diffusion with higher ADC value than malignant one. hemorrhagic adnexal masses and should be obtained routinely The texture of the ovaries was clearly imaged in women of in patients with infertility in the axial and sagittal planes, reproductive age. The ovaries displayed hypointense stroma (repetition time (TR)/echo time (TE) 800/8; number of acqui- with hyperintense follicles on T2 weighted images. Normal fal- sitions 2 using a multislice technique, FOV: 38 · 38, echo train lopian tubes were not routinely imaged because of their small length (ETL) of 3. Slice thickness varied from 4 to 6 mm with a diameter and tortuous course. On T1-weighted images, the 0.8 to 1.2 mm interslice gap, matrix 512 · 256, bandwidth normal pelvic musculature and viscera demonstrated homoge- 62 KHz. neous low to medium signal intensity (3). 3-T2-weighted imaging in the axial, coronal and sagittal All participants signed a written consent. The study was planes (repetition-time (TR) msec/echo time (TE) msec approved by the ethics committee of Tanta University 4500/80, echo train length (ETL) of 8–18, number of acquisi- Hospitals. tions 3, FOV: 38 · 38, bandwidth: 41 KHz, matrix: 320 · 224. Some images were obtained in the coronal plane 5. Statistical analysis to assess the extent of the tumor. Diffuse weighted imaging (DWI): in some patients addi- Statistical analysis was performed to evaluate the sensitivity, tional DWI was performed, using a single-shot echo-planner specificity, positive and negative predictive values and diagnos- imaging (EPI) sequence, TR/TE inversion time (TI): tic accuracy of MRI in correlation to intervention; laparoscop- 7000/80 ms; FOV: 38 · 38 cm, matrix 192 · 224, bandwidth: ic or traditional operation, abdominal or vaginal and 250KHz, slice thickness 8 mm, interslice gap 0 mm, b value histopathological study of biopsies or excised lesions (4). of 0, 500, 800, and 1000 s/mm2. Apparent diffuse coefficient (ADC) map: ADC maps were generated on the scanner console using the b _ 1000,500, 800 6. Results and b _ 0 images. For each patient a region of interest (ROI) was 0.6-187; mean 20.2 cm2. The ROI was then manually cop- The study population comprised 76 infertile patients. Twenty- ied to the corresponding ADC map on the display monitor, five cases were found to have hydrosalpinx, 11 cases had pyos- from which the mean and lowest ADC values were obtained. alpinx, 3 patients had hematosalpinx, 24 patients had ovarian We investigated ADC values for each region. endometriomas (19 bilateral and 5 unilateral), 6 patients had 168 M.E. Badawy et al. deep endometriosis (urinary bladder 3 cases and rectosigmoid 7. Discussion 3 cases), endometriosis of the abdominal wall 1case, 1 patient had cystadenocarcinoma and 5 patients had benign cystic ter- Nondilated fallopian tubes are not usually seen on MR images atoma. The pelvic lesions found in these patients were accu- unless they are outlined by pelvic fluid. Dilated fallopian tubes rately diagnosed by MR imaging as the diagnosis was manifest as fluid-filled tubular structure that arises from the confirmed after operative interference and histopathological upper lateral margin of the uterine fundus and is separate from examination. Laparoscopic surgery was resorted for treatment the ipsilateral ovary. On MR imaging, they appear as of tubal lesions and ovarian endometriomas. Traditional sur- gery was used for treatment of other pelvic lesions. Magnetic resonance imaging failed to diagnose peritubal adhesions in 22 out of 39 cases (56.41%) of tubal lesions and peritoneal implants of endometriosis in 12 out of 31cases (38.70%) of pelvic endometriotic lesions. They were discovered during the surgical treatment of the tubal and ovarian lesions through laparoscopy or laparotomy.

Table 1 shows the clinical characteristics of the studied women. Table 2 shows MR imaging pattern of dilated tubes. MR imaging was able to discriminate hydrosalpinx from ovar- ian cysts and differentiated between hydrosalpinx and pyos- alpinx and hematosalpinx (Figs. 1–3). Table 3 depicts the features of MR imaging of ovarian endometriomas and endometriosis of abdominal wall, uri- nary bladder and rectosigmoid (Fig. 4a–c). Table 4 shows the MRI signal pattern of 6 ovarian neo- plasm, 1 serous cystadenocarcinoma and 5 benign cystic teratoma. Table 5 shows the accuracy, sensitivity, specificity, positive and negative predictive values of MR imaging for diagnosis of tubal diseases, endometriotic lesions, ovarian neoplasm, peritubal adhesions and peritoneal implants of endometriosis.

Table 1 Clinical characteristics of 76 studied women. Clinical characteristics Age (years) 26.33 (23–37) Body mass index (kg/m2) 26.40 (24–39) Figure 1 Bilateral pyosalpinx. (a)- contrast enhanced axial T1- Infertility Primary 43(56.5%) weighted image showing bilateral intrapelvic multilocular hypoin- Secondary 33 (43.42%) tense structures with irregular thick walls showing enhancement. Duration of infertility 1.2 (1–3) (b)- T1 weighted fat suppressed image showing high signal Gravidity 4.67 (2–7) intensity within the right adnexa which is likely due to high protein content of the fluid.

Table 2 Signal pattern of magnetic resonance imaging of disease. Pathology T1- weighted image T1- weighted fat suppression T2- weighted image Contrast – enhanced axial T1- weighted image Pyosalpinx Hyperintense tubular High signal intensity within Irregular thick wall tubular Mild enhancement of the N =11 structure the content which is hyperintense mass. wall. proteinaceous Hematosalpinx Hyperintense mass with No signal loss which is High signal intensity mass No change N =3 folds consistent with hemorrhage with folds Hydrosalpinx Low signal intensity cystic No change High signal intensity mass in No change N =25 mass has a folded and the adnexa, had a paddle and tortuous appearance. tortuous appearance. MRI for diagnosis of pelvic lesions 169

Figure 2 Hematosalpinx. (a)- axial T1 weighted image showing hyperintense mass with folds. (b)- axial T2 weighted image showing less hyperintense mass with folds compared to (a). (c)- axial T1 weighted with fat suppressed image showing hyperintense mass consistent with hemorrhage. retort-, sausage-, C-, or S-shaped cystic masses on MR imag- with low signal intensity, but the tubal content may have high ing. Thin longitudinally oriented folds along the interior of signal intensity if it is hemorrhagic (hematosalpinx) or the tube represent incompletely effaced mucosal or submucos- proteinaceous (pyosalpinx) (7). On T2-weighted images fluid al plicae. However, in severe cases, the plicae may be either content of hydrosalpinx has high signal intensity and slightly flattened or absent (5). hypointense relative to urine in the presence of hemorrhage If the diameter of the dilated tube reaches 10 cm, not an (hematosalpinx) or pus (pyosalpinx) (8). uncommon finding, hydrosalpinx may mimic a multilocular Hydrosalpinx may be a component of a complex adnexal ovarian tumor such as cystadenoma, the multiplanar capa- lesion. Pelvic inflammatory disease (PID) is one of the most bility of MR imaging can help to determine whether a multi- common causes of tubal and peritubal damage. In the acute locular cystic structure is actually the dilated tube, which phase of the disease, the fallopian tubes are swollen, edema- may be differentiated from an or dilated bowel tous and congested. As acute suppurative ensues, loop by its serpentine appearance. Identifying an ipsilateral the tubal lumen fills with pus and tubal fimbriae may adhere ovary that is separate from a lesion also may assist in the dif- to the ovaries. This may result in salpingo-, a condi- ferential diagnosis of hydrosalpinx and ovarian cyst (5). tion in which the tubes and ovaries partially adhere to one On T2-weighted images, the cystic nature of the lesion with another but remain largely separate and identifiable on images. its incomplete internal septa, which have low-signal intensity, If the fimbriae close before the ovary becomes extensively may be seen. A thickened wall of a dilated fallopian tube that involved, the inflamed tube forms a pyosalpinx. As the inflam- displays variable or heterogeneous signal intensity may be mation subsides, the pus undergoes slow proteolysis and the indicative of pyosalpinx or a component of a tubo-ovarian tube is filled with a thin serous fluid, the pyosalpinx thus is abscess. The wall and septa demonstrate contrast enhance- transformed into hydrosalpinx (9). ment, especially when inflammation is present. Presence of The detection of a hydrosalpinx within a complex adnexal an enhancing solid component is suggestive of fallopian tube mass may aid in the differential diagnosis as malignant tumors carcinoma. Enhancement of the dilated tubal wall surrounding are not associated with dilated tubes (10). a saclike cystic mass may be indicative of a tubal pregnancy. Endometriosis may affect the fallopian tube at two separate MR imaging may identify the cause of dilatation of the tube locations, with different consequences (11). The most common (5). types of tubal endometriosis are serosal and subserosal. Hydrosalpinx may be an isolated finding. On MR imaging Functional serosal and subserosal endometrial implants cause the signal intensity of the tubal fluid depends on the cause of repeated bouts of intralesional hemorrhage and fibrosis with obstruction (6). On T1-weighted images, the signal intensity resultant formation of peritubal adhesions and, thus, hydros- of the content of hydrosalpinx usually is that of simple fluid alpinx. A less common type of tubal endometriosis is 170 M.E. Badawy et al. /s 2 /s /s /s 2 2 2 10.3 mm mm mm mm · 3 3 3 10 10 10 · · · 0.88 0.77 0.93 Heterogeneous hyperintense mass of the posteriorurinary wall bladder of Hyperintense mass of the rectosigmoid Heterogeneous mass located dorsally to the sheathrectus of muscle Diffusion weighted imageHypointense cyst ADC map High signal intensity cyst ADC value 0.85 – 1.16 wall of the bladder rectosigmoid Hypointense mass located dorsally to the sheath of rectus muscle

Figure 3 Bilateral hydrosalpinx and hematosalpinx. (a)- Bilateral hydrosalpinx: axial T2-weighted image demonstrates bilateral intrapelvic wall multilocular hyperintense structure with thick irregular walls. (b)- Hematosalpinx: axial T2-weighted image with higher intensity of the contents compared to hydrosalpinx. image to hemorrhagic content NoNo Hypointense mass in the posterior Hypointense mass in the Isointense signal of the lesion to muscle. intraluminal. In this type, the tubal lumen in the intramural and isthmic segments is filled with endometrial tissue. Repeated intraluminal hemorrhages end in perifimbrial adhe- sions and hematosalpinx. Hyperintense tubal fluid seen on T1-weighted images is sug- gestive of hematosalpinx. On T2-weighted images, the signal intensity within the fallopian tube, with hematosalpinx, gener- ally is very high relative to the typical signal intensity of an endometrioma, which displays moderate to marked T2 shad- ing (12). Hematosalpinx has been reported to be an indication of endometriosis in certain cases and it may be the only imag- T2- weighted image Fat suppressed T1-weighted Low signal intensity cyst High signal intensity cyst due Hypointense mass in thewall posterior of the bladder Hyperintense mass in the rectosigmoid Heterogeneous lesion located dorsally to the sheathmuscle. of rectus ing finding indicative of endometriosis (4,12). In the present study we encountered 25 cases of hydros- alpinx due to PID, 3 cases of hematosalpinx due to tubal endometriosis as no cause was detected on MR imaging and =1

11 cases of pyosalpinx (Table 2, Figs. 1–3). The diagnosis of N

these lesions was confirmed on laparoscopy together with Signal pattern of magnetic resonance imaging of endometriotic lesions. histopathological examination. Thus, in the present study, the accuracy, sensitivity, specificity, positive and negative pre- =24 dictive values of MR imaging for diagnosis of tubal dilatation, =3 =3 Table 3 Ovarian Endometriomas N Bladder endometriosis N Rectosigmoid endometriosis N ADC = apparent diffuse coefficient. type and cause, were 100%. Abdominal wall endometriosis MRI for diagnosis of pelvic lesions 171

Figure 4 Endometrioma of the ovary. (a)- Fat suppressed axial T1-weighted image shows high signal intensity cysts, indicating their hemorrhagic content. (b)- Axial T2-weighted image showing two endometrial cysts of low signal intensity. (c)- Apparent diffusion coefficient (ADC) maps showing two endometrial cysts of high signal intensity. ADC value was 1.16 · 10.32/s.

Endometriosis affects about 4–13% of all women in repro- and the viscosity of the contents in the endometrioma ductive age and 25–50% of women with infertility problem. It caused T2 shortening and produced shading. commonly occurs in pelvis, ovaries, ovarian fossa, pouch of Douglas and tubes. Deep infiltrating endometriosis may affect MR imaging showed a sensitivity of 92% and specificity of urinary bladder, ureters and bowel (13). 98% for diagnosis of endometrioma (17). In the present study The appearance of ovarian endometriomas on MR images the histopathological examination after operative treatment is variable and depends on the concentration of iron and pro- through laparoscopy denoted that the accuracy, sensitivity, tein in the fluid, products of blood degeneration. Most ovarian specificity, positive and negative predictive values of MR imag- endometriomas have the growth appearance of chocolate cysts ing for diagnosis of endometriomas were 100% (Table 5). represents highly concentrated blood products (13–15). In the MR imaging is not sensitive for superficial peritoneal present study 24 cases of ovarian endometriomas were discov- implants of endometriosis; therefore MR imaging should not ered, 19 bilateral and 5 unilateral (Table 2, Fig. 4a–c). In the be relied on to rule endometriosis, laparoscopy is the gold stan- present study on T1-weighted MR image ovarian endometri- dard for diagnosis of pelvic endometriosis but sometimes the oma had high signal intensity similar to that of fat. On fat sup- endometrial implants are concealed by dense adhesions. (18). pressed T1-weighted image the high signal intensity of According to the definition of Bazot et al. (19) deep infil- endometrioma did not decrease (Fig. 4a). This signal charac- trating endometriosis (DIE) is defined as an endometriotic teristic differentiates endometriomas from fatty adnexal mass- lesion penetrating into the retroperitoneal space or the wall es such as benign cystic teratomas (14). In the present study on of pelvic organs to a depth of at least 5 mm. The lesions of T2-weighted image ovarian endometriomas had low signal DIE can present as nodules, masses, cysts or adhesions intensity (Fig. 4b). This low signal intensity on T2-weighted between organs. Signal characteristics vary as a function of images is called shading and occasionally occurs in a gradient the distribution of stromal tissue versus glandular elements, from higher to lower signal intensities (15). This pattern of sig- the presence of hemorrhage, and the extent of an inflammatory nal intensities results from the high content of iron concentra- reaction. Pure fibrous lesions have hypointense signal intensity tion in the endometrioma and is rarely seen in other masses of on T1- and T2- weighted imaging whereas lesions with a strong any type (15). glandular component having hyperintense signals on both T1 and T2. The most frequently encountered lesions, however, The findings in the present study matched with the findings are mixed with the fibrous component manifesting hypointense in other studies (13–15). Imaoka et al. (16) found that signals on T1 and hemorrhagic foci demonstrating hyperin- endometriomas on MR imaging were high signal intensity tense signals on T2 (13). cysts on T1- and T2- weighted imaging or high signal inten- In the present study when the diagnosis of deep endometri- sity on T1-weighted imaging with or without fat suppres- otic lesions was not conclusive by transvaginal US. MR imag- sion and low signal intensity on T2-weighted imaging ing was performed and the diagnosis was conclusive and was (shading). The dense concentration of cyclic hemorrhage confirmed by laparoscopic or conventional surgery and 172 M.E. Badawy et al.

histopathological examination. The following varieties of deep endometriosis were encountered in the present study:

a. Three cases of endometriosis of the rectosigmoid colon, MR features showed that infiltrating endometriosis was /s

2 hyperintense on T2-weighted imaging and isointense or hypointense with no fat suppression on T1-weighted imag- mm 3

ing (Table 2; Fig. 5). They presented with rectal

10 and dyschesia. Rectosigmoid endometriosis was associated · with ovarian endometriosis in the three cases. intensity ADC value was 0.86 The rectosigmoid colon was the second most commonly affected site of DIE in the study of Chamie et al. (20) the inci- dence was 19%. In rectosigmoid assessment, the degree of rec- tal wall infiltration, the size of the rectal lesion, and the distance from the anal border can influence the surgical man- agement (21). All these factors were accurately identified by MR imaging. The diagnosis was confirmed after endoscopical- ly directed biopsy and histopathological examination with accuracy of 100% (20) and 89.1% (21).

Hyperinten-se mass The tumor is of low signal In the present study the accuracy, sensitivity, specifici- ty, positive and negative predictive values of MR imaging for diagnosis of endometriosis of the rectosigmoid colon were 100% (Table 5), the same as that of Chamie et al. (20).

b. Three cases of urinary bladder endometriosis manifested as a mass on the posterior aspect of the bladder, the mass was hypointense on T2-weighted imaging and isointense with no fat suppression on TI-weighted imag-

Heterogeneous high-signal intensity lesion pelvic complex and a heterogeneous intermediate to bright small intracystic solid component and signal, complete uniform septum is seen at its inferoposterior aspect ing (Table 3; Fig. 6).

The patients were complaining of infertility and urge incon- tinence that was marked in the premenstrual and menstrual periods. Cystoscopic examination (inter-menstrual) did not reveal any abnormalities. MR imaging discovered the condi- tion. Cystoscopy, during menses, demonstrated the mass that did not invade the bladder mucosa. The mass, in one case, was excised through laparoscopy and the other two cases were submitted to laparotomy. Histopathological examination con- firmed the diagnosis. The accuracy, sensitivity, specificity, Markedly suppressed the small peripheral linear hyperintense lesion on the T1 fat saturated sequence No Hyperintense well defined positive and negative predictive values of MR imaging for diagnosis of this lesion were 100% in the present study. Bladder endometriosis defined as bladder wall involvement with invasion of the detrusor muscle, is a rare condition consti- tuting roughly 6% of all endometriosis cases (22). In 40% of cases the urinary symptoms (urgency, frequency, hematuria, or urge incontinence) are cyclic related to menses, 60% will present with non-cyclic symptoms. Typical MR fea- tures include localized or diffuse bladder wall thickening involving the dome/posterior wall and heterogeneous T2- Hyperinte-nse well defined pelvic complex predomin- antly cystic mass and hypointense small intracystic solid component A low signal intensitywith lesion small hyperinten-se peripheral linear lesion isointensity with occasional T1 hyperintense foci. MR also evaluates the distal ureters which may become obstructed, more often by associated mass effect rather than by direct invasion by endometriosis. This is pertinent to surgical plan- ning as re-implantation of the ureter may be required (23) The accuracy of MR imaging for diagnosis endometriosis of the urinary bladder was 98.6% (22) and 99.2% (23). Signal pattern of magnetic resonance imaging of 6 ovarian tumors.

c. One case of abdominal wall endometriosis, consequent =1 =5 to repeated cesarean section, the lesion was located dor- Table 4 Malignant ovarian neoplasm (serous cystadenocarcinoma) N PathologyBenign cystic teratoma N T1-weighted Fat suppressed T1-weighted T2-weighted Diffusion weighted imaging ADC map ADC = apparent diffuse coefficient. sally to the rectal muscle. It showed heterogeneous MRI for diagnosis of pelvic lesions 173

Table 5 Accuracy of MR imaging for diagnosis of pelvic lesions associated with infertility. Accuracy (%) Sensitivity (%) Specificit (%)y PPV (%) NPV (%) Hydrosalpinx 100 100 100 100 100 Pyosalpinx 100 100 100 100 100 Hematosalpinx 100 100 100 100 100 Peritubal adhesions 0 0 0 0 0 Endometriosis: -endometriomas 100 100 100 100 100 -urinary bladder 100 100 100 100 100 -rectosigmoid 100 100 100 100 100 -abdominal wall 100 100 100 100 100 -peritoneal implants 0 0 0 0 0 Ovarian neoplasm: -benign 100 100 100 100 100 -malignant 100 100 100 100 100

Figure 5 Endometriosis of the rectosigmoid. Axial T1-weighted image showing bilateral adnexal cysts displaying high signal intensity, one cyst is infiltrating the rectosigmoid and producing Figure 7 Endometriosis of the lower abdominal wall. Axial T2- rectal bleeding. weiaghted image shows a hypointense lesion located dorsally to rectus muscle.

Figure 6 Urinary bladder endometriosis: axial T2- weighted images showing a hypointense mass at the posterior aspect of the Figure 8 Benign cystic teratoma: axial T2 showing heteroge- urinary bladder. neous high-signal intensity lesion. 174 M.E. Badawy et al.

Figure 9 cystadenocarcinoma: a- Axial T1-weighted image showing hypointense pelvic complex predominantly cystic mass with a hypointense intracystic component. b and c- Axial and sagittal T2-weighted images showing hyperintense pelvic complex mass predominantly cystic with heterogeneous intermediate to bright intracystic solid component and evidence of a septum is seen at its inferoposterior aspect.

signal intensity on T2-weighted imaging and isointense the presence of both solid and cystic areas within the mass; signal to muscle on fat suppressed T1-weghted imaging necrosis (hyperintense on T1- and T2- weighted MR imaging) (Table 3; Fig. 7). within a solid lesion; papillary projections from the wall or sep- tum of a cystic lesion; an irregular septum or wall; multiple The patient with abdominal wall endometriosis presented thickened (>3 mm) septations; a large size (>6 cm); bilateral with a palpable mass in the scar. The mass was cyclic. lesions; ascites and enlarged pelvic and para-aortic lymph Endometriosis of the abdominal wall may be difficult to diag- nodes which can be easily detected on MR imaging and thus nose; it is often mistaken clinically and with diagnostic imaging assist in staging a malignant ovarian tumor (25). On explo- for other abnormal conditions such as a suture granuloma, an ration the surgical staging of cystadenocarcinoma, in the pre- incisional hernia or primary or metastatic cancer. Accuracy of sent study, was 1a. The early stage may explain the absence diagnosis with MR imaging approached 100% (24), the same of many features of malignancy on MR imaging of this malig- degree of accuracy in the present study. nant tumor. Certain features of MR imaging predominate for each type After excision by laparotomy and histopathologic study of of ovarian tumor. Knowledge of these key features of ovarian the six ovarian tumors included in the present study the accu- tumors may allow a specific diagnosis or substantial narrowing racy, sensitivity, positive and negative predictive values of MR of the differential diagnosis. Characterization of ovarian imaging for diagnosis of these ovarian tumors were 100%. The tumors can aid to surgical planning, whether exploration or accuracy of MR imaging for diagnosis of the specific type of laparoscopic excision, and may help distinguish benign from ovarian tumors both benign and malignant by Jung et al. malignant tumors and thus avoid inappropriate management (25) was 98.7%. (25). In the present study magnetic resonance imaging failed to In the present study conclusive diagnosis, with 100% accu- diagnose peritubal adhesions in 22 out of 39 cases (56.41%) racy, of 5 cases of benign cystic teratomas, and 1 case of cys- of tubal lesions and peritoneal implants of endometriosis in tadenocarcinoma was obtained as they showed the 12 out of 31cases (38.70%) of pelvic endometriotic lesions. characteristic MR imaging features of these tumors (Table 4; They were discovered during the surgical treatment of the Figs. 8 and 9a–c). tubal and ovarian lesions through laparoscopy or laparotomy. Other studies reported similar features on MR imaging for The accuracy of MR imaging for diagnosis of pelvic lesions the two types of ovarian tumors. They found that manifesta- associated with infertility was 100%, This may be explained by tions of malignant adnexal mass on MR imaging included MRI for diagnosis of pelvic lesions 175 the small number of women studied and the use of all modes of (10) Ha HK, Lim GY, Cha ES. MR imaging of the tubo-ovarian MR imaging and of the phased array coil with spatial and tis- abscess. Acta Radiol 1995;36:510–4. sue contrast resolution and multiplanner capability. (11) Clement PB. Diseases of the peritoneam. In: Blaustein A, editor. The limitation of this work was the small number of cases Blaustein’s pathology of the genital tract. NY, New York: Springer-Verlag; 1984. p. 729–89. enrolled in the study. (12) Outward EK, Stiegelman ES, Chiorvanich P. Dilated fallopian tubes: imaging characteristics. Radiology 1998;208:463–9. 8. Conclusion (13) Gougoutas CA, Seigelman ES, Hunt J, Outward EK. Pelvic endometriosis: various manifestations and MR imaging finding. The following pelvic lesions associated with female infertility AJR 2000;175:353–8. were not conclusively diagnosed by HSG or/ and US but were (14) Practice Committee of the American Society for Reproductive Medicine: Endometriosis and Infertility. Fertil Steril 2006; 5 precisely diagnosed by MR imaging with 100% accuracy: (Suppl 1): S156-S161. Pyosalpinx, hydrosalpinx and hematosalpinx, ovarian and (15) Bia KG, Vrachliotis TG, Agrawal R, Shetty AN, Maximovich R, deep infiltrating endometriosis and benign and malignant Hricak H. Pelvic endometriosis. MR imaging spectrum with ovarian tumors as benign cystic teratoma and serous laparoscopic correlation and diagnostic pitfalls. Radiographics cystadenocarcinoma. 1997;17:639–55. (16) Imaoka I, Wada A, Matsuo M. MR imaging of disorders Conflict of interest associated with female infertility; use in diagnosis, treatment and management. Radiograghics 2003;23:1401–21. (17) Siegelman ES, Outward EK. Tissue characterization in the female No conflict of interest declared. pelvis by means of MR imaging. Radiology 1999;212:5–18. (18) Siegelman ES, Outward EK, Wang T, Mitchell DG. Solid pelvic References masses caused by endometriosis: MR imaging features. AJR 1994;165:357–61. (1) Gibbs KS, Karian TY, Haney AF, Nygaard I. Danforth’s (19) Bazot M, Darai E, Hourani R, Thomassin I, Cortez A, Uzan S. Obstertrics and Gynecology. 10th ed. Philadelphia, Pa: Deep pelvic endometriosis: MR imaging for diagnosis and Lippincott Williams & Wilkins; 2012, pp. 649–667. prediction of the extension of disease. Radiology (2) Woodward PJ, Wagner BJ, Farley TE. MR imaging in the 2004;232:379–89. evaluation of female infertility. Radiographics 1993;13:293–310. (20) Chamie LP, Blasbalg R, Goncalves MOC, Carvalho FH, Abroa (3) Steinkeler JA, Woodfield CA, Lazarus E, Hillstrom MM. Female MS, Oliveira IS. Accuracy of magnetic resonance imaging for infertility: a systemic approach to radiologic imaging and diagnosis and preoperative assessment of deeply infiltrating diagnosis. Radiographics 2009;29:1353–70. endometriosis. Int J Gynaecol Obstet 2009;106:198–200. (4) Anderson MC, Robby SJ, Russell P, editors. Pathology of the (21) Chapron C, Fauconnier A, Viera M, Baracat M, Dousset B, female reproductive tract. England: Churchill Livingstone; 2002. Pansioni V, Lacha-Lavenu MC, Dubuisson JB. Anatomical p. 415–44. distribution of deeply infiltrating endometriosis: surgical implica- (5) Kim MY, Rha SE, Oh SN, Jung SE, Lee YJ, Kim YS, Byun JY, tion and proposition for a classification. Hum Reprod Lee A, Kim MR. MR imaging findings of hydrosalpinx: a 2003;18:157–61. comprehensive review. Radiographics 2009;29:495–507. (22) Fauconnier A, Chapron C, Dubuisson JB, Viera M, Dousset B, (6) Brown MA, Ascher SM. Adnexa. In: Semelka RC, editor. Breat G. Relation between symptoms and the anatomic location Abdominal-pelvic MRI. Hoboken, NL: Willey-Liss; 2006. p. of deep infiltrating endometriosis. Fertil Steril 2002;78:719–26. 1334–79. (23) Kinkel K, Chapron C, Bailleyquier C, Faitel X, Dubuisson JB, (7) Hamm B, Forstner R, Kim E. MRI and CT of female pelvis. J Moreau JF. Magnetic resonance imaging characteristic of deep Nucl Med 2008;49:779–862. endometriosis. Hum Reprod 1999;18:1081–6. (8) Forstner R, Sattlegger P. In: Heuck A, Reiser M, editors. (24) Streelatha S, Tejaswini BM, Nethra HN, Saidhar L. Case report Abdominal and pelvic MRI. Berlin, Germany: Springer-Verlag; on rectus sheath endometrioma with chocolate cyst of the ovary. 1998, p. 247–81. Scholar J Medial Case Rep 2014;2:1112–3. (9) Wheeler JE. Pathology of the fallopian tube. In: Blaustein A, (25) Jung SE, Lee JM, Rha SE, Byun JY, Jung JI, Hahn ST. CT and editor. Blaustein’s pathology of the genital tract. New York, NY: MR imaging of ovarian tumors with emphasis on differential Springer-Verlag; 1984, p. 393–411. diagnosis. Radiographics 2000;22:1305–23.