Variations in Appearance of Endometriomas

Transcription

1 CME rticle Variations in ppearance of Endometriomas Elizabeth sch,, Deborah Levine, MD Objective. ecause of the range of patient ages with endometriosis, the persistence of endometriomas, and the degradation of internal blood products over time, the appearance of endometriomas is variable. The purpose of this study was to identify the prospective diagnoses in cases of surgically proven endometriomas and to illustrate the variety of appearances of endometriomas. Methods. Sonographic images from 325 women with histologic confirmation of an adnexal mass during a 16- month period were reviewed. Eighty-seven endometriomas in 71 women were found. Prospective diagnoses were scored as follows: 1, the diagnosis was endometrioma only; 2, the differential diagnosis listed endometrioma first; 3, the differential diagnosis listed endometrioma but not first; 4, the diagnosis did not include endometrioma but included a complex or hemorrhagic cyst; and 5, the diagnosis did not include endometrioma. Lesions with the classic appearance of an endometrioma (hypoechoic mass with diffuse low-level internal echoes) or an atypical appearance were chosen for illustration. Results. Thirty-five (60.3%) of 58 endometriomas measuring at least 2 cm in greatest dimension included endometrioma in the differential diagnosis at sonography. In 6 (10.3%) of 58, the lesions were described as complex or hemorrhagic cysts. typical endometriomas included cases with retracted clots that appeared solid but without blood flow, endometriomas in pregnant patients, and endometriomas in postmenopausal women with heterogeneous internal echoes and central calcification. ruptured endometrioma was interpreted as pelvic inflammatory disease. Five endometriomas were suspected to be malignancies because of a solid appearance, blood flow, surrounding adhesions, and, in 1 pregnant patient, extensive decidualization. Conclusions. Recognition of the varied appearance of endometriomas should aid the interpreting physician in giving an appropriate prospective diagnosis of endometriomas. Key words: adnexa; endometrioma; endometriosis; ovary; sonography. Received March 12, 2007, from Harvard Medical School, oston, Massachusetts US (E..); and Department of Radiology, eth Israel Deaconess Medical Center, oston, Massachusetts US (D.L.). Revision requested March 29, Revised manuscript accepted for publication pril 13, This study was supported by grants from the CRICO Risk Management Foundation and the Harvard Medical School PSTEUR and Office of Enrichment programs. ddress correspondence to Deborah Levine, MD, Department of Radiology, eth Israel Deaconess Medical Center, 330 rookline ve, oston, M US. [email protected] CME rticle includes CME test Sonography is the imaging method of choice for initial diagnosis of endometriosis. 1 ecause of the range of ages affected by endometriosis, the persistence of endometriomas, and the degradation of internal blood products over time, there are a variety of appearances of endometriomas. The characteristic appearance of an endometrioma on sonography is that of a hypoechoic mass containing diffuse low-level internal echoes, the so-called chocolate cyst. However, many other appearances have been described, including an anechoic cyst, a solid-appearing mass, solid elements in a cyst with low-level internal echoes, and punctuate echogenic foci in the wall of the cyst. 2 The most common misdiagnoses of endometriomas by sonography are hemorrhagic cysts and dermoids. 1 It is important to accurately diagnose endometriomas because the treatment or surgical approach may differ if this diagnosis is suspected before surgical removal. ecause of the risk of malignant transformation, 3 endometriomas are fre by the merican Institute of Ultrasound in Medicine J Ultrasound Med 2007; 26: /07/$3.50

2 Variations in ppearance of Endometriomas quently removed. This study assessed the spectrum of sonographic appearances and preoperative impressions of endometriomas in a series of surgically removed adnexal masses. Materials and Methods This study was approved by our Institutional Review oard and was compliant with the Health Insurance Portability and ccountability ct. Given the retrospective nature of the study, informed consent was not obtained. Patient and Study Selection Pathology reports from surgeries performed at eth Israel Deaconess Medical Center between January 1, 2004, and May 9, 2005, containing the words pelvic, pelvis, uterus, ovary(ies), tube(s), adnexa, and adnexal were searched. Of these reports, those from women who had at least 1 pelvic sonographic examination at eth Israel Deaconess Medical Center in the 5 years preceding surgery and for whom an adnexal mass was reported at pathologic examination were included. Up to 3 sonograms preceding surgery were included for each patient. total of 325 women with 650 sonographic studies were reviewed. The final diagnosis was determined by a combination of histologic and surgical findings. In 79 lesions, the findings of endometrioma at histologic examination were used for final diagnosis. When the findings at pathologic examination did not correlate with the imaging findings, the surgical note was reviewed. In 8 cases, endometriomas were diagnosed at surgery (and not by histologic examination) by the visual finding of a thick-walled chocolate cyst, usually accompanied by adhesions. 4 Patient Data Data from sonography reports included the date of the study, indication for imaging study, patient age, pregnancy status, lesion size, and differential diagnosis. Data from pathology reports included the date of surgery, indication for surgery, lesion size, and histologic diagnoses. Patients included in Results were those with a final diagnosis of endometrioma or those for whom the written impression suggested endometrioma. Report Review The sonography reports were reviewed for the size of the lesion and the impression. Impressions were scored as follows: 1, giving the diagnosis of only endometrioma for the lesion; 2, giving the diagnosis of endometrioma first in a list of differential diagnoses; 3, giving the diagnosis of endometrioma in a list of differential diagnoses; 4, not listing endometrioma but listing a complex or hemorrhagic cyst; and 5, not listing endometrioma at all. The impressions of those without endometriomas listed in the differential diagnosis were tabulated. We repeated this assessment for only those lesions larger than 2 cm in maximum diameter. Image Review Cases with classic and atypical appearances of endometriomas were chosen from this group of histologically and surgically proven endometriomas to illustrate the various appearances of endometriomas. Examples with retractile clots, endometriomas in pregnancy, endometriomas in postmenopausal women, and endometriomas with associated lesions, dense adhesions, a solid appearance, and ruptures were chosen to show the range of appearance of endometriomas. solid malignancy that arose in an endometrioma was not included in the tally of atypical appearances of endometriomas because the malignant appearance was more important than the classification as endometriosis. Results Patient Data Endometriomas were the most common surgically removed lesions, with 87 endometriomas in 71 women. ges ranged from 23 to 73 years, with a mean age ± SD of 42 ± 10.5 years. Intervals between the most recent sonographic study and surgery ranged from 0 to 4.5 years with an average interval of days. Clinical indications for surgery included adnexal cysts (n = 25), adnexal masses (n = 21), pain (n = 6), pelvic masses (n = 5), endometriosis (n = 7), uterine fibroids (n = 3), and 1 case each, an enlarged ovary, bleeding, cancer, and a dermoid. Four women were pregnant at the time of imaging, 3 at gestational ages of 5, 7, and 33 weeks and 1 with 2 exams during pregnancy at 6 and 7 weeks. 994 J Ultrasound Med 2007; 26:

3 sch and Levine Two diagnoses of endometrioma occurred in women with concurrent histologic diagnoses of adnexal malignancy, 1 ipsilateral (serous papillary adenocarcinoma arising from an endometrioma), and 1 contralateral (neuroendocrine tumor). Other concurrent adnexal lesions included 2 cases of inclusion cysts and 1 case each of cystadenofibroma, adenofibroma, fibroma, a corpus luteal cyst, and hydrosalpinx. In the 79 lesions reported at pathologic examination, the sizes of the endometriomas ranged from 0.3 to 11 cm in largest dimension, as well as 1 microscopic focus of endometriosis. The average lesion size was 4.4 ± 2.78 cm. In the 67 lesions where a cyst was measured on the sonogram closest to the surgery, the sizes ranged from 0.9 to 13.5 cm, with an average size of 4.9 ± 2.87 cm. Of these, 58 lesions were larger than 2 cm. Report Review total of 138 sonograms of endometriomas were reviewed in 71 women. Twenty-eight women had 1 study; 19 had 2 studies; and 24 had 3 studies. Indications for the studies preceding surgery in order of frequency included known ovarian cysts (n = 26), pain (n = 14), bleeding (n = 7), known adnexal masses (n = 6), suspected/known endometriosis (n = 5), fibroid uteri (n = 4), pelvic masses (n = 3), changes in ovarian size (n = 2), and 1 case each of a previous endometrioma, an enlarged uterus, abdominal distention, and an elevated cancer antigen 125 level. Table 1 gives the sizes of lesions with respect to how they were described on the sonography reports closest in proximity to the surgery. The following scores were given to the written reports for endometriomas measuring larger than 2 cm: 1 (n = 23), 2 (n = 6), 3 (n = 6), 4 (n = 6), and 5 (n = 17). Figures 1 and 2 show examples of lesions diagnosed as endometriomas prospectively. Thirty-five (60.3%) of 58 endometriomas measuring at least 2 cm in greatest dimension included endometrioma in the differential diagnosis at sonography (Table 2). In an additional 6 (10.3%) of 58, the lesions were described as complex or hemorrhagic cysts. Table 3 lists the first item on the differential diagnosis for endometriomas measuring larger than 2 cm. Image Review Image review of 87 cases of surgically proven endometriomas revealed 13 lesions that were considered atypical. Of these, 6 atypical endometriomas were diagnosed correctly on prospective sonography reports. These included 2 endometriomas with retracted clots, which appeared solid but without blood flow (Figures 3 and 4), 2 endometriomas in pregnancy (Figures 5 and 6), and 2 endometriomas in postmenopausal women (Figure 7). Of the 2 endometriomas in postmenopausal women, 1 was considered atypical because of heterogeneous internal echoes, and the other was considered atypical because of central calcification. Seven atypical endometriomas were diagnosed incorrectly on prospective sonography reports. One endometrioma in a 73-year-old woman was diagnosed as a complex cyst without specific etiology. ruptured endometrioma was interpreted as pelvic inflammatory disease because of the patient s febrile status and a pelvic fluid collection seen on computed tomography (Figure 8). Five atypical endometriomas appeared suspicious for malignancy on sonography. The lesions in these patients appeared of high concern because of extensive adhesions (n = 2; Figure 9) Table 1. Prospective Diagnosis of Endometriomas With Respect to Size (87 Endometriomas in 71 Women) Mean Size (Range) at Mean Size (Range) on Differential Diagnosis (ddx) No. of Cases (%) Pathologic Examination, cm Most Recent Sonogram, cm Endometrioma only lesion listed in ddx 28 (32.2) 4.7 ( ) 4.7 ( ) Endometrioma listed first in a list of ddx 10 (11.5) 4.0 ( ) 4.4 ( ) Endometrioma mentioned but not first in list of ddx 9 (10.3) 4.8 ( ) 6.4 ( ) No mention of endometrioma but complex/hemorrhagic 9 (10.3) 5.2 ( ) 5.6 ( ) cyst listed in ddx No mention of endometrioma or complex/hemorrhagic 31 (35.6) 4.0 ( ) 4.6 ( ) cyst in ddx J Ultrasound Med 2007; 26:

4 Variations in ppearance of Endometriomas Figure 1. Classic endometrioma. Transvaginal transverse image shows a classic-appearing endometrioma (calipers) with diffuse lowlevel internal echoes. The prospective diagnosis was endometrioma. and a solid appearance and blood flow (n = 3; Figures 10 and 11). 4-cm papillary serous carcinoma contained in an endometriotic cyst was not classified as an atypical endometrioma because it was a superimposed lesion (Figure 12). Discussion Diffuse low-level internal echoes are the characteristic sonographic findings in endometriomas. Patel et al 5 showed this sonographic appearance Figure 2. Classic endometrioma. Transvaginal transverse image shows a complex ovarian cyst (calipers) with a histologic diagnosis of endometrioma. The prospective differential diagnosis listed endometrioma first, followed by hemorrhagic cyst, dermoid, and mucinous tumor. It is possible that the slightly more coarse appearance of these echoes made the differential diagnosis longer than that of the classic endometrioma in Figure 1. in 95% of endometriomas in a retrospective review. That study also noted that adnexal masses with diffuse low-level internal echoes, as well as hyperechoic wall foci and multilocularity, without other neoplastic features were 32 times more likely to be endometriomas than any other adnexal lesions. 5 prospective study by Dogan et al 6 determined the positive predictive value of sonography in diagnosis of surgically proven endometriomas to be 91.5% overall and 97% for classic-appearing endometriomas (round shape, regular margins, thick walls, and homogeneous low-level internal echoes) but only 70.9% for atypical-appearing lesions (anechoic appearance, internal septations, and irregular margins). In our review of reports, 60% (35/58) of endometriomas measuring at least 2 cm in greatest dimension included endometrioma in the differential diagnosis at sonography. This is likely due to the fact that we assessed the prospective reports, the length of time between imaging and surgery (up to 5 years), and the wide range of expertise of the radiologists interpreting the sonograms at our institution. Lesions that were not diagnosed definitively as endometriomas by sonography were frequently diagnosed nonspecifically as complex cysts (6/58 [10%]), hemorrhagic cysts (4/58 [7%]), and neoplasms (4/58 [7%]). lthough punctuate calcifications have been described in the walls of endometriomas, it is important to recognize that endometriomas also may have central calcification. Therefore, although the presence of focal calcification suggests the diagnosis of teratoma, this does not exclude endometrioma. 7 n endometrioma can be distinguished from a teratoma by assessment of the distribution of hyperechoic and hypoechoic fluid within the cyst if a fluidfluid level is present. In an endometrioma, the supernatant fluid layer should be hypoechoic, with a hyperechoic dependent layer representing blood. In a dermoid, the supernatant layer will be echogenic, representing fat (Figure 13). 7 typical endometriomas may also be mistaken for malignancy, particularly in pregnant patients with decidualized endometriosis. 8 Sonographic findings include solid, vascular- 996 J Ultrasound Med 2007; 26:

5 sch and Levine Table 2. Prospective Diagnosis of Endometriomas Larger Than 2 cm With Respect to Size (57 Endometriomas) Mean Size (Range) at Mean Size (Range) on Differential Diagnosis (ddx) No. of Cases (%) Pathologic Examination, cm Most Recent Sonogram, cm Endometrioma only lesion listed in ddx 23 (39.7) 5.0 ( ) 5.1 ( ) Endometrioma listed first in a list of ddx 6 (10.3) 4.9 ( ) 4.9 ( ) Endometrioma mentioned but not first in list of ddx 6 (10.3) 6.1 ( ) 6.6 ( ) No mention of endometrioma but complex/hemorrhagic 6 (10.3) 5.2 ( ) 7.0 ( ) cyst listed in ddx No mention of endometrioma or complex/hemorrhagic 17 (29.3) 5.2 ( ) 5.4 ( ) cyst in ddx ized components arising from the wall and extending into the cyst. 9 Cancer antigen 125 levels are not diagnostically useful during pregnancy, particularly during weeks 11 through 14, during which cancer antigen 125 levels are physiologically elevated. 10 Color Doppler sonography is also not effective in distinguishing decidualization of endometriosis from malignancy because flow can be present in the decidualized tissue. 9 In a study by lcázar et al, 11 neither the use of color Doppler sonography nor that of pulsed Doppler sonography improved the diagnostic accuracy of transvaginal sonography alone in the diagnosis of endometrioma. limitation of this study was the retrospective nature reporting only on surgically removed lesions and not including the large number of classic endometriomas that were followed sonographically but not removed. However, it is the surgical patients in whom it is most important to give an appropriate differential diagnosis before intervention. This study evaluated that subset of patients. In summary, we found that 60% of surgically proven endometriomas larger than 2 cm were appropriately diagnosed prospectively. However, as shown in this series, there were many cases of surgically removed endometriomas with atypical features. Some of these atypical endometriomas were appropriately diagnosed prospectively. However, atypical endometriomas can have a variety of sonographic characteristics typically attributed to other adnexal masses. Knowledge of this spectrum of appearances is important so that endometrioma is not excluded from the differential diagnosis if internal calcification is present, particularly in a postmenopausal woman. dnexal masses in pregnant patients with low-level internal echoes and solid vascular components arising from the wall of the cyst may represent decidualized endometriosis rather than malignancy. Recognition of the varied appearances of endometriomas should aid the interpreting physician in giving an appropriate prospective diagnosis of endometriomas. Table 3. Primary Prospective Diagnosis of dnexal Lesions Confirmed Histologically as Endometriomas Measuring Larger Than 2 cm in Greatest Dimension Primary Sonographic Diagnosis No. of Cases Endometrioma 29 Complex cyst 6 Hemorrhagic cyst 4 Neoplasm 4 Cystadenoma 3 Physiologic cyst 1 Dermoid 1 Solid ovarian mass 1 Complex solid mass* 1 enign-appearing cyst 1 Ovarian cyst, not otherwise specified 1 Hydrosalpinx 2 Calcification in otherwise normal- 1 appearing ovary Not specified 1 Normal 2 Total 58 *Serous papillary adenocarcinoma arising from an endometrioma. The differential diagnosis included pyosalpinx, dermoid, and endometrioma. Sonography was performed 4.5 years before surgery. J Ultrasound Med 2007; 26:

6 Variations in ppearance of Endometriomas Figure 3. Endometrioma with clot., Transvaginal sagittal image shows a complex 5.5-cm cystic- and solid-appearing mass. These features were considered prospectively as suggestive of neoplasm. Note that calipers are around the entire ovary and endometrioma, and 1 caliper is incorrectly placed around some surrounding fluid., Transvaginal transverse power Doppler image shows no internal blood flow. Therefore, the solid elements of this cyst are less suggestive of malignancy and likely represent blood products. Figure 4. Evolving appearance of an endometrioma., Transvaginal transverse image shows a complex cyst (calipers)., Power Doppler image shows no central blood flow. The prospective diagnosis was hemorrhagic cyst or solid ovarian tumor. It is important to recognize that if blood flow is not seen within a lesion, it cannot be assumed to be solid. Note also the through-transmission that suggests that this lesion is a cyst. C, Transvaginal transverse image shows the ovary (+ calipers) 3 months later with a more classic-appearing endometrioma ( calipers). C 998 J Ultrasound Med 2007; 26:

7 sch and Levine Figure 5. Endometrioma in pregnancy., Transvaginal transverse image in a patient 18 weeks pregnant shows a 5.8-cm endometrioma (calipers). Note the scattered bright echoes along the wall of the cyst., Transabdominal image at 33 weeks shows a slightly more coarse appearance to the endometrioma (calipers). Figure 6. Lobulated appearance of an endometrioma. Transvaginal image shows a lobulated endometrioma in a patient with a nonviable pregnancy. Note the gestational sac in the uterus. Figure 7. Endometrioma in a postmenopausal woman., Transvaginal transverse image shows an endometrioma (calipers) with punctuate calcification within the cyst., Transvaginal sagittal image in a different postmenopausal woman shows a poorly defined complex cyst. It may be that a coarse echo texture and calcification are more common in older lesions. J Ultrasound Med 2007; 26:

8 Variations in ppearance of Endometriomas Figure 8. Ruptured endometrioma. Transabdominal transverse image () and sagittal color Doppler image () in a patient with fever and an elevated white blood cell count show heterogeneous complex fluid with multiple septations. Some of the septations show blood flow. The prospective diagnosis was pelvic inflammatory disease. Note that these images were obtained 2 days after those from the same patient in Figure 3. Figure 9. Complex mass appearance of an endometrioma in a patient with prior hysterectomy., Transvaginal sagittal image shows a complex cyst (calipers) with multiple septations., Power Doppler image shows blood flow in the septations. t surgery, extensive adhesions with endometriosis were identified. The prospective diagnosis was strongly suggestive of neoplasm J Ultrasound Med 2007; 26:

9 sch and Levine Figure 10. ilateral atypical-appearing endometriomas., Transvaginal transverse image shows a complex left adnexal cyst (calipers) with an irregular mural nodule., Transvaginal transverse color Doppler image of a complex cystic- and solid-appearing lesion shows blood flow in the solid portion. It may be that the flow is predominately in some adjacent normal ovarian tissue. Given the bilateral complex appearance of these lesions, neoplasms were suspected prospectively. Figure 11. Decidualized endometriosis in a patient 7 weeks pregnant. Transvaginal sagittal color Doppler image shows a cyst with diffuse low-level internal echoes and mural irregularity. The histologic diagnosis was decidualized endometriosis. Figure 12. Papillary serous carcinoma arising in an endometriotic cyst in a postmenopausal woman. Transverse power Doppler image shows a 4-cm solid adnexal mass. J Ultrasound Med 2007; 26:

10 Variations in ppearance of Endometriomas Figure 13. Endometrioma with fluid layer. Transabdominal () and transvaginal () images show an endometrioma with layering echogenic material. It is possible if the patient is supine for a long enough period that blood products will layer, as shown in these images. Note how the hyperechoic component layers inferiorly, unlike in a dermoid, where the hyperechoic component tends to be located superiorly. References 1. Moore J, Copley S, Morris J, Lindsell D, Golding S, Kennedy S. systematic review of the accuracy of ultrasound in the diagnosis of endometriosis. Ultrasound Obstet Gynecol 2002; 20: Callen PW. Ultrasonography in Obstetrics and Gynecology. 4th ed. Philadelphia, P: W Saunders Co; slam N, Ong C, Woelfer, Nicolaides K, Jurkovic D. Serum C125 at weeks of gestation in women with morphologically normal ovaries. JOG 2000; 107: lcázar JL, Laparte C, Jurado M, López-García G. The role of transvaginal ultrasonography combined with color velocity imaging and pulsed Doppler in the diagnosis of endometrioma. Fertil Steril 1997; 67: Heaps JM, Nieberg RK, erek JS. Malignant neoplasms arising in endometriosis. Obstet Gynecol 1990; 75: Kurman RJ. laustein s Pathology of the Female Genital Tract. 4th ed. New York, NY: Springer-Verlag; Patel MD, Feldstein V, Chen DC, Lipson SD, Filly R. Endometriomas: diagnostic performance of US. Radiology 1999; 210: Dogan MM, Ugur M, Soysal SK, Soysal ME, Ekici E, Gokmen O. Transvaginal sonographic diagnosis of ovarian endometrioma. Int J Gynaecol Obstet 1996; 52: Jain K. Endometrioma with calcification simulating a dermoid on sonography. J Ultrasound Med 2006; 25: Sammour RN, Leibovitz Z, Shapiro I, et al. Decidualization of ovarian endometriosis during pregnancy mimicking malignancy. J Ultrasound Med 2005; 24: Fruscella E, Testa C, Ferrandina G, et al. Sonographic features of decidualized ovarian endometriosis suspicious for malignancy. Ultrasound Obstet Gynecol 2004; 24: J Ultrasound Med 2007; 26: