A Practical Approach to the Ultrasound … Practical Approach to the Ultrasound Characterization of...

Copyright @ Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

A Practical Approach to the Ultrasound Characterizationof Adnexal Masses

Douglas L. Brown, MD

Abstract: Because pelvic ultrasound is commonly used to evaluate

adnexal masses, it is important to understand the most useful

sonographic features for predicting benign and malignant masses.

Determining whether an adnexal mass is of ovarian or extraovarian

origin is key in arriving at the most likely diagnosis. Most adnexal

masses are benign, and each of the most common benign ovarian

lesions has a typical sonographic appearance. Additionally, most

malignant ovarian neoplasms have a solid component with

detectable flow by Doppler ultrasound, allowing one to strongly

suggest the diagnosis. We will review an approach to the ultrasound

diagnosis of adnexal masses that progresses through a series of

4 questions to help lead one to the most likely diagnosis.

Key Words: ovary, adnexa, neoplasm, cancer, Doppler

(Ultrasound Quarterly 2007;23:87Y105)

LEARNING OBJECTIVES1. Describe the most predictive ultrasound features of the

5 most common benign ovarian masses.2. Identify the sonographic features most predictive of

ovarian malignancies.3. Describe a systematic ultrasound approach to determine

the most likely diagnosis for an adnexal mass.

Pelvic ultrasound is frequently the first imaging modalityused when abnormalities are suspected of the ovaries or

adnexa. It is incumbent on sonographers and physicians whoperform and interpret pelvic ultrasound to understand themost useful sonographic features. It is of little use to ourpatients or referring physicians to give an exceedingly longdifferential diagnosis or report a mass as just a Bcomplexcyst^ without further description or characterization. Mostadnexal masses are benign, and we now have enoughaccumulated ultrasound knowledge to confidently recognizemost benign adnexal masses. Many adnexal masses have sucha typical ultrasound appearance that a reasonably confident

diagnosis can be made by ultrasound.1 Using the power of thisknowledge will serve our patients well by allowing for moreconservative management when appropriate and decreasingpatient and physician anxiety. Most malignant masses canalso be recognized by ultrasound. There will be occasionalproblematic cases, but we can serve our patients best byknowing the common and predictive sonographic appear-ances of both benign and malignant adnexal masses.

In this article, we will review a practical approach to theultrasound diagnosis of adnexal masses. This approachprogresses through a series of 4 general questions in anattempt to arrive at the most likely diagnosis. Those 4questions are the following:1. Is the adnexal mass extraovarian in origin?2. If the mass is ovarian in origin, is it one of the BBig 5^?3. Does the clinical history help?4. Is it a complex cystic mass, or is it a solid mass?

There admittedly will be cases that remain problematicafter this or any approach, but they are infrequent.Occasionally, other imaging modalities such as computedtomography (CT), positron emission tomography, or mag-netic resonance (MR) scanning can be helpful. Magneticresonance, in particular, has a role in the characterization ofso-called indeterminate adnexal masses.2 However, it is myview that the designation of a mass as indeterminate isvariable and inconsistent, with the potential to designate toomany masses as indeterminate if the interpreting physician isnot actively involved in evaluating these masses. I believethat the number of seemingly Bindeterminate^ adnexalmasses will be minimal if one is aware of the sonographicfeatures to be discussed here.

IS THE ADNEXAL MASS EXTRAOVARIANIN ORIGIN?

Although most adnexal masses arise from the ovary,diagnostic errors will be made if one assumes an adnexalmass is of ovarian origin. It is important to try to make thedistinction of ovarian versus extraovarian origin during thereal-time scan. Seeing ovarian tissue with follicles extendingaround the mass will help confirm an ovarian origin in thepremenopausal patient. Ovarian origin may be more difficultto determine with large masses or in the postmenopausalpatient with a small ovary without follicles. Recognition of anipsilateral ovary separate from the mass allows for readydetermination of an extraovarian origin. The most problem-atic cases occur when no ovarian tissue is seen around themass or separate from the mass. In such cases, the origin may

CORE CURRICULUM IN SONOGRAPHYCME ARTICLE

Ultrasound Quarterly & Volume 23, Number 2, June 2007 87

Received for publication December 29, 2006; accepted March 16, 2007.Professor of Radiology, Mayo Clinic, Department of Radiology, Rochester,

MN.Dr Brown has disclosed that he has no financial interests in or relationships

with any commercial companies pertaining to this educational activity.Lippincott Continuing Medical Education Institute, Inc. has identified and

resolved all faculty conflicts of interest regarding this educational activity.Reprints: Douglas L. Brown, MD, Mayo Clinic, Department of Radiology, 200

First St, SW, Rochester, MN 55905 (e-mail: [email protected]).

Copyright * 2007 by Lippincott Williams & Wilkins


not be clear, and one may need to consider both ovarian andextraovarian causes. Occasionally, in these cases, the masshas features typical of, for instance, a dermoid or ahydrosalpinx that will allow one to suggest the likelydiagnosis even without knowing the origin of the mass forcertain.

Although one cannot always distinguish ovarian fromextraovarian origin, there are a few things that can be doneduring the scan to increase the likelihood of making thisdistinction. During transvaginal scanning, gentle pressurewith the transvaginal transducer and/or with a hand on thepatient’s abdomen may be useful. When some ovarian tissueis seen immediately adjacent to an adnexal mass, suchpressure may help to separate the two in cases of anextraovarian mass. Additionally, one should look with atransabdominal transducer, even if the bladder is empty, asthis might reveal a separate ovary superiorly or laterally in thepelvis. Another feature that may be useful is the vascularsupply of the lesion. This is most likely to be helpful withpedunculated fibroids. If one can show vessels bridgingbetween the uterus and the solid adnexal mass, then apedunculated fibroid is likely.3,4

Most extraovarian adnexal masses are benign.5 Severalextraovarian masses have characteristic features that allowone to make an accurate diagnosis. Table 1 lists someextraovarian masses that one may encounter. It is not an all-inclusive list, and the designations of common versusuncommon and typical imaging appearance may be open todebate. We will consider several of the more commonextraovarian masses listed in Table 1.

Paraovarian CystIt is not unusual to encounter a paraovarian or paratubal

cyst during a pelvic sonogram.6 The term paraovarian cystwill be used here, realizing that one cannot typicallydetermine exactly where they arise by ultrasound, and thedistinction does not seem to be clinically important. Thesetypically appear as simple cysts by ultrasound (Fig. 1) and areindistinguishable from ovarian cysts if one does not recognizethe extraovarian location.7,8 Occasionally, paraovarian cystshave internal echoes due to hemorrhage.9

Paraovarian CystadenomaParaovarian cystadenomas or cystadenofibromas

(Fig. 2) are uncommon but should be considered when anextraovarian cyst contains a mural nodule or septation.6,10,11

Malignancy is rare in such lesions, although they are oftenremoved surgically. This seems to be the same lesionhistologically that occurs with increased frequency in theepididymis of men with von Hippel-Lindau syndrome.12,13

TABLE 1. Causes of Extraovarian Masses in the Female Pelvis

Predominately cystic masses

Common Paraovarian cyst

Hydrosalpinx

Peritoneal inclusion cyst

Uncommon Paraovarian cystadenoma

Tarlov cyst

Bladder diverticulum

Lymphocele

Abscess

Varices

Predominately solid masses

Common Pedunculated fibroid

Uncommon Neural neoplasms

Lymphadenopathy

Rudimentary horn of unicornuate uterus

Variable sonographic appearance

Common Tubal pregnancy

Uncommon Fallopian tube neoplasm

Gastrointestinal tract lesions

Neoplasms

Appendiceal mucocele

Appendicitis

Meckel diverticulum

Distended iliopsoas bursa

FIGURE 1. Paraovarian cyst. Image from a transvaginalultrasound of the right adnexa reveals a simple cyst (betweenarrows). The cyst is adjacent to, but separate from, the rightovary (O).

FIGURE 2. Paraovarian cystadenoma. Image from atransvaginal ultrasound shows cystic mass with a nodular solidcomponent (arrow). The ipsilateral ovary (not shown) wasseparate from the cystic mass.

Brown Ultrasound Quarterly & Volume 23, Number 2, June 2007

88 * 2007 Lippincott Williams & Wilkins


However, it seems uncommon in women, and the strength ofits association with von Hippel-Lindau syndrome in womenis not clear.

HydrosalpinxA hydrosalpinx should be considered when one

encounters an elongated cystic mass with a tubular shape14

that is not a vessel (Fig. 3A). Incomplete septations (Fig. 3B),which, in a hydrosalpinx, likely represent the wall of thedistended tube as it folds on itself, were initially reported tobe a reliable indicator of a hydrosalpinx.15,16 A recent study,however, found that an incomplete septation, although morecommon in hydrosalpinges than in other lesions, was not areliable discriminator.17 That study found the waist sign,defined as indentations along opposite sides of the cysticmass, to be the most discriminating ultrasound feature of ahydrosalpinx (Fig. 3C).17 One may also encounter smallnodular-appearing areas along the wall of a hydrosalpinx(Fig. 3B), referred to as Bbeads-on-a-string^ which are prob-ably due to abnormal endosalpingeal folds.16

One should be careful not to mistake small tortuousveins with slow flow for a hydrosalpinx. The flow may be soslow that it is not detectable by Doppler techniques, and thus,one might assume that it is a hydrosalpinx. Increasing thegray scale gain slightly and looking for movement of low-level echoes in the lumen will help one recognize such veinswith slow flow.

Fallopian tube carcinoma is rare tumor18 and is anuncommon exception to the rule that extraovarian adnexalmasses are benign lesions. It should be considered if oneencounters a solid mass or papillary projections within anextraovarian tubular cystic mass (Fig. 4).19,20 It may rarelybe difficult to distinguish such a solid component that isworrisome for a tubal carcinoma from the projectionsdescribed as Bbeads-on-a-string^ in a chronic hydrosalpinx.In my experience, the nodular areas in a chronic hydro-salpinx seem to be smaller and multiple, whereas the solidcomponent in a fallopian tube neoplasm is larger and oftenis solitary.

Peritoneal Inclusion CystThe peritoneal inclusion cyst (PIC) has also been

referred to by other terms, including peritoneal pseudocyst

FIGURE 4. Transvaginal ultrasound with color Dopplerultrasound image (shown in grayscale) of the left adnexashows a solitary, relatively large, solid mass (M) within a dilatedfallopian tube (between arrows). Flow was detected within thesolid component.

FIGURE 3. Hydrosalpinx. A, Image from a transvaginalultrasound reveals a tubular cystic mass with a partial septation(arrow). A normal ovary (not shown) was seen separate fromthe mass. B, Image from a transvaginal ultrasound in a differentpatient. This hydrosalpinx contains a partial septation(arrowheads) and multiple small nodular areas (arrows) alongthe wall because of thickened endosalpingeal folds. If one didnot recognize that this lesion was extraovarian in location, itcould be mistaken for an ovarian carcinoma. C, Image from atransvaginal ultrasound in another patient with hydrosalpinxreveals a tubular cystic mass with the ‘‘waist sign’’ (arrows).The calipers mark the adjacent ovary.

Ultrasound Quarterly & Volume 23, Number 2, June 2007 Ultrasound Characterization of Adnexal Masses

* 2007 Lippincott Williams & Wilkins 89


and inflammatory cyst of the peritoneum. It is thought to becaused by peritoneal adhesions that trap fluid produced by afunctioning ovary. Most patients have a history of prior pelvicsurgery, endometriosis, or pelvic inflammatory disease (PID)that results in pelvic adhesions.

Several articles have described the ultrasound appear-ance of PICs.21Y25 The ovary typically appears to be inside thePIC by sonography, often at the periphery of the cyst. MostPICs have septations. In 1 study, gentle pressure with thetransvaginal transducers showed movement of the septationsin 48% of PICs.25 Such mobility of septations has beensuggested as more common in PICs than in neoplasms25;however, to my knowledge, this has not been confirmed inother studies. The fluid is typically anechoic, but there can beinternal echoes. Peritoneal inclusion cysts may have bizarreshapes25 and sometimes conform to the shape of thesurrounding structures, a feature which may be more evidenton CT or MR scanning. Thus, when one encounters a sep-tated cystic pelvic mass in a patient with a history of pelvicsurgery, endometriosis, or PID, a careful search should bemade for a normal ovary within or at the periphery of thecystic mass (Fig. 5).

Tarlov CystTarlov cysts, also referred to as perineural or root sleeve

cysts, are uncommon, but it is important to recognize thembecause they can be mistaken for gynecologic pathology andtreated inappropriately. Patients could potentially undergounnecessary or harmful surgery if these are not recognized.Most Tarlov cysts are asymptomatic and discovered inciden-tally.26 Rarely, surgery may be performed if they are felt to bethe cause of pain.26

Tarlov cysts may be unilateral or bilateral. Small Tarlovcysts are not typically seen by ultrasound, and it is usually thelarger cysts that are seen sonographically. Tarlov cysts almostalways have internal echoes,27Y29 although it is not clear whatthese echoes represent (Fig. 6). The cysts tend to appear asslightly elongated or beaded cystic masses posteriorly. Thismight present some confusion with a hydrosalpinx, butTarlov cysts do not tend to be as elongated and tubular as ahydrosalpinx, although this admittedly is a subjectiveevaluation that may be difficult. One would also probablynot expect to see an incomplete septation or Bwaist sign^ in aTarlov cyst. Computed tomography or MR can be used toconfirm the diagnosis of a Tarlov cyst and readily shows theirorigin from sacral neural foramina.

FIGURE 6. Tarlov cysts. A, Transverse transabdominalultrasound image reveals 3 adjacent cystic masses posteriorlyon the left. B, Transvaginal ultrasound image of one of theposterior left adnexal masses in the same patient revealsinternal echoes within the cystic mass (between arrows). Theleft ovary (not shown) was seen separate from all 3 lesions.B indicates bladder; U, uterus.

FIGURE 5. Peritoneal inclusion cyst. A, Transverse image froma transabdominal ultrasound of the pelvis shows cystic masswith septations (arrows). B, Image from a transvaginalultrasound in the same patient shows normal ovary (betweenarrows) at the edge of the cystic mass (labeled C), with themass extending around the ovary. U indicates uterus.




IF THE MASS IS OVARIAN IN ORIGIN, IS IT ONEOF THE BIG 5?

Once one has confirmed an ovarian origin of the mass,the most common lesions need to be considered. Most ovarianmasses encountered in clinical imaging practices are benign.Simple cysts, hemorrhagic cysts, corpora lutea, endometrio-mas, and dermoids are likely to be the most frequent ovarianmasses that one will encounter. I refer to these as the BBig 5^because they are so common and usually have a typicalultrasound appearance that allows one to make a confidentdiagnosis. A small percentage of these lesions will not havethe classic appearance. In such cases, knowledge of the morefrequent atypical findings will allow one to favor a particulardiagnosis, although with a lesser degree of confidence thanwhen the lesion has the classic appearance.

Simple CystAn ovarian follicle typically reaches a mean diameter

of 20 to 24 mm (range, 17Y27 mm) at the time ofovulation.30,31 Although follicles technically are simplecysts, I believe it is better to refer to them as Bfollicles^rather than Bcysts^ because the latter has potential to bemisinterpreted as something that may be abnormal or needfollow-up.32 Thus, I would generally consider a simple cystless than about 25 mm in mean diameter to be normal in theovary of a premenopausal woman. A simple cyst in the ovaryshould meet the same criteria as a simple cyst in any organ:anechoic, thin wall, and distal acoustic enhancement.

As many ovarian simple cysts (Fig. 7) will resolve, thecause is not often proven. Many are probably follicular cysts.A small percentage of corpus luteal cysts are simple cysts bysonography.33 Serous cystadenoma can appear as a simplecyst and is a more likely consideration with simple cysts thatare large and those in older patients. The approach to follow-up of simple ovarian cysts in premenopausal women is notstandardized. It has been suggested that simple cysts of lessthan 3 cm do not require followup.34

Simple ovarian cysts occur in 3.5% to 17% ofpostmenopausal women.35 On follow-up ultrasound, mostsuch cysts either disappear or remain stable.36Y39 Only a smallpercentage, generally less than approximately 10%, increasein size.36Y39 In many lefts, a simple cyst that remains less than5 cm in size is managed with follow-up ultrasound.34,40Y42 Ifthe cyst becomes larger than 5 cm or changes appearance,surgery is usually warranted.34,41 Of those cysts treatedsurgically, the majority are serous cystadenomas.42

This brings up the question of the frequency ofmalignancy in a simple ovarian cyst. We generally assumethat this is uncommon, although there are occasional reportsof malignancy in simple ovarian cysts.43Y45 These weretypically older studies using transabdominal transducers, orthe sonographic methods were not always clear. In a morerecent study of a large number of cases with well-definedmethods and use of transvaginal scanning, the frequency ofmalignancy in simple ovarian cysts was 0.7% in premeno-pausal women and 1.6% in postmenopausal women.40

Interestingly, all malignancies occurred in cysts of largerthan 7.5 cm in mean diameter. Pathologically, there wereusually small mural nodules that were not seen sonographi-cally. Therefore, for cysts that are well seen with transvaginalscanning (and also with transabdominal scanning whenneeded for larger cysts), the likelihood of malignancy insimple cysts of less than 7.5 cm is virtually nil. The authorssuggested that the 5-cm threshold for conservative versussurgical management remains valid.40 Thus, we shouldrecognize that, for larger cysts, there is a slight chance thatsmall nodular areas along the cyst wall can be missed,probably because of the larger surface area to be evaluated.Sonographers and sonologists should carefully scrutinize theentire inner cyst wall in these lesions.

Hemorrhagic CystHemorrhagic cysts occur commonly in the ovaries of

premenopausal women. The patient may present for medicalcare because of acute pain. A recent study found 2 particularsonographic features to be reliable.46 Hemorrhagic cystsoften have internal echoes with a pattern of fine interdigitat-ing lines46 that have been variously described as Breticular,^Bfishnet,^ or Blacelike^ and are believed to be caused by fibrinstrands that result when clot hemolyzes.34,46 This reticular orfishnet pattern (Fig. 8A) due to fibrin strands was found tohave a sensitivity of 90%, specificity of 98%, and positivelikelihood ratio (LR) of 40 for hemorrhagic cysts.46 Fibrinstrands have the potential to be confused with septations.46

However, fibrin strands tend to be very thin, rarely extend allthe way across a cyst, and arc within the cyst, as opposed totrue septations that have more substance and are hencethicker and extend to opposite walls of the cyst in a linearor slight curvilinear fashion. The other reliable ultrasoundfeature which is retracting clot (solid echoes with a con-cave margin) was found to have an even higher LR (967)and specificity (100%), but its sensitivity was only 30%(Fig. 8B). The combination of fibrin strands, no septations,and smooth walls was found to have an LR of 200, withsensitivity of 90% and specificity of 100%.46 It has beensuggested that follow-up ultrasound of classic hemorrhagic

FIGURE 7. Simple ovarian cyst. Transvaginal ultrasoundimages shows a simple cyst (labeled C) arising from the rightovary.




cysts is probably not necessary unless the patient remainssymptomatic.46,47

Occasionally, hemorrhagic cysts have other appear-ances. In these instances, one is not likely to confidently makethe diagnosis on a single sonogram but would at least considerthe possibility and suggest a follow-up sonogram. Mosthemorrhagic cysts will resolve within 6 to 8 weeks.34,47

Infrequently, the clot may have convex outer margins andsimulate a solid area of a neoplasm (Fig. 8B). Absence of flowin the seemingly solid area by Doppler ultrasound isreassuring, but follow-up ultrasound is generally needed. Aminority of hemorrhagic cysts will have heterogeneousinternal echoes and seem to be solid (Fig. 8C); this likely iscaused by acute or subacute clot.47 Doppler ultrasound can beuseful to consider this possibility and not send the patientdirectly to surgery for a solid ovarian neoplasm before ob-taining a follow-up ultrasound. Fluid-fluid levels occasionallyoccur in hemorrhagic cysts.47 In my experience, the moreechogenic fluid is located dependently in hemorrhagic cysts.

Corpus LuteumIt is important to recognize the corpus luteum (CL) as a

normal finding in premenopausal women and not mistake itfor something more sinister. One expects to see the CL in thesecretory phase of the menstrual cycle and in the first trimesterof pregnancy. It may also be seen in the subsequent follicularphase.48 The appearance of the CL has been described in a fewstudies using transvaginal sonography.33,48,49 Unfortunately,the ability to discriminate the CL from true abnormalities bysonography has not, to my knowledge, been investigated.Different schemes to describe its appearance have also beenused. Although these technically are cysts, usually complexcysts, it has been suggested that we refer to them as a CL,rather than a corpus luteal cyst, unless they are larger thanapproximately 4 to 5 cm.32 The CL is typically less thanapproximately 3 cm in diameter.49

The CL often appears as a small cystic lesion with acrenulated wall and low-level internal echoes (Fig. 9A),34,50

although it may have a variable appearance. In a study of theCL during the normal menstrual cycle, the majority had acentral cystic cavity.48 The minority had no such cysticcavity48 and appeared as a hypoechoic to isoechoic area in theovary. Central cystic areas became less frequent withincreasing time from ovulation. The wall was described asthickened and irregular but was not a specific feature thatwas evaluated.48

In another study of the CL in early pregnancy, 34%of patients had a CL that was hypoechoic compared withthe ovary (Fig. 9B), 27% had a thick-walled cyst withanechoic center, 23% had a cyst with internal echoes, and15% had a simple cyst.33 The wall is usually hypoechoic orisoechoic compared with adjacent ovarian parenchyma.51

The isoechoic or slightly hypoechoic lesions may beoverlooked.33 In my opinion, small lesions that have oneof these typical appearances in a premenopausal patientshould be considered as the normal CL. I admit, however,that the discriminatory ability of such features has not beenproven, and there will be some cases where one is not

FIGURE 8. Hemorrhagic ovarian cysts. A, Transvaginalultrasound image shows a complex cystic mass (betweenarrows) with a reticular or fishnet pattern of internal echoes.This is the classic appearance of a hemorrhagic ovarian cyst. B,Transvaginal ultrasound image shows 2 hemorrhagic cysts inthe left ovary. Seemingly solid area in one of the cysts has aconcave margin (arrow), which is suggestive of clot. The otherseemingly solid area (arrowhead) did not have concavemargins. This can also occur with clot but makes it harder todistinguish from true solid tissue of a neoplasm. Both cysticlesions resolved on follow-up ultrasound 1 month later. C,Transvaginal ultrasound image of a hemorrhagic cyst (betweenarrows) in the acute to subacute stage. As opposed to lesion inpart A, the echoes in this cyst are more heterogeneous andmore echogenic, potentially leading one to suspect a solidmass. Color Doppler ultrasound (not shown) showed no flowwithin the lesion, and it resolved on follow-up ultrasound.




reasonably certain that the lesion is a CL, and follow-upultrasound may be appropriate. The CL is typically a veryvascular structure, and one should not be surprised to seeabundant flow peripherally49 by color or power Dopplerultrasound (Fig. 9C). Corpora lutea may also incidentallybe identified on CT, typically as lesions of less than 3 cmwith thick, crenulated, enhancing walls.52

EndometriomaUltrasound is not reliable for identifying the small

deposits of endometriosis but is useful for identifyingendometriomas.53 Although the deposits of endometrialtissue may occur in many locations and are common onperitoneal surfaces, endometriomas usually occur in the

FIGURE 9. Corpus luteum. A, Transvaginal ultrasound imageshows a CL (between the calipers), appearing as a complexcystic mass with a crenulated wall and internal echoes in acentral cystic cavity. Small amount of free fluid (labelled FF) isalso present. B, Transvaginal ultrasound image in a differentpatient shows a hypoechoic appearance to the CL (marked bycalipers). This can result when the internal echoes are similar inechogenicity to the wall of the lesion. Corpora lutea with thisappearance can be overlooked because they are often similar inechogenicity to the adjacent ovary. C, Transvaginal ultrasoundimage with color Doppler ultrasound from the same patientas in part B shows abundant flow around the periphery ofthe CL (arrows). FF indicates free fluid. FIGURE 14. Color plates for images 14A, 14B and 14C.




ovary.54 Endometriomas (Fig. 10A) typically have homo-geneous low- to medium-level internal echoes.55Y58 Theappearance of hemorrhagic cysts and endometriomas over-laps58Y60 in a small percentage of cases, but endometriomastend to have more homogeneous internal echoes as opposed tothe more heterogeneous and reticular pattern in hemorrhagiccysts. This homogeneous appearance has sometimes beenreferred to as a Bground glass^ appearance.1

Although diffuse homogeneous internal echoes are verycommon and quite typical of endometriomas, other lesionswill occasionally have low-level internal echoes.58,61 Acous-tic streaming, a technique using Doppler ultrasound to showmovement or lack of movement of the internal echoes, hasbeen suggested as a potential method to help distinguishendometriomas from other lesions with internal echoes.62

Initial results seem promising because most endometriomasare reported not to have acoustic streaming, but furtherexperience is needed.

Septations with resulting multilocularity are common,occurring in 45% of endometriomas in 1 study.58 Hyperechoicwall foci were observed in 35% of endometriomas (Fig. 10B)and were found to be a useful feature for diagnosingendometriomas.58 To my knowledge, no other studies haveevaluated this feature of endometriomas. Wall nodularity(Fig. 10C) occurs in approximately 20% of endometriomas58

and is problematic because this feature typically raisesconcern for a neoplasm. Color Doppler ultrasound or MRmay be helpful in such instances,58 but their value in thissituation has not been proven. It has been speculated thatthese solid areas are caused by clot or fibrin.1 If so, onewould expect there to be no detectable flow by colorDoppler ultrasound. I have anecdotally seen a few caseswhere the nodular area was caused by endometrial tissue, inwhich case, flow may be detectable by Doppler ultrasound andthen confused for a neoplasm. At least 1 case of anendometrioma with a nodular area containing flow has beenreported.63 Infrequently, an endometrioma may have a solidappearance.1,34,61 In my experience, this solid appearance(Fig. 10B) tends to occur with the more chronic ovarianendometriomas. It is not always clear whether they are reallysolid lesions or just appear that way because of older bloodproducts and/or fibrosis. Additionally, extraovarian depositsof endometriosis may infrequently be large enough to beidentified by ultrasound, and such deposits tend to besolid.54,64

When one detects a cystic mass with diffuse, homo-geneous, low-level internal echoes and no nodular areas, anendometrioma is very likely. Multilocularity and/or hyper-echoic wall foci may further increase the likelihood of anendometrioma.

FIGURE 10. Endometrioma. A, Transvaginal ultrasoundimage of a typical endometrioma with homogeneous low- tomedium-level internal echoes. This lesion has 2 locules(labeled 1 and 2). Endometriomas may be unilocular ormultilocular. B, Transvaginal ultrasound image of 3 adjacentendometriomas in the right ovary in a different patient.Endometriomas 1 and 2 are essentially isoechoic to the ovaryand could be mistaken for solid lesions; both of these havesmall echogenic foci (arrows) along part of their periphery.Endometrioma 3 is not homogeneous, likely because ofmore recent internal hemorrhage. C, Transvaginalultrasound image of an endometrioma in another patient.Homogeneous internal echoes (labeled E) are present, butthere is also a solid appearing area (arrow) that is slightlyhyperechoic compared with the internal echoes. Suchsolid-appearing areas are probably caused by a clot or a focalarea of endometrial tissue.




Mature Cystic TeratomaMature cystic teratomas, often referred to as dermoids,

are the most common ovarian neoplasm.34,65 Most dermoidshave one of several typical sonographic findings, allowing aconfident diagnosis by ultrasound. Most dermoids can beaccurately characterized by ultrasound.66 For the rare casewhere one is not sure if it has typical sonographic features of adermoid, CT or MR can be performed. Either imagingmodality should allow easy recognition of macroscopic fatand confirm the diagnosis of a dermoid. My personal view isthat MR is generally the better alternative imaging modality.If there is no macroscopic fat, CT typically provides little inthe way of further clues to the diagnosis. Endometriomas,however, are one of the more common false positivediagnoses for dermoid, and MR will often allow a confidentdiagnosis of an endometrioma. It is also important to note thatmost dermoids do not have detectable flow within the lesionby color Doppler ultrasound. As dermoids are prone totorsion, it is important to assess for flow in some adjacentnormal ovary rather than in the dermoid itself when torsionof a dermoid is being considered.

A hyperechoic area (Fig. 11A), particularly one thatattenuates sound (Fig. 11B), is a common sonographic findingin dermoids. This feature is believed to be caused by amixture of hair and sebum.1,67 The terminology used todescribe this hyperechoic area is variable and includesdermoid plug,68 Rokitansky protuberance,68 densely echo-genic tubercle,69,70 echogenic nodule or mass,67,71 hyper-echoic solid component,72 and regional bright echoes.66 TheBtip of the iceberg^ sign (Fig. 11B) refers to such markedshadowing from the hyperechoic area that the deeper part ofthe mass is not seen.73 The hyperechoic solid area is typicallyof high echogenicity, although not as intensely echogenic ascalcification. This high degree of echogenicity is usuallyobvious after one has seen typical examples, but there may beoccasional cases where one is not sure if the area is echogenicenough to be typical of a dermoid. Magnetic resonanceshould be considered in such cases. Occasionally, the entiremass may be hyperechoic (Fig. 11C).67 The presence of distalacoustic shadowing, often originating in the hyperechoicarea, further increases the likelihood of a dermoid.66,71 Theshadowing may have the appearance of a progressive fadingof the sound beam, probably because of a mixture of hair andsebum, or can appear as a sharp well-defined shadow causedby calcification.67 Shadowing in dermoids is most frequentlydue to attenuation of sound by the mixture of hair and sebum,rather than due to calcification.73

Another typical feature of dermoids is the Bdermoidmesh^ appearance that consists of hyperechoic lines and dots(Fig. 11D).74,75 It is believed to be caused by hair withinnonfatty fluid.1,74,75 This is a useful sign, although fibrinstrands in a hemorrhagic cyst or air in an abscess occasionallymay cause confusion. Fortunately, it seems that, in mostdermoids with this mesh appearance, there is also anothertypical finding such as hyperechoic areas, shadowing, or a fat-fluid level.

A fluid-fluid level (Fig. 11E) has also been described indermoids76,77 but is an infrequent finding.66,71 As an isolatedfinding, it may not have much diagnostic value, but there are

often other sonographic features also present that are typicalof a dermoid.66,77 One might expect that the fatty fluid wouldbe more echogenic and would be the nondependent layer offluid. In fact, when there is a fluid-fluid level in a dermoid,most often, the nondependent fluid is hypoechoic, and thedependent fluid is hyperechoic.77 Although the nondependentfluid is always the fatty fluid by CT or MR, the echogenicityof the fatty fluid by ultrasound is variable.77 Pure sebum,which is liquid at body temperature, has been reported to beanechoic or hypoechoic.67 This may be the explanation indermoids where the nondependent fluid is less echogenic.Other lesions such as hemorrhagic cysts, endometriomas, andovarian neoplasms will occasionally have fluid-fluid levels,but in all such cases, the dependent fluid is more echogenic.77

Thus, as an isolated finding, a fluid-fluid level where thedependent fluid is more echogenic than the nondependentfluid is not very predictive of any particular pathology.However, the nondependent fluid will be more echogenic in aminority of dermoids with a fluid-fluid level (30% in 1study77). Thus, when the fluid-fluid level has this appearancewith a more echogenic nondependent layer, a dermoid seemsvery likely because this pattern was not seen in any otherlesions.77 If one recognizes a floating nodule at the interfacebetween the 2 fluid layers, this also seems predictive of adermoid.77

There is one other appearance that seems verypredictive of a dermoid, but it is uncommon. Differentterminology has been used to describe this appearance, but itconsists of multiple floating globules, spherules, or ballswithin a cystic mass (Figs. 11F, 11G).78Y82 Because this is anuncommon appearance, its predictive value has not beencarefully evaluated, although it has been suggested to bepathognomonic of a dermoid.78 Such lesions are typicallylarge cystic masses (usually greater than approximately10 cm and often 20 to 30 cm), and the globules tend to befairly uniform in size. They rarely have any of the othertypical sonographic features of a dermoid as discussed above.The floating balls are often described as hyperechoic ormarkedly echogenic, although, to my review, the echogeni-city is not always as hyperechoic as typical of most dermoids.The globules have been reported to contain hair, sebaceousdebris, keratin, and fibrin.78Y81,83Y87 Occasionally, MR showsno evidence of macroscopic fat in the floating globules.81

Most ovarian teratomas are benign. Malignant trans-formation is reported to occur in 1% to 3% of ovarianteratomas,88Y91 but this is based mostly on pathological series.Even these low rates seem high in my experience, and Isuspect the frequency is considerably less in clinical practicebecause small dermoids are sometimes followed clinically.Lower malignancy rates of 0.17% to 0.3% for ovarianteratomas have also been reported.92,93 The malignancy ratemay be higher in postmenopausal women and was reported at15% in 1 study.94 There is little information in the literaturedescribing sonographic features that one might use to predictmalignancy of ovarian teratomas. Branching isoechoic solidcomponents have been suggested as a feature of malig-nancy.95 The most common histology with malignanttransformation of ovarian teratomas is squamous cellcarcinoma.93 Serum squamous cell carcinoma antigen levels







have been suggested as a possible parameter to diagnosemalignancy in ovarian teratomas, but it may be of limitedvalue.96 With MR imaging, solid components with transmuralextension or contrast enhancement suggest malignancy.97,98

Malignant teratomas are more likely than benign teratomas tohave detectable blood flow within solid areas by colorDoppler ultrasound.96 However, the use of this feature inclinical practice is not clear because malignant teratomas areuncommon, and 20% of benign teratomas have some area ofdetectable flow by color Doppler ultrasound.96 It was notclear, however, where the flow was detected in these 20% ofpatients with benign teratomas. Another report found thatnone of the benign teratomas had detectable flow in theBcenter^ of tumor.99 Additionally, struma ovarii, althoughalso uncommon, may have detectable blood flow. Strumaovarii, a benign monodermal form of teratoma composedmostly of thyroid tissue, has a different pathologicalappearance from most mature cystic teratomas.100 Itsultrasound appearance is also variable,100 and struma ovariiis one of the more difficult lesions to reliably characterize byultrasound.101

DOES THE CLINICAL HISTORY HELP?If the above questions do not lead to a likely diagnosis,

there are a few instances when the clinical history can bequite useful. This question is mainly to consider thepossibility of TOAs because their sonographic appearanceis quite variable. Generally, one will already be aware of thispossibility based on the clinical presentation before doing thesonogram. History is also helpful for theca lutein cysts,although they usually have a characteristic ultrasoundappearance.

Tubo-ovarian AbscessTubo-ovarian abscess (TOA) results when a lower

genital tract infection ascends and results in salpingitis andthen ovarian inflammation to the degree that there isbreakdown of normal appearance of the adnexal structures.16

The term tubo-ovarian complex has been suggested whenthere are inflammatory changes, but the ovary remainsrecognizable by ultrasound.16 The term tubo-ovarian abscessis suggested when there is further progression of inflamma-tory change with resulting breakdown of the adnexalstructures such that the ovary is not separately recognized.16

The distinction is important because an abscess may requiredrainage, whereas a complex may not.34

Despite ultrasound being a common imaging method inpatients with a suspected TOA, there is surprisingly little inthe literature evaluating ultrasound’s overall usefulness inthis setting.102 There are several studies describing ultrasoundfor patients with PID but little actual description of thesonographic findings for a TOA.16,103Y106 The ultrasoundappearance of TOAs is variable1 (Fig. 12), but they usuallyhave a cystic component1,106,107 and may have septations,thick walls, internal echoes, and/or seemingly solid areas. Theappearance overlaps with other lesions including neoplasms34

and hemorrhagic cysts. A minority of TOAs have a tubularcomponent suggestive of a pyosalpinx.106 The presence of anassociated pyosalpinx, which appears like a hydrosalpinxbut with internal echoes in the fluid is a suggestive sono-graphic finding.104,106,107 Unless there is a pyosalpinxcomponent, there does not seem to be a predictive ultra-sound appearance for a TOA. Given this variable ultra-sound appearance, the clinical findings with signs andsymptoms of PID are helpful,106 if not necessary, to makethe diagnosis of TOA.

Theca Lutein CystsTheca lutein cysts typically appear as bilateral cystic

masses with multiple septations.107 They are believed to becaused by an ovarian response to high levels of humanchorionic gonadotropin. One would likely suspect thecalutein cysts by their typical appearance (Fig. 13), althoughhaving an appropriate clinical history is helpful in making amore confident diagnosis. Theca lutein cysts are associatedwith gestational trophoblastic disease but can occasionally beseen with a multiple pregnancy, fetal hydrops, or a normal

FIGURE 11. Mature cystic teratoma. A, Transvaginal ultrasound image of a complex cystic mass reveals that a portion of themass consists of a hyperechoic solid component (between the arrows), typical of a mature cystic teratoma (also known as adermoid). In this case, there is little, if any, distal shadowing from the hyperechoic component. B, Transvaginal ultrasoundimage shows a complex cystic mass (outlined by the calipers) in a different patient. A portion of it consists of a hyperechoicnodule (arrowheads), from which distal acoustic shadowing (between the arrows) originates. The acoustic shadowing obscuresthe deeper portion of the hyperechoic nodule, which has been referred to as the ‘‘tip of the iceberg’’ sign. C, Transvaginalultrasound image with color Doppler ultrasound (shown in grayscale) in another patient reveals a uniformly hyperechoicmass (arrows). Note that there is no flow detected within the mass. Although dermoids are prone to torsion, one should notexpect to detect flow within most dermoids by color Doppler ultrasound. Therefore, when considering ovarian torsion in apatient with a dermoid, one needs to look for flow in a portion of normal ovary adjacent to the dermoid. D, Transvaginalultrasound image in another patient reveals a complex cystic mass with the dermoid mesh appearance, which consistsof hyperechoic lines and dots (some of which are indicated by arrowheads). This mass also contains a hyperechoic solid area(arrow). E, Transvaginal ultrasound image in a different patient reveals a fluid-fluid level (arrows), with the hyperechoic fluidbeing nondependent. F, Transabdominal ultrasound image of another patient demonstrates multiple solid nodules(asterisks) nondependently within a large cystic mass. This has been referred to as floating fat balls and other names. Thesenodules are not as hyperechoic as in some of the more typical dermoids as illustrated above, but the ‘‘floating’’ nature of thenodules is suggestive of a dermoid. G, Sagittal T2-weighted image with fat saturation of the same patient as in part Fagain demonstrates the multiple ‘‘floating’’ nodules. Patient was scanned in the supine position. In this case, there was no fatevident on T1-weighted sequences with fat saturation. Figure 11F and 11G courtesy of Dr. Peter Hanson and Dr. Karl Stein,Wabasha, MN.




singleton pregnancy (known as hyperreactio luteinalis).108

This appearance is also the same as seen with ovarianhyperstimulation syndrome during assisted reproductivetechniques.108 After the high human chorionic gonadotropinstimulus for their formation is removed, it may take severalweeks for the cysts to resolve.107

IS IT A COMPLEX CYSTIC MASS, OR IS IT ASOLID MASS?

If the series of questions have not lead to a likelydiagnosis by this point, ovarian neoplasm (other than themature cystic teratoma as discussed above) is becoming morelikely. One will not always be able to make a specificdiagnosis at this point, but knowledge of the most frequent

appearance of the more common ovarian neoplasms willusually allow one to favor a particular diagnosis.

Ovarian neoplasms are often grouped into 4 generalhistological categories: surface epithelial neoplasms, germcell neoplasms, sex cordYstromal neoplasms, and metastaticneoplasms. Although there are malignant forms in the othercategories, most ovarian cancers are epithelial neoplasms.Epithelial neoplasms include serous, mucinous, endome-trioid, and clear-cell types. Serous and mucinous neoplasmshave benign and malignant forms, whereas endometrioid andclear cell neoplasms are nearly always malignant.109

Endometrioid and clear cell carcinomas are the most commoncarcinomas to arise in endometriomas.54 Transitional celltumors, sometimes referred to as Brenner tumors, areinfrequent epithelial tumors and are usually benign.109

Additionally, borderline tumors have been described for allsurface epithelial cell types but are, by far, most common inthe serous and mucinous cell types.110 Borderline tumors,sometimes called tumors of low malignant potential oratypical proliferating tumors, have less stromal invasion thanfrankly malignant tumors. Borderline tumors are malignant,but they carry a better prognosis than frankly malignanttumors.110 Of the germ cell group of neoplasms, mature cysticteratomas are by far the most common89 and have beendiscussed above. Less frequent neoplasms in this groupinclude dysgerminoma, yolk sac tumor, and embryonalcarcinoma. These other germ cell tumors are usually solidbut can have areas of cystic change.111 Fibroma is themost common neoplasm in the sex cordYstromal group;other neoplasms in this group include thecomas, granulosacell tumors, and Sertoli-Leydig cell tumors.112 Metastaticneoplasms constitute the fourth group and will be furtherdiscussed below.

We will review sonographic features that suggest anovarian neoplasm, particularly malignant neoplasms, and thenapply those to the different histological groups of ovarianneoplasms. The presence of a solid component (Figs. 14AYC),which is not the typical hyperechoic component of a dermoid,

FIGURE 12. Tubo-ovarian abscess. A, Transvaginal ultrasoundimage shows a complex cystic mass (outlined by calipers).At least 2 locules are seen, each with internal echoes. Thisultrasound appearance may suggest hemorrhagic cysts orendometriomas, but the clinical setting was suggestive of PID.B, Transvaginal ultrasound image in another patient shows acomplex cystic mass. This mass contains a mixed cystic andsolid area that may raise concern for a neoplasm. The clinicalsetting in this case, however, was suggestive of PID.

FIGURE 13. Theca lutein cysts. Transabdominal ultrasoundimage reveals a large cystic ovarian mass (outlined by calipers)with multiple locules, typical of theca lutein cysts. The otherovary (not shown) had a similar appearance.




is the most useful feature for identifying ovarian malig-nancy.72,113 Other terminology has been used to refer to thesesolid areas, including papillary vegetation, papillary projec-tions, and nodularity. Ovarian cancer is often unilocular andmay not therefore have any septations.72 If septations arepresent, thicker septations (typically considered as greaterthan 2 to 3 mm) are more typical of malignancy (Fig. 14B).Ascites, other than the normal trace amount that is frequentlyseen in premenopausal women, is also predictive of ovariancancer but tends to occur with later stage disease.72 Althoughovarian malignancy is associated with larger masses,5 thesonographic morphology of a mass is a more useful predictorof malignancy than its size.72,114 Upon reviewing our datafrom a previously published study of 28 ovarian malignan-

cies,72 12 malignant masses (43%) were less than 5 cm inmean diameter, 8 (29%) were less than 4 cm in mean diameter,and 6 (21%) were less than 3 cm in mean diameter. In one ofthe well-recognized studies of ovarian cancer screening withsonography, the authors reported volume rather than dia-meters.115 If one assumes a spherical shape (which I realize isnot likely but would probably skew the calculation to only asmall degree) and calculates the diameter corresponding withthose volumes, 8 (47%) of 17 ovarian malignancies in thatstudy would have been less than 4 cm in diameter. Althoughthere is a trend for malignant ovarian masses to be larger thanbenign ovarian masses, small ovarian lesions can still bemalignant. Thus, one should not disregard a mass just becauseit is small, and should still carefully evaluate its morphology.

FIGURE 14. Epithelial ovarian neoplasms. A, Transabdominal ultrasound image with color Doppler ultrasound shows a large complexcystic pelvic mass containing a solid nodular area (between the arrows) that has flow by color Doppler ultrasound. This was aserous adenocarcinoma of the right ovary. B, Transvaginal ultrasound image in another patient demonstrates a complex cystic mass.One might consider an endometrioma, in this case, due to the internal echoes; however, there were 2 solid areas that had flowby color Doppler ultrasoundVone of which is shown in this image (arrow). This was a serous borderline tumor of the right ovary.C, Transvaginal ultrasound image with color Doppler ultrasound from another patient reveals a complex cystic mass with asmall nodular solid area (arrow) along the wall. Color Doppler ultrasound shows only a small focus of flow at the base of the nodule.This was a papillary cystadenofibroma of the right ovary. D, Transvaginal ultrasound image in a different patient reveals acomplex cystic adnexal mass with a septation (arrowhead) and solid nodular areas (arrows). This was a borderline serous tumorof the left ovary. E, Transvaginal ultrasound image from a different patient shows a complex cystic mass (between arrows)with multiple septations. No solid areas were seen. This was a mucinous cystadenoma. F, Transvaginal ultrasound image shows alarge adnexal mass (between the arrows) that is partly cystic but has a large heterogeneous solid component and a few septations.This was a clear cell adenocarcinoma. See color plates for images 14A, 14B and 14C on page 93.




For large masses that seem to be simple cysts, careful scrutinyof all the inner wall is important because tiny solid nodules canbe missed.40

Doppler evaluation initially received lots of interest,but the pulsed Doppler ultrasound parameters of resistiveindex, pulsatility index, peak systolic velocity, and time-averaged maximum velocity do not allow any significant

improvement over morphological assessment. Thus, the useof pulsed Doppler analysis is limited.72,116 Optimal char-acterization occurs with the combination of gray scaleultrasound for morphological assessment and color Dopplerimaging.117,118 Assessing for solid areas by gray scaleultrasound and by color or power Doppler imaging to detectflow within any solid area (Figs. 14A, 14C) are the 2 mostuseful features for identifying malignancy.72,116,118Y120

Although one can often predict an ovarian neoplasm by thiscombined approach, it may be difficult to distinguish a benignfrom malignant neoplasm (Fig. 14D) because benignneoplasms may also have detectable flow in solid areas.119

The risk of malignancy increases with larger and moreirregular solid components.1 Although probably not widelydone, Doppler ultrasound assessment of vessel size andorientation may be helpful, with erratic course and changingcaliber of vessels more typical of malignant masses.121

Published studies have evaluated various methods forcategorizing the ultrasound features of ovarian masses,including gray scale ultrasound alone and in combinationwith Doppler imaging. Scoring systems, mathematical modelsof probability, and artificial neural networks have beendeveloped. Although these work fairly well, they are some-what cumbersome to use. Results of these and other studies,however, have provided us with the most important featuresto assess. Knowing those features and evaluating them in asubjective method have proven to be useful. Subjectiveassessment by an experienced examiner, using provensonographic features, is a highly reliable method to differ-entiate benign from malignant adnexal masses and performsbetter than at least some of the mathematical models.122,123

I believe this is the easiest method to incorporate intomost clinical practices, and it has been shown to be reliable.Thus, it behooves the ultrasound practitioner to be cognizantof these more reliable ultrasound features.

For complex cystic lesions that are more cystic thansolid, epithelial neoplasm should generally be the firstconsideration. Although there are occasional exceptions,serous and mucinous neoplasms tend to be predominatelycystic with a smaller solid component (Figs. 14AYE).Endometrioid and clear cell carcinomas (Fig. 14F) tend tohave a larger solid component. Mucinous cystadenomas tend

FIGURE 15. Ovarian fibroma. A, Transvaginal ultrasoundimage demonstrates a 7-cm solid mass (between the arrows) inthe right adnexa. It was difficult to know if this was of ovarianorigin because the right ovary was not seen. B, Anothertransvaginal ultrasound image from the same patient shows atiny echogenic focus (arrow) at the periphery of the solid mass.Such echogenic foci in the ovaries are not usually thought to beof clinical significance but, as in this case, can sometimes behelpful in identifying the ovary.35 C, Transvaginal ultrasoundimage in a different patient reveals an area of marked acousticshadowing originating within the right adnexa. One might notrecognize the solid mass (labeled M) accounting for theshadowing and could discount this as bowel gas. However, thiswas a consistent well-defined focal area of shadowing duringthe scan. Notice that the shadowing does not originate from abright interface, as might be expected with shadowing fromcalcification, a dermoid, or bowel gas.




to appear as cystic masses with multiple septations and nosolid areas (Fig. 14E), whereas serous cystadenomas tend tobe unilocular or bilocular cystic masses with no solidareas.1,118,124,125 However, the appearance of these benignlesions overlaps with each other and occasionally withmalignancy, making it difficult to render a confident diagnosisof serous or mucinous cystadenomas in some cases.1

When a tumor is predominately solid, one shouldconsider the sex cordYstromal group of tumors. Althoughsome of the tumors in this group, particularly if larger, canhave a cystic component; many are solid. The most common

neoplasm in this group is the fibroma. It usually appears as asolid mass that can be homogeneous or heterogeneous andmay attenuate sound (Fig. 15).126Y130 Occasionally, it canhave marked attenuation of sound (Fig. 15C).126,128,130

Fibromas are almost always benign tumors. There is aspectrum with thecomas, and both histological types canoccur in the same tumor. The thecoma component can secreteestrogens. Fibromas have a similar appearance to uterineleiomyomas. Thus, when one identifies a solid adnexal massbut does not identify an ipsilateral ovary, the primarydifferential diagnosis is ovarian fibroma versus pedunculatedleiomyoma. Granulosa cell tumors are low-grade malignan-cies and are the most common ovarian tumor to secreteestrogen.127 In my experience and that of others,127 granulosacell tumors are commonly solid, although a complex cysticappearance with solid areas or septations has beendescribed.131 Estrogen secretion from granulosa cell tumorscan cause endometrial thickening and abnormal vaginalbleeding.131 Sertoli-Leydig cell, and other related, tumors arealso part of the sex cordYstromal group. It is the mostcommon ovarian tumor to secrete androgens and can causevirilizing symptoms.127,132 Most virilizing ovarian neoplasmsare small solid tumors that can be subtle (Fig. 16).127,132

When a virilizing ovarian neoplasm is being considered, oneshould be suspicious of any subtle abnormality in the ovaryand at least obtain a follow-up ultrasound. Occasionally,these tumors can have a cystic component.127 When facedwith a seemingly solid lesion, one should also considermimics of solid ovarian masses: pedunculated uterineleiomyoma, chronic endometrioma, and subacute hemorrha-gic cyst. Brenner tumor is an uncommon benign epithelialtumor that also tends to be solid and may coexist with otherovarian neoplasms.133,134

FIGURE 16. Virilizing stromal neoplasm. A 39-year-old patientpresented with virilizing symptoms and elevated testosteronelevels. Transvaginal ultrasound image demonstrates a subtle,small, isoechoic mass (labeled M) in the left ovary. Such a masscould be easily overlooked or attributed to a corpus luteum.Surgical resection revealed an 11 � 6 � 4-mm Leydig celltumor of the left ovary.

FIGURE 17. Metastasis to the ovary. Transvaginal ultrasoundimage reveals a solid mass (labeled M) replacing the left ovary.The right ovary (not shown) appeared similar. This was causedby metastasis from breast cancer.

TABLE 2. Four Questions to Consider When Evaluatingan Adnexal Mass

1. Is the adnexal mass extraovarian in origin?

If so, consider more common lesions such as

paraovarian cyst

hydrosalpinx

Peritoneal inclusion cyst

2. If the mass is ovarian in origin, is it one of the BBig 5^?

Evaluate for suggestive features of a

simple cyst

hemorrhagic cyst

corpus luteum

endometrioma

mature cystic teratoma (dermoid)

3. Does the clinical history help?

Mainly to consider tubo-ovarian abscess or theca lutein cyst

4. Is it a complex cystic mass or is it a solid mass?

Evaluate for solid areas not typical of a dermoid and use color or powerDoppler ultrasound to assess for flow in the solid areas. Thesefindings are useful for predicting neoplasms.

If more cystic than solid, consider epithelial neoplasms first.

If mostly or entirely solid, consider sex cordYstromal neoplasms first ifunilateral and metastases if bilateral.




Metastatic neoplasms have a variable appearance.When considering all types of ovarian neoplasms, thereare few features useful in distinguishing primary versussecondary ovarian malignancy.135 Multilocularity is morecommon, however, in primary, versus secondary, neo-plasms.135 The most common primary sources of metastasisto the ovary include breast, colon, gastric, and appendicealcarcinomas.136Y140 When considering these more frequentprimary sources, the presence of bilateral solid masses(Fig. 17) is suggestive of metastases.34,141Y143 Metastasis tothe ovaries from breast cancer usually occurs in the settingof advanced stage disease.144,145 It is more variable whethermetastasis from other primary tumors have a knownprimary carcinoma before the metastasis is discovered.135

SUMMARYKnowing the most useful sonographic features for

common benign tumors, which are different from thesonographic features of most malignant tumors, will allowone to make a confident diagnosis or generate a short list ofdifferential diagnoses in most cases. Progressing throughthese series of 4 questions is 1 approach to the sonographicdiagnosis of adnexal masses that I have found to work well.Table 2 briefly summarizes this approach. There will beoccasional problematic cases, whether because of limitedscan quality, confusing sonographic appearance, or rarelesions that are not easily characterized, but this approachworks well for most adnexal masses.

REFERENCES1. Valentin L. Use of morphology to characterize and manage common

adnexal masses. Best Pract Res Clin Obstet Gynaecol.2004;18:71Y89.

2. Adusumilli S, Hussain HK, Caoili EM, et al. MRI of sonographicallyindeterminate adnexal masses. AJR Am J Roentgenol. 2006;187:732Y740.

3. Kim SH, Sim JS, Seong CK. Interface vessels on color/power DopplerUS and MRI: a clue to differentiate subserosal uterine myomas fromextrauterine tumors. J Comput Assist Tomogr. 2001;25:36Y42.

4. Madan R. The bridging vascular sign. Radiology. 2006;238:371Y372.5. Lee SI. Radiological reasoning: imaging characterization of bilateral

adnexal masses. AJR Am J Roentgenol. 2006;187(suppl 3):S460YS466.6. Savelli L, Ghi T, De Iaco P, et al. Paraovarian/paratubal cysts: comparison

of transvaginal sonographic and pathological findings to establishdiagnostic criteria. Ultrasound Obstet Gynecol. 2006;28:330Y334.

7. Barloon TJ, Brown BP, Abu-Yousef MM, et al. Paraovarian andparatubal cysts: preoperative diagnosis using transabdominal andtransvaginal sonography. J Clin Ultrasound. 1996;24:117Y122.

8. Kim JS, Woo SK, Suh SJ, et al. Sonographic diagnosis of paraovariancysts: value of detecting a separate ipsilateral ovary. AJR Am JRoentgenol. 1995;164:1441Y1444.

9. Alpern MB, Sandler MA, Madrazo BL. Sonographic features ofparovarian cysts and their complications. AJR Am J Roentgenol.1984;143:157Y160.

10. Korbin CD, Brown DL, Welch WR. Paraovarian cystadenomas andcystadenofibromas: sonographic characteristics in 14 cases. Radiology.1998;208:459Y462.

11. Ghossain MA, Braidy CG, Kanso HN, et al. Extraovariancystadenomas: ultrasound and MR findings in 7 cases. J Comput AssistTomogr. 2005;29:74Y79.

12. Korn WT, Schatzki SC, DiSciullo AJ, et al. Papillary cystadenoma ofthe broad ligament in von Hippel-Lindau disease. Am J Obstet Gynecol.1990;163:596Y598.

13. Funk KC, Heiken JP. Papillary cystadenoma of the broad ligament in a

patient with von Hippel-Lindau disease. AJR Am J Roentgenol.1989;153:527Y528.

14. Tessler FN, Perrella RR, Fleischer AC, et al. Endovaginal sonographicdiagnosis of dilated fallopian tubes. AJR Am J Roentgenol.1989;153:523Y525.

15. Guerriero S, Ajossa S, Lai MP, et al. Transvaginal ultrasonographyassociated with colour Doppler energy in the diagnosis of hydrosalpinx.Hum Reprod. 2000;15:1568Y1572.

16. Timor-Tritsch IE, Lerner JP, Monteagudo A, et al. Transvaginalsonographic markers of tubal inflammatory disease. Ultrasound ObstetGynecol. 1998;12:56Y66.

17. Patel MD, Acord DL, Young SW. Likelihood ratio of sonographicfindings in discriminating hydrosalpinx from other adnexal masses.AJR Am J Roentgenol. 2006;186:1033Y1038.

18. Pectasides D, Pectasides E, Economopoulos T. Fallopian tubecarcinoma: a review. Oncologist. 2006;11:902Y912.

19. Kurjak A, Kupesic S, Jacobs I. Preoperative diagnosis of the primaryfallopian tube carcinoma by three-dimensional static and powerDoppler sonography. Ultrasound Obstet Gynecol. 2000;15:246Y251.

20. Patlas M, Rosen B, Chapman W, et al. Sonographic diagnosis ofprimary malignant tumors of the fallopian tube. Ultrasound Q.2004;20:59Y65.

21. Guerriero S, Ajossa S, Mais V, et al. Role of transvaginal sonographyin the diagnosis of peritoneal inclusion cysts. J Ultrasound Med.2004;23:1193Y1200.

22. Hoffer FA, Kozakewich H, Colodny A, et al. Peritoneal inclusioncysts: ovarian fluid in peritoneal adhesions. Radiology. 1988;169:189Y191.

23. Kim JS, Lee HJ, Woo SK, et al. Peritoneal inclusion cysts and theirrelationship to the ovaries: evaluation with sonography. Radiology.1997;204:481Y484.

24. Kurachi H, Murakami T, Nakamura H, et al. Imaging of peritonealpseudocysts: value of MR imaging compared with sonography and CT.AJR Am J Roentgenol. 1993;161:589Y591.

25. Savelli L, de Iaco P, Ghi T, et al. Transvaginal sonographic appearanceof peritoneal pseudocysts. Ultrasound Obstet Gynecol. 2004;23:284Y288.

26. Slipman CW, Bhat AL, Bhagia SM, et al. Abdominal pain secondary toa sacral perineural cyst. Spine J. 2003;3:317Y320.

27. Forsnes EV, Segna RA. Dural ectasia presenting as an adnexal mass.Obstet Gynecol. 1996;88:712Y713.

28. Raza S, Klapholz H, Benacerraf BR. Tarlov cysts: a cause of complexbilateral adnexal masses on pelvic sonography. J Ultrasound Med.1994;13:803Y805.

29. McClure MJ, Atri M, Haider MA, et al. Perineural cysts presenting ascomplex cystic adnexal masses on transvaginal sonography. AJR Am JRoentgenol. 2001;177:1313Y1318.

30. Bakos O, Lundkvist O, Wide L, et al. Ultrasonographical and hormonaldescription of the normal ovulatory menstrual cycle. Acta ObstetGynecol Scand. 1994;73:790Y796.

31. Ritchie WGM. Sonographic evaluation of normal and inducedovulation. Radiology. 1986;161:1Y10.

32. Timor-Tritsch IE, Goldstein SR. The complexity of a Bcomplex mass^and the simplicity of a Bsimple cyst^. J Ultrasound Med. 2005;24:255Y258.

33. Durfee SM, Frates MC. Sonographic spectrum of the corpus luteum inearly pregnancy: gray-scale, color, and pulsed Doppler appearance.J Clin Ultrasound. 1999;27:55Y59.

34. Dill-Macky M, Atri M. Ovarian sonography. In: Callen P, ed.Ultrasonography in Obstetrics and Gynecology. 4th ed. Philadelphia:WB Saunders Company; 2000:857Y896.

35. Parsons A. Whither the simple cyst in postmenopausal women?(opinion). Ultrasound Obstet Gynecol. 2002;20:112Y116.

36. Conway C, Zalud I, Dilena M, et al. Simple cyst in the postmenopausalpatient: detection and management. J Ultrasound Med. 1998;17:369Y372.

37. Goldstein SR, Subramanyam B, Snyder JR, et al. The postmenopausalcystic adnexal mass: the potential role of ultrasound in conservativemanagement. Obstet Gynecol. 1989;73:8Y10.

38. Kroon E, Andolf E. Diagnosis and follow-up of simple ovarian cystsdetected by ultrasound in postmenopausal women. Obstet Gynecol.1995;85:211Y214.




39. Levine D, Gosink BB, Wolf SI, et al. Simple adnexal cysts: thenatural history in postmenopausal women. Radiology. 1992;184:653Y659.

40. Ekerhovd E, Wienerroith H, Staudach A, et al. Preoperative assessmentof unilocular adnexal cysts by transvaginal ultrasonography: acomparison between ultrasonographic morphologic imaging andhistopathologic diagnosis. Am J Obstet Gynecol. 2001;184:48Y54.

41. Oyelese Y, Kueck AS, Barter JF, et al. Asymptomatic postmenopausalsimple ovarian cyst. Obstet Gynecol Surv. 2002;57:803Y809.

42. Bailey CL, Ueland FR, Land GL, et al. The malignant potential of smallcystic ovarian tumors in women over 50 years of age. Gynecol Oncol.1998;69:3Y7.

43. Hall DA, McCarthy KA. The significance of the postmenopausalsimple adnexal cyst. J Ultrasound Med. 1986;5:503Y505.

44. Luxman D, Bergman A, Sagi J, et al. The postmenopausal adnexalmass: correlation between ultrasonic and pathologic findings. ObstetGynecol. 1991;77:726Y728.

45. Osmers RGW, Osmers M, von Maydell B, et al. Preoperative evaluationof ovarian tumors in the premenopause by transvaginosonography.Am J Obstet Gynecol. 1996;175:428Y434.

46. Patel MD, Feldstein VA, Filly RA. The likelihood ratio of sonographicfindings for the diagnosis of hemorrhagic ovarian cysts. J UltrasoundMed. 2005;24:607Y614. quiz 615.

47. Jain KA. Sonographic spectrum of hemorrhagic ovarian cysts.J Ultrasound Med. 2002;21:879Y886.

48. Baerwald AR, Adams GP, Pierson RA. Form and function of the corpusluteum during the human menstrual cycle. Ultrasound Obstet Gynecol.2005;25:498Y507.

49. Bourne TH, Hagstrom H, Hahlin M, et al. Ultrasound studies ofvascular and morphological changes in the human corpus luteum duringthe menstrual cycle. Fertil Steril. 1996;65:753Y758.

50. Swire MN, Castro-Aragon I, Levine D. Various sonographicappearances of the hemorrhagic corpus luteum cyst. Ultrasound Q.2004;20:45Y58.

51. Frates MC, Visweswaran A, Laing FC. Comparison of tubal ring andcorpus luteum echogenicities: a useful differentiating characteristic.J Ultrasound Med. 2001;20:27Y31.

52. Borders RJ, Breiman RS, Yeh BM, et al. Computed tomography ofcorpus luteal cysts. J Comput Assist Tomogr. 2004;28:340Y342.

53. Woodward PJ, Sohaey R, Mezzetti TP. Endometriosis:radiologic-pathologic correlation. Radiographics. 2001;21:193Y216.

54. Clement PB. Diseases of the peritoneum. In: Kurman RJ, ed.Blaustein’s Pathology of the Female Genital Tract. New York:Springer-Verlag; 2002:729Y789.

55. Dogan MM, Ugur M, Soysal SK, et al. Transvaginal sonographicdiagnosis of ovarian endometrioma. Int J Gynaecol Obstet.1996;52:145Y149.

56. Kupfer MC, Schwimer SR, Lebovic J. Transvaginal sonographicappearance of endometriomata: spectrum and findings. J UltrasoundMedicine. 1992;11:129Y133.

57. Mais V, Guerriero S, Ajossa S, et al. The efficiency of transvaginalultrasonography in the diagnosis of endometrioma. Fertil Steril.1993;60:776Y780.

58. Patel MD, Feldstein VA, Chen DC, et al. Endometriomas: diagnosticperformance of ultrasound. Radiology. 1999;210:739Y745.

59. Pascual MA, Tresserra F, Lopez-Martin L, et al. Role of color Dopplerultrasonography in the diagnosis of endometriotic cyst. J UltrasoundMed. 2000;19:695Y699.

60. Bhatt S, Kocakoc E, Dogra VS. Endometriosis: sonographic spectrum.Ultrasound Q. 2006;22:273Y280.

61. Kurjak A, Kupesic S. Scoring system for prediction of ovarianendometriosis based on transvaginal color and pulsed Dopplersonography. Fertil Steril. 1994;62:81Y88.

62. Clarke L, Edwards A, Pollard K. Acoustic streaming in ovarian cysts.J Ultrasound Med. 2005;24:617Y621.

63. Guerriero S, Ajossa S, Mais V, et al. The diagnosis of endometriomasusing colour Doppler energy imaging. Hum Reprod. 1998;13:1691Y1695.

64. Siegelman ES, Outwater E, Wang T, et al. Solid pelvic masses causesby endometriosis: MR imaging features. AJR Am J Roentgenol.1994;163:357Y361.

65. Koonings PP, Campbell K, Mishell DR, et al. Relative frequency of

primary ovarian neoplasms: a 10-year review. Obstet Gynecol.1989;74:921Y926.

66. Patel MD, Feldstein VA, Lipson SD, et al. Cystic teratomas of theovary: diagnostic value of sonography. AJR Am J Roentgenol.1998;171:1061Y1065.

67. Sheth S, Fishman EK, Buck JL, et al. The variable sonographicappearances of ovarian teratomas: correlation with CT. AJR Am JRoentgenol. 1988;151:331Y334.

68. Quinn SF, Erickson S, Black WC. Cystic ovarian teratomas: thesonographic appearance of the dermoid plug. Radiology.1985;155:477Y478.

69. Cohen L, Sabbagha R. Echo patterns of benign cystic teratomas bytransvaginal ultrasound. Ultrasound Obstet Gynecol.1993;3:120Y123.

70. Mais V, Guerriero S, Ajossa S, et al. Transvaginal ultrasonography inthe diagnosis of cystic teratoma. Obstet Gynecol. 1995;85:48Y52.

71. Caspi B, Appelman Z, Rabinerson D, et al. Pathognomonic echopatterns of benign cystic teratomas of the ovary: classification,incidence and accuracy rate of sonographic diagnosis. UltrasoundObstet Gynecol. 1996;7:275Y279.

72. Brown DL, Doubilet PM, Miller FH, et al. Benign and malignantovarian masses: selection of the most discriminating gray-scale andDoppler sonographic features. Radiology. 1998;208:103Y110.

73. Guttman PH Jr. In search of the elusive benign cystic ovarian teratoma:application of the ultrasound Btip of the iceberg^ sign. J ClinUltrasound. 1977;5:403Y406.

74. Malde HM, Kedar RP, Chadha D, et al. Dermoid mesh: a sonographicsign of ovarian teratoma (letter). AJR Am J Roentgenol. 1992;159:1349Y1350.

75. Bronshtein M, Yoffe N, Brandes JM, et al. Hair as a sonographic markerof ovarian teratomas: improved identification using transvaginalsonography and simulation model. J Clin Ultrasound.1991;19:351Y355.

76. Owre A, Pedersen JF. Characteristic fat-fluid level at ultrasonographyof ovarian dermoid cyst. Acta Radiol. 1991;32:317Y319.

77. Kim HC, Kim SH, Lee HJ, et al. Fluid-fluid levels in ovarian teratomas.Abdom Imaging. 2002;27:100Y105.

78. Kawamoto S, Sato K, Matsumoto H, et al. Multiple mobile spherulesin mature cystic teratoma of the ovary. AJR Am J Roentgenol.2001;176:1455Y1457.

79. Rao JR, Shah Z, Patwardhan V, et al. Ovarian cystic teratoma:determined phenotypic response of keratocytes and uncommonintracystic floating balls appearance on sonography and computedtomography. J Ultrasound Med. 2002;21:687Y691.

80. Umesaki N, Nagamatsu A, Yada C, et al. MR and ultrasound imaging offloating globules in mature ovarian cystic teratoma. Gynecol ObstetInvest. 2004;58:130Y132.

81. Canda AE, Astarcioglu H, Obuz F, et al. Cystic ovarian teratoma withintracystic floating globules. Abdom Imaging. 2005;30:369Y371.

82. Tongsong T, Wanapirak C, Khunamornpong S, et al. Numerousintracystic floating balls as a sonographic feature of benign cysticteratoma: report of 5 cases. J Ultrasound Med. 2006;25:1587Y1591.

83. Otigbah C, Thompson MO, Lowe DG, et al. Mobile globules in benigncystic teratoma of the ovary. BJOG. 2000;107:135Y138.

84. Gol M, Saygili U, Uslu T, et al. Mature cystic teratoma with intracysticfat balls in a postmenopausal woman. Eur J Obstet Gynecol ReprodBiol. 2005;119:125Y126.

85. Muramatsu Y, Moriyama N, Takayasu K, et al. CT and MR imaging ofcystic ovarian teratoma with intracystic fat balls. J Comput AssistTomogr. 1991;15:528Y529.

86. Jantarasaengaram S, Siricharoenthai S, Vairojanavong K. Cysticovarian teratoma with intracystic fat balls. Ultrasound Obstet Gynecol.2003;22:102Y103.

87. Rathod K, Kale H, Narlawar R, et al. Unusual Bfloating balls^appearance of an ovarian cystic teratoma: sonographic and CT findings.J Clin Ultrasound. 2001;29:41Y43.

88. Pantoja E, Rodriguez-Ibanez I, Axtmayer RW, et al. Complications ofdermoid tumors of the ovary. Obstet Gynecol. 1975;45:89Y94.

89. Talerman A. Germ cell tumors of the ovary. In: Kurman RJ, ed.Blaustein’s pathology of the female genital tract. 5th ed. New York:Springer-Verlag; 2002:967Y1033.

90. Caruso PA, Marsh MR, Minkowitz S, et al. An intense clinicopathologic




study of 305 teratomas of the ovary. Cancer. 1971;27:343Y348.

91. Bloomfield TH. Benign cystic teratomas of the ovary: a review ofseventy-two cases. Eur J Obstet Gynecol Reprod Biol. 1987;25:231Y237.

92. Ayhan A, Aksu T, Develioglu O, et al. Complications and bilaterality ofmature ovarian teratomas (clinicopathological evaluation of 286 cases).Aust N Z J Obstet Gynecol. 1991;31:83Y85.

93. Comerci JT, Licciardi F, Bergh PA, et al. Mature cystic teratoma: aclinicopathologic evaluation of 517 cases and review of the literature.Obstet Gynecol. 1994;84:22Y28.

94. Fenghua W, Zhanhong J, Chunhua Y. Analysis of 20 mature ovariancystic teratoma cases in postmenopausal women. Chin Med J.2001;114:137Y138.

95. Mlikotic A, McPhaul L, Hansen GC, et al. Significance of the solidcomponent in predicting malignancy in ovarian cystic teratomas.J Ultrasound Med. 2001;20:859Y866.

96. Emoto M, Obama H, Horiuchi S, et al. Transvaginal color Dopplerultrasonic characterization of benign and malignant ovarian cysticteratomas and comparison with serum squamous cell carcinomaantigen. Cancer. 2000;88:2298Y2304.

97. Rha SE, Byun JY, Jung SE, et al. Atypical CT and MRI manifestationsof mature ovarian cystic teratomas. AJR Am J Roentgenol.2004;183:743Y750.

98. Kido A, Togashi K, Konishi I, et al. Dermoid cysts of the ovary withmalignant transformation: MR appearance. AJR Am J Roentgenol.1999;172:445Y449.

99. Zalel Y, Seidman DS, Oren M, et al. Sonographic and clinicalcharacteristics of struma ovarii. J Ultrasound Med. 2000;19:857Y861.

100. Outwater E, Siegelman ES, Hunt JL. Ovarian teratomas: tumor typesand imaging characteristics. Radiographics. 2001;21:475Y490.

101. Valentin L, Ameye L, Jurkovic D, et al. Which extrauterine pelvicmasses are difficult to correctly classify as benign or malignant on thebasis of ultrasound findings and is there a way of making a correctdiagnosis? Ultrasound Obstet Gynecol. 2006;27:438Y444.

102. Horrow MM. Ultrasound of pelvic inflammatory disease. UltrasoundQ. 2004;20:171Y179.

103. Cacciatore B, Leminen A, Ingman-Friberg S, et al. Transvaginalsonographic findings in ambulatory patients with suspected pelvicinflammatory disease. Obstet Gynecol. 1992;80:912Y916.

104. Taipale P, Tarjanne H, Ylostalo P. Transvaginal sonography insuspected pelvic inflammatory disease. Ultrasound Obstet Gynecol.1995;6:430Y434.

105. Boardman LA, Peipert JF, Brody JM, et al. Endovaginal sonography forthe diagnosis of upper genital tract infection. Obstet Gynecol.1997;90:54Y57.

106. Varras M, Polyzos D, Perouli E, et al. Tubo-ovarian abscesses:spectrum of sonographic findings with surgical and pathologicalcorrelations. Clin Exp Obstet Gynecol. 2003;30:117Y121.

107. Fried AM, Kenney CM, Stigers KB, et al. Benign pelvic masses:sonographic spectrum. Radiographics. 1996;16:321Y334.

108. al-Harbi O, al-Saleem A, al-Tayeb O, et al. Recurrent bilateral thecalutein cysts in association with normal pregnancy. Ultrasound ObstetGynecol. 1998;11:222Y224.

109. Seidman JD, Russell P, Kurman RJ. Surface epithelial tumors of theovary. In: Kurman RJ, ed. Blaustein’s Pathology of the Female GenitalTract. 5th ed. New York: Springer-Verlag; 2002:791Y904.

110. Hart WR. Borderline epithelial tumors of the ovary. Mod Pathol.2005;18(suppl 2):S33YS50.

111. Brammer HM, Buck JL, Hayes WS, et al. Malignant germ cell tumorsof the ovary: radiologic-pathologic correlation. Radiographics.1990;10:715Y724.

112. Young RH, Scully RE. Sex cordYstromal, steroid cell, and otherovarian tumors with endocrine, paraendocrine, and paraneoplasticmanifestations. In: Kurman RJ, ed. Blaustein’s Pathology of the FemaleGenital Tract. 5th ed. New York: Springer-Verlag; 2002:905Y966.

113. Granberg S, Wikland M, Jansson I. Macroscopic characterization ofovarian tumors and the relation to the histological diagnosis: criteria tobe used for ultrasound evaluation. Gynecol Oncol. 1989;35:139Y144.

114. Tailor A, Jurkovic D, Bourne TH, et al. Sonographic prediction ofmalignancy in adnexal masses using multivariate logistic regressionanalysis. Ultrasound Obstet Gynecol. 1997;10:41Y47.

115. van Nagell J, DePriest P, Reedy M, et al. The efficacy of transvaginalsonographic screening in asymptomatic women at risk for ovariancancer. Gynecol Oncol. 2000;77:350Y356.

116. Guerriero S, Ajossa S, Risalvato A, et al. Diagnosis of adnexalmalignancies by using color Doppler energy imaging as asecondary test in persistent masses. Ultrasound Obstet Gynecol.1998;11:277Y282.

117. Kinkel K, Hricak H, Lu Y, et al. US characterization of ovarian masses:a meta-analysis. Radiology. 2000;217:803Y811.

118. Buy J, Ghossain MA, Hugol D, et al. Characterization of adnexalmasses: combination of color Doppler and conventional sonographycompared with spectral Doppler analysis alone and conventionalsonography alone. AJR Am J Roentgenol. 1996;166:385Y393.

119. Schelling M, Braun M, Kuhn W, et al. Combined transvaginal B-modeand color Doppler sonography for differential diagnosis of ovariantumors: results of a multivariate logistic regression analysis. GynecolOncol. 2000;77:78Y86.

120. Cohen LS, Escobar PF, Scharm C, et al. Three-dimensional powerDoppler ultrasound improves the diagnostic accuracy for ovarian cancerprediction. Gynecol Oncol. 2001;82:40Y48.

121. Kurjak A, Kupesic S, Anic T, et al. Three-dimensional ultrasound andpower Doppler improve the diagnosis of ovarian lesions. GynecolOncol. 2000;76:28Y32.

122. Valentin L, Hagen B, Tingulstad S, et al. Comparison of patternrecognition and logistic regression models for discrimination betweenbenign and malignant pelvic masses: a prospective cross validation.Ultrasound Obstet Gynecol. 2001;18:357Y365.

123. Timmerman D, Schwarzler P, Collins WP, et al. Subjective assessmentof adnexal masses with the use of ultrasonography: an analysis ofinterobserver variability and experience. Ultrasound Obstet Gynecol.1999;13:11Y16.

124. Guerriero S, Mallarini G, Ajossa S, et al. Transvaginal ultrasoundand computed tomography combined with clinical parameters andCA-125 determinations in the differential diagnosis of persistentovarian cysts in premenopausal women. Ultrasound Obstet Gynecol.1997;9:339Y343.

125. Buy JN, Ghossain MA, Sciot C, et al. Epithelial tumors of theovary: CT findings and correlation with US. Radiology.1991;178:811Y818.

126. Stephenson WM, Laing FC. Sonography of ovarian fibromas. AJR Am JRoentgenol. 1985;144:1239Y1240.

127. Outwater EK, Wagner BJ, Mannion C, et al. Sex cordYstromaland steroid cell tumors of the ovary. Radiographics. 1998;18:1523Y1546.

128. Conte M, Guariglia L, Panici PB, et al. Ovarian fibrothecoma:sonographic and histologic findings. Gynecol Obstet Invest.1991;32:51Y54.

129. Bazot M, Ghossain MA, Buy JN, et al. Fibrothecomas of the ovary: CTand US findings. J Comput Assist Tomogr. 1993;17:754Y759.

130. Atri M, Nazarnia S, Bret PM, et al. Endovaginal sonographicappearance of benign ovarian masses. Radiographics.1994;14:747Y760.

131. Sharony R, Aviram R, Fishman A, et al. Granulosa cell tumors of theovary: do they have any unique ultrasonographic and color Dopplerflow features? Int J Gynecol Cancer. 2001;11:229Y233.

132. Outwater EK, Marchetto B, Wagner BJ. Virilizing tumors of the ovary:imaging features. Ultrasound Obstet Gynecol. 2000;15:365Y371.

133. Athey PA, Siegel MF. Sonographic features of Brenner tumor of theovary. J Ultrasound Med. 1987;6:367Y372.

134. Green GE, Mortele KJ, Glickman JN, et al. Brenner tumors of theovary: sonographic and computed tomographic imaging features.J Ultrasound Med. 2006;25:1245Y1251. quiz 1252Y1254.

135. Brown DL, Zou KH, Tempany CM, et al. Primary versus secondaryovarian malignancy: imaging findings of adnexal masses in theradiology diagnostic oncology group study. Radiology. 2001;219:213Y218.

136. Moore RG, Chung M, Granai CO, et al. Incidence of metastasis to theovaries from nongenital tract primary tumors. Gynecol Oncol.2004;93:87Y91.

137. Demopoulos RI, Touger L, Dubin N. Secondary ovarian carcinoma: aclinical and pathological evaluation. Int J Gynecol Pathol.1987;6:166Y175.




138. Holtz F, Hart WR. Krukenberg tumors of the ovary: a clinicopathologicanalysis of 27 cases. Cancer. 1982;50:2438Y2447.

139. Ulbright TM, Roth LM, Stehman FB. Secondary ovarian neoplasia:a clinicopathologic study of 35 cases. Cancer. 1984;53:1164Y1174.

140. Yakushiji M, Tazaki T, Nishimura H, et al. Krukenberg tumors of theovary: a clinicopathologic analysis of 112 cases. Acta Obstet GynaecolJpn. 1987;39:479Y485.

141. Choi BI, Choo IW, Han MC, et al. Sonographic appearance ofKrukenberg tumor from gastric carcinoma. Gastrointest Radiol.1988;13:15Y18.

142. Athey PA, Butters HE. Sonographic and CT appearance of Krukenbergtumors. J Clin Ultrasound. 1984;12:205Y210.

143. Salem S, Wilson SR. Gynecologic ultrasound. In: Rumack C, WilsonSR, Charboneau JW, eds. Diagnostic Ultrasound. 3rd ed.Philadelphia: Elsevier Mosby; 2005:527Y587.

144. Hann LE, Lui DM, Shi W, et al. Adnexal masses in womenwith breast cancer: US findings with clinical and histopathologiccorrelation. Radiology. 2000;216:242Y247.

145. Simpkins F, Zahurak M, Armstrong D, et al. Ovarian malignancy inbreast cancer patients with an adnexal mass. Obstet Gynecol. 2005;105:507Y513.



Date post:	23-Apr-2018
Category:	Documents
Upload:	ngoliem
View:	218 times
Download:	0 times

A Practical Approach to the Ultrasound … Practical Approach to the Ultrasound Characterization of...

Documents