Gynecologic imaging: Current and emerging applicationsVR Iyer, SI Lee
Department of Radiology, Massachusetts General Hospital, White 270, 55 Fruit Street, Boston MA 02114, USA
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0022-3859.65285
Source of Support: None, Conflict of Interest: None
Common diagnostic challenges in gynecology and the role of imaging in their evaluation are reviewed. Etiologies of abnormal uterine bleeding identified on pelvic sonography and sonohysterography are presented. An algorithmic approach for characterizing an incidentally detected adnexal mass and use of magnetic resonance imaging for definitive diagnosis are discussed. Finally, the role of F18-fluorodeoxyglucose positron emission tomography in the management of gynecological malignancies, and pitfalls associated with their use are examined.
Keywords: Endometrial polyp, endometrioma, ovarian cyst, ovarian dermoid, pelvic ultrasound
Imaging plays a pivotal role in resolving common complaints that present to a gynecologist's practice. It has a central role in the workup of abnormal uterine bleeding. Imaging also has wide applications in diagnosing incidentally detected adnexal masses. Positron emission tomography/computed tomography (PET/CT) has several emerging applications in the management of gynecological malignancies. These key issues that radiologists must keep abreast of to reach conclusive diagnoses are discussed in this review.
Although abnormal uterine bleeding (AUB) is most commonly caused by hormonal imbalance, it can be a presenting feature of endometrial polyps, hyperplasia or cancer of the cervix or endometrium.  About 20% of postmenopausal bleeding is a manifestation of gynecologic malignancy, most commonly arising from the cervix or endometrium.  Endometrial cancer accounts for approximately 10% of AUB in postmenopausal women (and is likely to be lower in premenopausal women). However, as AUB is the most common presenting sign of endometrial cancer, which is curable in early stages, imaging is directed towards its diagnosis.
Imaging for diagnosis of AUB begins with transvaginal ultrasonography (TVUS), which is widely available, well-tolerated and noninvasive. The appearance of normal and abnormal endometrium in pre- and postmenopausal women is described below.
In a premenopausal woman, the endometrium consists of a superficial functional layer and a deep basal layer. Thickness and echogenicity of normal endometrium change during the menstrual cycle. , In the early proliferative phase (Days 5 to 9), the endometrium is relatively thin and displays an echogenicity similar to that of the myometrium. In the late proliferative phase (Days 10 to 14), the superficial portion of the endometrium is hypoechoic, whereas the deep portion of the endometrium appears as a uniformly echogenic band.  This gives the characteristic multilayered appearance of the periovulatory endometrium. Endometrial thickness and echogenicity steadily increase in the secretory phase (Days 14 to 27). By the end of the secretory phase, the entire endometrium is hyperechoic. Endometrial thickness should be measured on a sagittal image of the uterus.  The thickness of the endometrium in different menstrual phases is presented in [Table 1].  However, this normal variation in endometrial thickness compromises the utility of TVUS in the evaluation of AUB in this population. In general, a thickness of >16 mm in a symptomatic patient is considered abnormal but with sensitivity of 67% and specificity of 75%. 
The normal postmenopausal endometrium should appear thin, homogeneous and echogenic. Double-layer endometrial thickness less than 5 mm without focal thickening is consistent with atrophy.  Endometrial thickness greater than 5 mm (measured on TVUS) is highly sensitive for detecting endometrial cancer in postmenopausal women with AUB. It is more sensitive than other invasive diagnostic methods such as endometrial biopsy or sonohysterography for diagnosing endometrial cancer in this population [Table 2]. ,,,,,
The endometrium in a woman undergoing hormone replacement therapy (HRT) may vary up to 3 mm if cyclic estrogen and progesterone therapy is being used. In asymptomatic patients, a thickness of up to 8 mm is considered normal. If a patient receiving HRT presents with AUB, endometrial thickness ≥5 mm warrants further workup.  However, more recent findings suggest that a thickness of 8 mm, and not 5 mm, on TVUS may be a more appropriate cutoff point for endometrial biopsy in symptomatic patients. 
Tamoxifen is a selective estrogen receptor modulator used in the treatment of breast cancer and has a weak estrogen-agonist effect in the uterus. In asymptomatic postmenopausal women treated with tamoxifen, endometrial thickness greater than 6 mm on TVUS suggests an abnormality.  In symptomatic patients, however, a cutoff of 5 mm warrants further investigation. The thickness of the endometrium increases with duration of tamoxifen therapy. 
If the endometrium appears abnormal in thickness or morphology, the next step in the diagnosis of AUB is nonfocal endometrial biopsy. In premenopausal women with AUB unresponsive to hormonal therapy and risk factors for endometrial cancer (age >35 years, morbid obesity, chronic diabetes or hypertension, or chronic tamoxifen usage), endometrial biopsy is recommended irrespective of TVUS findings.
If TVUS shows normal endometrial thickness, or if biopsy findings are benign, a search must be initiated for focal pathology to explain the AUB. Sonohysterography and hysteroscopy are both highly accurate in detecting focal endometrial pathology, such as subendometrial fibroid and endometrial polyps [Table 2]. Sonohysterography involves placement of a 5F catheter through the cervix and distension of the uterine cavity with sterile normal saline (<20 ml) under ultrasound visualization. It is a minimally invasive procedure, well-tolerated with no sedation and not associated with major complications.
Endometrial polyps [Figure 1] typically appear as well-defined, homogeneous, polypoidal lesions that are isoechoic to the endometrium with preservation of the endometrial-myometrial interface. They are usually attached to the endometrium with a stalk containing a well-defined vascular supply. Polyps with atypical features, such as cystic components, multiplicity, broad base, and hypoechogenicity or heterogeneity, are sometimes seen. 
Submucosal fibroids appear as broad-based, hypoechoic, well-defined solid masses with shadowing. An overlying layer of echogenic endometrium confirms their subendometrial location, thus distinguishing them from endometrial polyps. In contrast to endometrial polyps, submucosal fibroids distort the endometrial-myometrial interface and show acoustic attenuation.  Pedunculated submucosal fibroids, fibroids that have prolapsed into the endocervical canal and fibroids with a multilobulated surface are unusual presentations. 
Sonohysterography is less useful for characterizing diffuse lesions like hyperplasia and endometrial carcinoma. Endometrial hyperplasia appears as a diffuse thickening of the echogenic endometrial stripe without focal abnormality. Distinguishing focal endometrial hyperplasia from polyps is difficult because of similar characteristics of focal endometrial thickening. 
Endometrial cancer should be suspected when the single layer of endometrium is thicker than 8 mm, irregular, broad-based, or the endometrial-myometrial junction is lost. Single layer thickness <2.5 mm is rarely associated with malignancy.  Chronic tamoxifen therapy is associated with higher rates of endometrial pathologies, including cancer. In addition, cystic subendometrial atrophy is also seen which results in an apparently increased thickness on TVUS evaluation. In such patients sonohysterography is useful in localizing the pathology to either the endometrium or the inner myometrium. However, cystic hypertrophy is, at times, extensive enough to be mistaken for diffuse or focal endometrial thickening, even at sonohysterography. 
Incidental adnexal masses encountered on sonography pose a challenging problem because features of benign and malignant adnexal masses overlap. Majority of malignant ovarian neoplasms have a rapid rate of growth, and complete surgical resection has been shown to improve prognosis. Thus, surgery is usually indicated if imaging features cannot adequately characterize the lesion as benign. In order to prevent unnecessary oophorectomies, while not missing early malignancy, the goal of imaging must be to confidently assess a lesion as benign and refer all indeterminate lesions expeditiously for surgical evaluation [Figure 2].
A wide range of sensitivity and specificity have been reported with US ranging from 84-97%, and 56-95%, respectively for the detection of ovarian malignancies. ,,,,, While readily available and a powerful first step, US performance is limited by operator variability and patient body habitus. Thus, when an adnexal mass is indeterminate on US, MRI has been shown to be more useful than CA-125 or CT as the follow-up test  as it can definitively characterize as benign certain lesions that are indeterminate on ultrasound.
Extra-ovarian adnexal lesions, whether solid or cystic, are likely to be benign and should be followed up with MRI, a useful adjunctive in reaching a definitive diagnosis. Solid-appearing ovarian lesions in both premenopausal and postmenopausal women and cystic ovarian lesions in postmenopausal women should be followed up with MRI, which can accurately characterize them as benign.
Most intra-ovarian cysts in premenopausal women are functional. A premenopausal woman with a cystic ovarian lesion with some features of benign lesions should be followed up with MRI, which can aid in reaching a conclusive diagnosis. Lesions with features suggestive of malignancy, e.g. irregular and thick walls and septa, papillary projections or large amount of ascites, should be referred for surgical evaluation.  However, if the lesion is a complex cyst with no definite features of either malignancy or benignity, it should be followed up with sonography in six weeks, then referred for surgical evaluation if it persists.
Common extra-ovarian lesions include pedunculated and broad-ligament fibroids, hydro- or pyosalpinx, parovarian cyst and peritoneal inclusion cysts, all of which can be well-characterized by MRI. Certain intra-ovarian lesions like dermoids, endometriomas and fibromas also have some classic MRI features.
Although most fibroids can be identified on US, small subserosal pedunculated fibroids and fibroids arising from the broad ligament can be difficult to characterize on TVUS as arising outside the ovary. The US appearance is variable, ranging from hypoechoic or echogenic with or without acoustic shadowing. On MRI, fibroids are well-rounded, discretely-marginated structures with a whirling internal architecture. Signal intensity on T1 and T2 weighted images and enhancement pattern varies widely. Continuity with the uterine myometrium establishes the diagnosis of subserosal myoma. ,, MRI can demonstrate the presence of normal ovaries, separate from the adnexal mass.
These are the most common types of ovarian sex-cord stromal tumors, and together represent 8% of ovarian neoplasms.  On US a hypoechoic and attenuating mass within the ovary is characteristic. In cases where there is no attenuation and when the ovarian origin is doubtful, MRI features are helpful. On MRI, fibromas and fibrothecomas are well-marginated ovarian lesions with homogeneous internal architecture, isointense on T1 images, and hypointense on T2 images with low-level gadolinium enhancement. 
The most common US appearance of endometrioma [Figure 3] is that of a hypoechoechoic lesion within the ovary which has diffuse low-level internal echoes.  The presence of hyperechoic foci in the wall, different from the wall nodularity seen in malignant lesions, increases the confidence in diagnosis; 4.9% of endometriomas appear solid on sonography.  MRI helps in the characterization of endometriomas when US imaging features are inconclusive or if malignant degeneration is suspected. On MRI, endometriomas appear as "light-bulb" bright lesions on T1 weighted images, due to the high concentration of blood products. A T2 "shading" effect has been described due to the varying effect of the blood products [Figure 3]. Punctate or plaque-like T1 bright lesions in the pelvis, irrespective of the T2 signal also indicate endometriosis. Enhancement of the solid portions should raise concern for malignancy. 
Mature cystic teratomas are composed of tissue of ectodermal, mesodermal and endodermal origin, with a predominance of ectodermal tissue. They are bilateral in 10% of cases and more common in women in the reproductive age group. Dermoids are usually unilocular cysts filled with sebaceous material and with a solid mural nodule consisting of teeth, hair or sebaceous material. The sonographic appearance of a dermoid is a cystic lesion with a densely echogenic tubercle [Figure 4]. It can also appear as a diffusely or partially echogenic mass with sound attenuation or as a cystic mass with thin echogenic bands.  MRI identifies macroscopic fat in dermoids as T1 bright lesions  that lose signal on fat saturation sequences. A few dermoids do not contain macroscopic fat, but lipid is present in the cyst wall and in the Rokitansky nodule. In these cases, gradient-recalled opposed-phase imaging can demonstrate loss of signal within the wall or nodule.  The solid portion of dermoids demonstrate restricted diffusion on diffusion-weighted imaging (DWI). 
The natural history of pelvic inflammatory disease is progression from a phase of acute salpingitis to involvement of the ovary in the form of a tubo-ovarian complex. Further breakdown of the adnexal architecture results in a tubo-ovarian abscess (TOA). In the acute phase, the wall of the Fallopian tube More Details becomes edematous and thickened. The fimbrial occludes, causing the tube to fill with pus and fold on itself. On US, the folding of the tube gives rise to the appearance of incomplete septa on longitudinal sections. On cross-section, the "cog-wheel" appearance is noted due to the sonolucent tube with inflammed endosalpingeal folds. Chronic disease leads to hydrosalpinx which appears sonolucent with hyperechoic projections on the wall, due to the flattened endosalpingeal folds.  On CT, an adnexal mass with uniformly thick, enhancing walls, regular, thickened septa and a thickened mesosalpinx are indicative of a TOA. Gas bubbles, when seen in the mass, are diagnostic. 
On MRI, pyosalpinx and hydrosalpinx are tortuous, elongated adnexal lesions. Wall thickening and contrast enhancement,  as well as enhancement of the surrounding inflammation are seen in pyosalpinx. The signal intensity of the mass depends on the protein content. However, they are typically hyperintense on T2 weighted images with areas of shading and hypointensity in the periphery. On DWI, TOAs, similar to most abscesses, demonstrate restricted diffusion.
Inflammation and adhesions of the pelvic peritoneum traps fluid secreted by the normal ovary, giving rise to peritoneal inclusion cysts. The US appearance is that of a multilocular cyst, with thin walls and septation, surrounding the ovary.  Fine peritoneal adhesions extend to and distort the ovary, but do not penetrate the parenchyma. When the ovary is thus entrapped, it appears like a "spider in a web".  The cysts can be mistaken for hemorrhagic cysts, hydrosalpinx, pyosalpinx, serous or mucinous cystadenoma, malignant ovarian lesions or dermoid cyst.  MRI helps in definitive diagnosis of the entity when a normal ovary cannot be identified on sonography, ipsilateral to the multiloculated cyst. The T1 hypo- and T2 hyperintense cysts have a thin pseudo-wall formed by pelvic organs, bowel and the pelvic wall, and conform to the shape of the surrounding space.
Follicular or corpus luteal cysts can be seen on US depending on the menstrual phase. On US, follicular cysts are smaller than 5 cm in size, thin-walled, sharp-margined and with posterior acoustic shadowing. They are T1 hypointense, T2 bright and with postcontrast wall enhancement. Corpus luteal cysts and hemorrhagic cysts may have walls thicker than 3 mm. The presence of blood products may result in the T1 hyperintense appearance of hemorrhagic cysts. 
Parovarian cysts are developmental remnants of the mesonephric ducts. The features of parovarian cysts are similar to follicular cysts, except that they do not regress on US follow-up. On MRI, the ovary can be identified separate from the lesion.
The applications of 2-[fluorine 18] fluoro-2-deoxy-D-glucose (FDG) positron emission tomography (PET) imaging in the management of gynecological malignancy are manifold. PET, when concurrently performed and fused with CT images (PET/CT), helps in the detection of lymph node spread of cervical cancer, to assess tumor volume and to determine local control after radiotherapy.  The utility of PET in the detection of malignancy stems from the ability to differentiate tissues with increased metabolism, such as cancer cells. Several physiologic processes with increased metabolism show abnormally high PET activity leading to misdiagnoses. Both the normal endometrium and the ovary show varying uptake depending upon the menstrual phase. Therefore, for routine clinical interpretation of PET/CT images, one must know the appearance of normal physiologic uptake in the endometrium and ovary.
In a premenopausal woman with normal menstrual cycle and no gynecological malignancy, physiological endometrial uptake has two peaks. One peak is at mid-cycle  and the other on the first three days of the menstrual flow. ,, Standardized uptake values [SUV] of 5 ± 3.2 during menses and 3.7 ± 0.9 in ovulatory phase are normal.  The endometrial uptake appears like an inverted triangle on axial images and like a curvilinear or ellipsoid over the urinary bladder in sagittal images.  Neither hormonal contraceptives nor intrauterine contraceptive devices increase the endometrial uptake significantly.  In patients with menstrual irregularities, oligomenorrhea has been shown to have high SUV similar to those seen during the normal ovulatory phase endometrium, and amenorrhea has been shown to have values similar to postmenopausal endometrium. FDG uptake should not be seen in the endometrium of postmenopausal women. The SUVs in postmenopausal women taking HRT was not seen to be higher in a study by Lerman et al.
In premenopausal women, physiological ovarian uptake of FDG PET is seen in the late follicular, ovulatory and early luteal phase of the menstrual cycle.  Most commonly increased FDG uptake is seen during the early luteal phase [Figure 5]. In women with a known malignancy, FDG accumulations in corpus luteal cysts have been misinterpreted as metastasis to iliac lymph nodes. ,, Physiological ovarian uptake is often unilateral and appears round to oval with an SUV greater than 3. Lerman et al, found that an SUV of 7.9 separated benign from malignant ovarian uptake; however, the sensitivity for the detection of malignancy is only 57%. Disappearance on imaging performed soon after the subsequent menstrual cycle most reliably confirms the physiologic nature of the cyst. In order to avoid the confusion arising from functional ovarian uptake, PET/CT scans in premenopausal women should be scheduled in the first week after menses. Any FDG accumulation in the ovaries in a postmenopausal woman should be considered suspicious for malignancy.
Heterogeneous and moderately intense FDG uptake has been seen in uterine fibroids. ,,,,,, The cause of this increased uptake may be related to the presence of receptors for certain growth factors and to the high endometrial and cervical tissue glycogen in myomatous uteri. , No correlation was seen between the uptake in the fibroid and the menstrual phase.  Kitajima et al, recently showed that the size of the fibroid, and presence of degeneration did not significantly correlate with SUV. However, a negative correlation was seen with age.
The utility of PET/CT in cervical cancers is well studied. Although MRI is more sensitive in assessing local spread of cervical cancers, PET/CT has a high sensitivity in detecting nodal metastases. The SUV max of the primary tumor can predict the presence of lymph node involvement, tumor persistence following treatment and pelvic recurrence. A higher SUV max also indicates a more aggressive tumor, and hence, the need for more aggressive treatment. , In ovarian cancer patients, PET/CT is useful to detect macroscopic recurrence of ovarian cancer  and has been shown to be more accurate than CT alone.  In advanced disease, it can be used to predict response of to neoadjuvant therapy and survival.  PET/CT has been shown to modify treatment in post-therapy surveillance of endometrial cancer in 33-73% patients in various studies. ,, Preliminary studies investigating the utility of PET/CT in uterine sarcoma,  vulvar  and vaginal cancer  and gestational trophoblastic disease , have shown promising results. The role of PET/CT in gynecological malignancy is evolving from a diagnostic tool into one that predicts survival and modifies treatment.
In postmenopausal women with AUB, endometrial thickness >5 mm on TVUS warrants further investigation.
In pre- and postmenopausal women with AUB, sonohysterography is useful in detecting endometrial polyps and submucosal fibroids.
For adnexal masses indeterminate on US, MRI has been shown to be the next best test for definitive characterization, especially if the lesion is thought to be benign.
In premenopausal women imaged with PET/CT hypermetabolic tracer uptake is seen in the normal endometrium and ovary during certain phases of the menstrual cycle.
In postmenopausal women imaged with PET/CT, hypermetabolic tracer uptake in the endometrium or the ovary should be considered suspicious for underlying pathology.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2]