| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
EDUCATION EXHIBIT |
Department of Radiology, Mayo Clinic, Rochester, Minnesota
The morphologic effects of tamoxifen on the endometrium have greatly increased the complexity of interpreting transvaginal ultrasound (US) studies in these patients. Many of the imaging features are still incompletely understood. This dilemma is magnified by the large number of women currently undergoing tamoxifen therapy and the high likelihood that radiologists will encounter them in their daily practice.
Tamoxifen, a nonsteroidal antiestrogen, has been used effectively in the adjuvant treatment of receptor-positive breast cancer (1). Prophylactic use of this drug has also been approved for women at increased risk of the development of breast cancer (2). Unfortunately, long-term use of tamoxifen has been linked with several undesirable effects, most notably a higher prevalence of endometrial cancer among women who have used the drug (2–5). This association is enhanced by a patient history of prior hormone replacement therapy and greater body mass (6). Other proliferative lesions, including endometrial hyperplasia, polyps, and more recently, uterine sarcoma (7–9), have also been described in increased frequency in this patient population (10–14). These histopathologic changes are believed to arise from a somewhat paradoxical agonistic effect of the drug on the uterus. Although endometrial cancer and uterine sarcoma remain the most sinister complications of tamoxifen use, the majority of tamoxifen-related changes in the uterus are benign. The numbers of patients with diagnosed endometrial cancers in most studies are small. As a consequence, there is little consensus regarding the value and potential method of routine surveillance for malignancy in this population (15,16).
Postmenopausal women taking tamoxifen, whether symptomatic or not, have endometria that are frequently thicker than those of control subjects at screening with transvaginal sonography (11,12). Although this endometrial thickness seems to correlate with the duration of tamoxifen treatment (17), it unfortunately has been shown to be considerably less reliable as a predictor of underlying histopathologic change (15,18). Biopsies of the uterine cavities in these patients often reveal an atrophic endometrium (18). In part, this finding may reflect the limited role of transvaginal US in these studies: It is more a method to provide an endometrial measurement than a source for morphologic clues that may allow the radiologist to narrow the potential list of tamoxifen-related endometrial disorders and further guide diagnostic interventions (15,18,19). Most authors describe merely a thickened endometrium or one that is punctuated by cystic spaces (20,21). The reliability of this endometrial measurement obtained at transvaginal sonography is often further confounded in these women by the higher prevalence of adenomyosis (22). The heterogeneous echotexture and cysts that involve the inner myometrium in this process (23) may further obscure the interface between the endometrium and adjacent myometrium, frequently leading to overestimation of the true endometrial thickness (21). In addition, there is little consensus as to the appropriate cutoff point at which an endometrium is considered too thick and should be further investigated with a tissue-specific method of diagnosis (17,24–28). The usual 5-mm-thick endometrial stripe found to be a reliable end point in postmenopausal women (29) may be too conservative a criterion in this patient population. Symptomatic patients who present with a thickened central endometrial complex at transvaginal sonography are more likely to subsequently have a diagnosis of endometrial disease (30), but these remain in the minority. Most women undergoing tamoxifen therapy lack symptoms. In addition, some endometrial cancers among patients taking tamoxifen have been shown to remain clinically silent (10). All of these shortcomings have fostered some skepticism and reservation regarding the utility of transvaginal sonography as a method to provide surveillance of the endometrium in these patients.
As in postmenopausal women and those undergoing hormone replacement therapy, hysterosonography has the potential to resolve some of the discrepancies between information provided by transvaginal sonography and that obtained by tissue-specific methods of diagnosis in patients who undergo treatment with tamoxifen (24,25,31,32). As a complement to transvaginal sonography, fluid is introduced into the uterine cavity. This fluid may envelop and isolate focal masses such as polyps, facilitating their identification and characterization when scanned with the transvaginal probe. Processes such as adenomyosis that are extrauterine can also be more definitively localized and distinguished from intracavitary pathologic conditions (28,31). In effect, hysterosonography may complement the information and enhance the specificity provided by transvaginal sonography alone, both in postmenopausal patients and in those currently taking tamoxifen (26,31,33,34). By better defining the contents of the uterine cavity, it may obviate invasive diagnostic procedures in some cases (28,32).
In the preceding article, Fong and her colleagues have provided an excellent and comprehensive imaging review of the effects of tamoxifen on the endometrium when visualized with transvaginal US complemented by hysterosonography. Various histopathologic conditions promoted within the uterus in these patients are illustrated through a series of clear and concise cases (35). The authors have effectively shown the added specificity available with hysterosonography over transvaginal sonography in these patients, particularly in relation to the detection of polyps and overinterpretation of seemingly thickened endometria due to adenomyosis or cystic atrophy. While highlighting the potential pitfalls of using transvaginal sonography alone to decipher the endometrium, they also quote a relatively high rate of technically unsuccessful hysterosonography in their two patient groups. This finding is not an isolated experience. The percentage of technically aborted hysterosonography examinations approaches 30% in some series (32). Perhaps closer scrutiny of the actual morphology of the endometrium at initial transvaginal sonography may lend more specificity than originally anticipated in many of these patients and provide an alternative to hysterosonography in some cases.
When considering the potential uterine consequences of tamoxifen therapy, we must keep in mind that polyps outnumber and often precede other abnormalities (3,13,27,36,37). Despite often atypical histologic findings (38), most polyps are benign but not clinically insignificant. There is some evidence that polyps arising in patients taking tamoxifen may harbor a greater likelihood of evolving into cancer than polyps discovered in patients who are currently taking no supplemental hormones or in those undergoing hormone replacement therapy (3,39,40). Although polyps have been shown to be easily identified and differentiated from other endometrial or intracavitary lesions at hysterosonography (41)—a finding supported by Fong and her colleagues (35)—it may be possible to identify polyps prospectively at the initial transvaginal US study (20,42). Among patients taking tamoxifen, the high proportion of polyps punctuated internally by cysts (glandulocystic changes) (3,43) may facilitate their prospective diagnosis by means of transvaginal sonography (44). Because cysts have been described with a variety of uterine pathologic conditions in these patients, a confident diagnosis may be possible only if the cystic spaces are coupled with a hyperechoic line around the periphery of the central endometrial complex (45). This combination of findings localizes the process to the uterine cavity and should allow the radiologist to distinguish polyps from adenomyosis, which may mimic a thickened endometrium. These features, when present at the initial transvaginal US examination, may add some specificity to the diagnosis of a thickened endometrium. It may also conceivably decrease the number of unnecessary surgical diagnostic procedures prompted by adenomyosis masquerading as intracavitary disease in this group of patients. In adenomyosis, the cysts should not be framed by a hyperechoic line, since they are part of an extraluminal process. Revisiting and more thoroughly evaluating the initial transvaginal sonogram may prove an advantageous alternative to hysterosonography in certain patients taking tamoxifen. In this scheme, hysterosonography could be reserved for those patients with no distinguishing imaging characteristics at initial transvaginal sonography and in whom the procedure can be successfully performed.
Endometrial cystic atrophy is thought to account for many of the false-positive cases of endometrial disease in patients taking tamoxifen, mistakenly flagged at transvaginal US by an apparently thickened central endometrial complex (27,45). Biopsy or hysteroscopy in affected women has been shown to yield endometrial atrophy despite this finding. In cases of endometrial atrophy described as "cystic," more commonly seen in patients taking tamoxifen, the lining endometrial stroma is still compact but hypervascular, lined by flattened epithelial cells and punctuated by scattered protuberances due to dilated glands (43). Descriptions of imaging findings at hysterosonography in these patients likewise include persistent focal areas of thickened cystic protuberances within the uterine cavity (26). These imaging features were thought to simulate polyps or areas of focal endometrial thickening (26), prompting unnecessary diagnostic intervention for tissue sampling or removal. This finding may account for the slightly lower specificity of hysterosonography as a method to characterize endometrial disorders in these patients when compared with postmenopausal women not taking tamoxifen. There seems to be some discrepancy in the literature about whether these cystic changes reflect adenomyosis rather than actual intracavitary cysts due to cystic atrophy in some cases, as both disorders are more common among patients taking tamoxifen. This discrepancy could be potentially resolved if the cysts are contained within the hyperechoic line at transvaginal US, a sign predictive of a focal intracavitary process. Of related interest are the data collected by Cohen and colleagues (46), who recently compared the frequency of histologic findings between those endometrial specimens collected by means of hysteroscopy and those provided by means of hysterectomy. Their findings suggest that in postmenopausal breast cancer patients, investigation with hysteroscopy may yield a lower frequency of various positive endometrial histologic findings than specimens available from hysterectomy. This difference may have implications regarding the reportedly high number of false-positive transvaginal and hysterosonographic studies in these patients. Some cases of glandulocystic atrophy that are manifest as a thickened endometrium at transvaginal sonography and protuberances at hysterosonography may potentially reflect missed polyps related to sampling error (47,48).
A future algorithm for the evaluation of patients taking tamoxifen may potentially include a role for MR imaging. Despite there being little in the imaging literature about MR imaging in this capacity and patient population, the modality has been shown to facilitate the diagnosis of adenomyosis in patients who are not undergoing tamoxifen therapy (49). Because adenomyosis is known to be common in this group of patients, MR imaging could conceivably be useful as a means to distinguish this entity from endometrial polyps or cancer and prevent unnecessary diagnostic interventions (50). Given the recently described association of uterine sarcoma and tamoxifen use, MR imaging may also prove useful in the detection of this malignancy (51). Ascher and colleagues (52) have described two distinct imaging patterns within the uteri of patients taking tamoxifen, in addition to features known to be more specific for adenomyosis, fibroids, and other benign entities. Considerably more research is needed in this area to ultimately define a role, if present, for this modality in the imaging evaluation of these patients.
Although regular endometrial screening of patients taking tamoxifen is not currently advocated, many patients will probably undergo US of the uterus as part of a yearly gynecologic examination. A large proportion of patients may have a thickened endometrium. Although this finding can lack specificity, it might be argued that it is still possible to harvest more diagnostic information from the sonographic features of the central endometrial complex than a measurement alone. Radiologists may consider using the findings from the initial transvaginal US examination to stratify patients into three groups. Those patients with an endometrial measurement not exceeding 5 mm at transvaginal sonography would require no further studies. Patients with evidence of a focal process heralded by a hyperechoic line at transvaginal sonography, more likely related to a polyp if it is associated with cystic spaces, would undergo further evaluation with hysteroscopy. The remaining patients with evidence of isolated cystic spaces or an otherwise blandly thickened endometrium would remain candidates for hysterosonography and undergo appropriate management as directed by the results of that study. As described and illustrated by Fong et al in the preceding article, hysterosonography is superior to unenhanced transvaginal sonography in the characterization of possible endoluminal and myometrial disorders induced by tamoxifen in many patients but it might be more effectively employed selectively and as a complement to initial transvaginal sonography. Most important, radiologists should remain familiar with the variety of imaging manifestations that can occur in patients who are taking tamoxifen and recognize the importance of hysterosonography in the avoidance of misdiagnosis and overtreatment of clinically insignificant processes, while still remaining vigilant for those more menacing conditions such as endometrial cancer, sarcoma, and even polyps.
 |
References |
|---|
 Top References |
|---|
Related Article
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOGRAPHICS | RADIOLOGY | RSNA JOURNALS ONLINE |
