Use of intracavity saline instillation and transvaginal ultrasonography to detect tamoxifen-associated endometrial polyps

The effect of tamoxifen on the genital tract

Tamoxifen is a selective estrogen receptor modulator (SERM) that is widely used in the treatment of patients with breast cancer and for chemoprophylaxis in high risk women. Tamoxifen results in a spectrum of abnormalities involving the genital tract, the most significant being an increased incidence of endometrial cancer and uterine sarcoma. This article reviews the effects of tamoxifen on the genital tract and the strengths and weaknesses of various imaging modalities for evaluating the endometrium.

Cancer of the endometrium: current aspects of diagnostics and treatment

Background

Endometrial cancer represents a tumor entity with a great variation in its incidence throughout the world (range 1 to 25). This suggests enormous possibilities of cancer prevention due to the fact that the incidence is very much endocrine-related, chiefly with obesity, and thus most frequent in the developed world. As far as treatment is concerned, it is generally accepted that surgery represents the first choice of treatment. However, several recommendations seem reasonable especially with lymphadenectomy, even though they are not based on evidence. All high-risk cases are generally recommended for radiotherapy.

Methods

A literature search of the Medline was carried out for all articles on endometrial carcinoma related to diagnosis and treatment. The articles were systematically reviewed and were categorized into incidence, etiology, precancerosis, early diagnosis, classification, staging, prevention, and treatment. The article is organized into several similar subheadings.

Conclusions

In spite of the overall good prognosis during the early stages of the disease, the survival is poor in advanced stages or recurrences. Diagnostic measures are very well able to detect asymptomatic recurrences. These only seem justified if patients' chances are likely to improve, otherwise such measures increases costs as well as decrease the patients' quality of life. To date neither current nor improved concepts of endocrine treatment or chemotherapy have been able to substantially increase patients' chances of survival. Therefore, newer concepts into the use of antibodies e.g. trastuzumab in HER2-overexpressing tumors and the newer endocrine compounds will need to be investigated. Furthermore, it would seem highly desirable if future studies were to identify valid criteria for an individualized management, thereby maximizing the benefits and minimizing the risks.

Sonohysterography: a valuable tool in evaluating the female pelvis

A number of medical conditions, including abnormal uterine bleeding, endometrial cancer, monitoring tamoxifen therapy, infertility, and recurrent abortion, warrant investigation of the female genital tract. Diagnostic studies including hysterosalpingogram, ultrasound, and sonohysterography have proved useful in the investigation of these gynecologic conditions. This article discusses each of these tests with particular emphasis on sonohysterography and their current and potential contributions in both diagnostic and therapeutic applications. The utility of each as well as their comparative value to each other and existing gold standards is reviewed.

TARGET AUDIENCE

Obstetricians & Gynecologists, Family Physicians.

LEARNING OBJECTIVES

After completion of this article, the reader should be able to outline the current screening methods for uterine cavity and pelvic abnormalities, to list the advantages of sonohysterography, and to describe the clinical situations where sonohysterography can be used.

Risk factors of endometrial polyps resected from postmenopausal patients with breast carcinoma treated with tamoxifen

BACKGROUND

Endometrial polyps are the most common endometrial pathology described in association with postmenopausal tamoxifen exposure. Up to 3% of these polyps may show malignant changes. However, to the authors' knowledge no one has described any risk factor for the development of this pathology in postmenopausal patients with breast carcinoma treated with tamoxifen. OBJECTIVE. The objective of this study was to evaluate whether risk factors can be identified for the development of endometrial polyps in postmenopausal patients with breast carcinoma treated with tamoxifen.

METHODS

The authors reviewed the medical records of 54 postmenopausal patients with breast carcinoma in whom endometrial polyps were resected by hysteroscopy after at least 6 months of tamoxifen treatment (Group I). Demographic characteristics, health habits, risk factors for endometrial carcinoma, and clinical factors related to the primary breast disease were examined. The results were compared with those obtained from 210 similar patients in whom hysteroscopy did not reveal any endometrial pathology (Group II).

RESULTS

Age at menopause was significantly older, duration of breast disease was significantly longer, and body weight was significantly heavier among Group I patients compared with Group II patients (P = 0.0162, P = 0.0026, and P = 0.0364, respectively). Endometrial thickness, measured by transvaginal ultrasonography, was significantly thicker in Group I patients (16.3 +/- 7.2 mm) compared with that detected in Group II patients (11.8 +/- 6.3; P = 0.0001).

CONCLUSIONS

Various factors, such as older age at menopause, longer duration of breast disease, heavier weight, and thicker endometrium may contribute to the prediction of increased risk of development of endometrial polyps in postmenopausal patients with breast carcinoma treated with tamoxifen.

Psychological impact of endometrial monitoring in tamoxifen-treated postmenopausal breast cancer patients

In view of the higher incidence of endometrial pathology in tamoxifen-treated breast cancer patients, it has been recommended that endometrial surveillance be performed on these women by means of transvaginal sonography. Our study investigated how breast cancer patients experience the endometrial surveillance and which personality factors influence this experience. We also studied compliance with the recommended examination. Fifty-three consecutive asymptomatic postmenopausal breast cancer patients who had taken tamoxifen for at least 6 months were included. Our results show that 23% of the women felt very anxious just before the examination. One woman in five evaluated the procedure as annoying, unpleasant, invasive and awkward, but only 3% found it really unacceptable. Difficulties in coping with mastectomy as well as anxiety negatively affected the experience of the examination. One in six women were doubtful about their ability to comply with such an examination in the future. Women who have difficulties in coming to terms with the mastectomy, anxious women, and women with a low tolerance towards common medical procedures are at especial risk of becoming dropouts. Suggestions are made for developing strategies that might improve these women's compliance.

Hysteroscopic evaluation of the endometrium in postmenopausal women taking tamoxifen

STUDY OBJECTIVE

To evaluate hysteroscopic endometrial changes due to tamoxifen therapy in postmenopausal women with breast cancer.

DESIGN

Retrospective study (Canadian Task Force classification II-2).

SETTING

University-affiliated hospital. Patients. Eighty-eight postmenopausal women (or with iatrogenic amenorrhea) receiving tamoxifen 20 mg/day for at least 1 year for breast cancer.

INTERVENTION

Record review of patients undergoing transvaginal sonography (TVS) and office hysteroscopy with eye-directed biopsy specimens obtained with a 5-mm, continuous-flow, operative hysteroscope.

MEASUREMENTS AND MAIN RESULTS

Patients with thickened endometrium and pathologic findings at hysteroscopy had taken tamoxifen for significantly longer times than those without such findings (p < 0.05).

CONCLUSION

Our findings confirm the estrogenic effect of tamoxifen on endometrium. Endometrial evaluation by TVS suggests further diagnostic procedures, but only hysteroscopy allows the surgeon to visualize endometrial lesions and obtain eye-directed biopsy tissue.

Tamoxifen-induced uterine abnormalities: the role of imaging

Tamoxifen citrate is an orally administered, nonsteroidal antiestrogen agent that is widely used for the treatment of breast cancer and that has recently been found to prevent breast cancer in some high-risk populations. Tamoxifen may, however, cause adverse effects at the uterine level. In this article, the authors review (a) the histopathologic uterine changes associated with tamoxifen therapy, (b) the spectrum of uterine imaging findings in women treated with tamoxifen, and (c) the recommendations of the American College of Obstetrics and Gynecology for women treated with tamoxifen. An algorithm for imaging evaluation of the uterus in women treated with tamoxifen is presented.

Malignant endometrial polyps in postmenopausal breast cancer tamoxifen-treated patients

BACKGROUND

Endometrial polyps are the most common endometrial pathology described in association with postmenopausal tamoxifen exposure. It is generally accepted that the occurrence of malignancy in endometrial polyps among healthy women is up to 0.5%. However, no one has yet described the incidence of this malignant transformation among postmenopausal breast cancer tamoxifen-treated patients. Objective. The aim of this study was to study the exact rate of malignant changes in endometrial polyps recovered from postmenopausal breast cancer tamoxifen-treated patients.

METHODS

We reviewed the pathological results and medical records of all postmenopausal breast cancer patients in whom endometrial polyps were recovered following at least 6 months of tamoxifen treatment in our institute. We also looked for the rate of malignant changes in polyps recovered from all healthy postmenopausal controls with endometrial polyps in our institute during the period of the study.

RESULTS

Two (3.0%) of 67 endometrial polyps recovered from postmenopausal breast cancer tamoxifen-treated patients revealed malignant features. None of the clinical variables tested, including risk factors for endometrial cancer, was significantly different between the groups. In the controls only 5 (0.48%) of 1034 polyps were malignant.

CONCLUSION

Up to 3.0% of endometrial polyps recovered from postmenopausal breast cancer tamoxifen-treated patients may show malignant changes. This rate is higher than that found in our controls as well as that reported in the general female population.

Discrepancy between ultrasonography and hysteroscopy and histology of endometrium in postmenopausal breast cancer patients using tamoxifen

BACKGROUND

The increased risk of endometrial carcinoma following the use of tamoxifen has stimulated studies on endometrial diagnostic screening methods. In tamoxifen users the endometrial thickening observed with transvaginal ultrasonography (TVU) frequently cannot be confirmed by hysteroscopy or histology.

OBJECTIVE

The aim was to investigate the relationship between TVU and hysteroscopic and histologic endometrial findings in postmenopausal patients using tamoxifen.

METHODS

Fifty-three asymptomatic postmenopausal tamoxifen-using breast cancer patients underwent a gynecological examination combined with TVU. Patients with an endometrial thickness of >5 mm were offered hysteroscopy and endometrial biopsy.

FINDINGS

Thirty-one patients (58%) had an endometrial thickness of >5 mm with enhanced, inhomogeneous echogenicity. Hysteroscopy was performed in 22 patients and 3 underwent hysterectomy. Seven of 22 patients had endometrial polyps, histologically characterized by cystically dilated glands lined with atrophic epithelium and periglandular stromal condensation. Histology of the three hysterectomy specimens showed a similar picture of atrophic luminal epithelium, covering dilated glands lined with atrophic epithelium and surrounded by dense stroma, which resembled the histology of the endometrial polyps. In all three specimens the histologically measured endometrial thickness corresponded with that on TVU.

INTERPRETATION

Tamoxifen can induce specific endometrial changes consisting of cystically dilated glands with periglandular stromal condensation while the overlying epithelium remains atrophic. The changes occur either in the endometrium itself or as a protrusion of the endometrium, i.e., as endometrial polyps. These findings explain the discrepancy between ultrasound, hysteroscopy, and histology. Due to the high number of false-positive findings, TVU is not an effective screening instrument in these patients.

Comparative study of transvaginal hysterosonography and biopsy for the evaluation of post-menopausal bleeding

BACKGROUND

The aim of this prospective study was to evaluate transvaginal hysterosonography (TVHS) in post-menopausal bleeding (PMB) as an alternative to endometrial biopsy. The study was conducted at the Zarka Military Hospital, Amman, Jordan, over a one-year period from 1996 to 1997.

PATIENTS AND METHODS

The study comprised 98 women presenting with post-menopausal bleeding who had been listed for diagnostic dilatation and curettage. Transvaginal sonography (TVS) and transvaginal hysterosonography were performed one week before operation. The mean age of the women was 57 years, and all of them had had their menopause for at least six months.

RESULTS

Sixty-one women (62%) demonstrated endometrial thickness of more than 5 mm by transvaginal sonography. All the women had transvaginal hysterosonography, except seven on whom hysterosonography could not be performed for technical reasons and who had to be excluded from the study, leaving a total of 54 women. TVS confirmed the presence of endoluminal mass in 30 of 54 women (60%). Twenty-two of the 30 endoluminal mass cases were pedunculated while eight were sessile. Sixteen of the pedunculated cases were endometrial polyps while the remaining six were fibroid polyps. Five of the sessile cases were fibroid, two were endometrial hyperplasia, and the last one endometrial carcinoma. The other 44 out of the 98 patients also underwent transvaginal hysterosonography. No pathology could be detected in these patients, but they were noted to have atrophic endometrium after dilatation and curettage.

CONCLUSION

The combination of transvaginal sonography and transvaginal hysterosonography is both sensitive and specific with regard to detecting and excluding endoluminal masses as the cause of post-menopausal bleeding. Diagnostic dilatation and curettage fails to detect a large percentage of some lesions, so TVS in combination with TVHS should be considered as the initial examination in the evaluation of all women with post-menopausal bleeding.

The role of ultrasound in the detection of endometrial pathologies in asymptomatic postmenopausal breast cancer patients with tamoxifen treatment

Transvaginal ultrasonography is an effective diagnostic procedure, and thus, it can reasonably be used to determine which asymptomatic, postmenopausal breast cancer patient with tamoxifen treatment will require endometrial sampling. Overall, it seems that the contribution of pulsed Doppler flow in the diagnosis of endometrial pathologies in such patients is nonconclusive. It is suggested that sonohysterography is a useful diagnostic tool for the assessment of specific endometrial pathologies, especially of space-occupying lesions in the endometrial cavity, in such patients.

A randomized trial on the use of ultrasonography or office hysteroscopy for endometrial assessment in postmenopausal patients with breast cancer who were treated with tamoxifen

OBJECTIVE

Our purpose was to compare the screening effectiveness and acceptability of transvaginal ultrasonography (with sonohysterography if endometrial thickness was >4 mm) with office hysteroscopy.

STUDY DESIGN

This randomized crossover study comprised 53 consecutive asymptomatic (without vaginal bleeding) postmenopausal patients with breast cancer who had taken tamoxifen (20 or 40 mg/day) for at least 6 months.

RESULTS

Two patients had endometrial cancer (1 primary, 1 breast secondary), both detected only by transvaginal ultrasonography. Twenty-six patients had at least 1 polyp (total 47, all benign). The women who had polyps were not significantly different in their age, body mass, months of tamoxifen intake, or the cumulative dose. The sensitivity and specificity of transvaginal ultrasonography were 85% and 100%, respectively. The corresponding values for office hysteroscopy were 77% and 92%. It was significant that more patients preferred transvaginal ultrasonography to hysteroscopy (P < .001).

CONCLUSION

Transvaginal ultrasonography (plus sonohysterography) may be more effective and acceptable than office hysteroscopy for detecting endometrial abnormalities in women taking tamoxifen.

The effects of norethisterone on endometrial abnormalities identified by transvaginal ultrasound screening of healthy post-menopausal women on tamoxifen or placebo

Tamoxifen (tam) is used extensively for treatment of patients with breast cancer and is being evaluated for chemoprevention in healthy women. It has, however, been reported to increase the risk of endometrial cancer in post-menopausal women, probably by an oestrogenic effect on the endometrium. It also causes endometrial cysts and polyps. The aims of this study were to identify the incidence of endometrial thickening, polyps and cysts by transvaginal ultrasound (TVUS) screening of a population of post-menopausal healthy women in the Royal Marsden tamoxifen chemoprevention trial and to evaluate the possible benefit from the use of intermittent norethisterone (NE) in women with persistent changes. Since 1990, we have undertaken regular TVUS, using an endovaginal B mode probe, of the 463 post-menopausal women in the trial randomized to tam (20 mg day(-1)) or placebo (plac), without breaking the randomization code. Endometrial thickening (ET) was defined as > or = 8 mm at the widest point across the myometrial cavity in the longitudinal plane, including any stromal changes. Cystic changes were defined as more than one hypoechogenic area > 1 mm. Polyps were identified using saline hydrosonography. Oral NE (2.5 mg day(-1)) was used for 21 days out of 28 for three consecutive cycles by women with persistent endometrium > or = 8 mm, including cystic and polypoid changes. TVUS was repeated after the three courses to evaluate any change caused by NE and endometrial biopsies, including hysteroscopy, was performed on those women with persistent abnormalities. A persistent ET > or = 8 mm was identified in 56 (24%) of the 235 women on tamoxifen compared with only 5 (2%) of 228 women on placebo (P <0.0005). Stromal changes, including cysts, were detected in 36 (15%) and polyps in 26 (11%) of the women on tamoxifen compared with only two (< 1%) of the women on placebo (P << 0.0005). After 3 months of cyclical norethisterone, 39 of 47 women (83%) on tamoxifen had persistent ultrasound changes. However, 45 (96%) had a progesterone withdrawal bleed. Hysteroscopy was performed in 39 women on tamoxifen (28 endometrial biopsy, 15 polypectomy), five of whom had histological evidence of a proliferative endometrium and a further three had an atypical hyperplastic endometrium (one of whom had a focus of invasive carcinoma). The cysts and polyps which were detected in women on tam could not be reversed by NE and were presumably stromal and not of malignant risk. However, 96% of the women had withdrawal NE bleeding, indicating an oestrogenically primed endometrium which could be a mechanism for an increased risk of endometrial cancer. Further studies are required to ascertain whether a progestin would protect against this risk. As in other studies, these results indicate that any increased risk of endometrial cancer caused by tamoxifen is low, and that TVUS screening is probably not justified for asymptomatic women on tamoxifen.

Tamoxifen and the female genital tract

Ugwumadu AHN, Carmichael P, Neven P. Tamoxifen and the female genital tract. Int J Gynecol Cancer 1998; 8: 6-15. Tamoxifen was originally developed by Imperial Chemical Industries (England) (ICI) in 1966 as an anti-estrogenic contraceptive. Ironically, it found a role in the treatment of anovulatory infertility, but its most important application to date is in adjuvant hormonotherapy for breast cancer. Tamoxifen has a complex and poorly understood mix of estrogenic and anti-estrogenic properties with variable and contrasting effects on hormone-sensitive target tissues, such as the endometrium. This article reviews the gynecologic lesions associated with tamoxifen therapy and discusses the merits and acceptability of endometrial surveillance tests and the role of progestogens.

Ultrasound in gynecology: endometrium

With the development and clinical application of transvaginal transducers/probes (TVS) the sonographic imaging of the endometrium was greatly enhanced compared with abdominal ultrasound. Also, the discomfort of a full bladder associated with abdominal ultrasound could be avoided. A shorter distance between probe and target allowed the use of higher frequency transducers, thereby achieving improved imaging. This review will only discuss the use of TVS, but it does not mean that abdominal ultrasound should not or cannot be used when dealing with the postmenopausal uterus. Transvaginal sonography (TVS) provides a valuable tool for the diagnosis of a wide range of gynecological disorders including those of the uterus and endometrium. The ability of TVS to depict the thickness and morphology of the endometrium has been established in both office and hospital settings. This article will discuss and illustrate the clinical and research applications of transvaginal sonography in relation to the endometrium in both symptomatic and asymptomatic postmenopausal women. The article is of particular relevance today given the number of women who are undergoing transvaginal ultrasonography in the absence of symptoms as a part of their routine check ups. There is a paucity of data relating to the management of apparent ultrasound abnormalities in such women. In symptomatic women or for women at risk of developing endometrial pathology, a technique that could reduce the number of biopsy procedures would be of value. Hysteroscopy, dilatation and curettage (D & C) as well as other endometrial sampling methods are all invasive, thus it would be of benefit if a way could be found to assess the endometrium using a relatively non-invasive approach. Such a technique would need to be relatively easy to learn and perform, as well as being well accepted by the patients. We believe that transvaginal sonography fulfills many of these requirements, the following review will attempt to put forward some of the evidence to support this view.

Tamoxifen and secondary tumours. An update

The nonsteroidal antiestrogen tamoxifen is the most widely used anticancer drug. In women with breast cancer, adjuvant therapy with tamoxifen reduces relapse and improves overall survival. In advanced breast cancer, the response rate is more than 50% in hormonal dependent disease. In women treated with adjuvant tamoxifen the incidence of new primary breast cancers is decreased. This latter observation has led to the initiation of prevention trials. In 1989 the first report from a large prospective randomised trial showed a significant increase of endometrial carcinoma among women treated with adjuvant tamoxifen. This effect may be linked to the somewhat paradoxical estrogenic properties of tamoxifen. The endometrial effects should be considered in the long term use of tamoxifen, and should also be taken into account in the evaluation of the prevention trials. Animal data indicate that tamoxifen can induce tumours in other organ systems, for example the liver, but no increase in primary liver cancer has been reported from the randomised trials. In some of these trials an increase in other gastrointestinal cancers (e.g. colon and gastric carcinoma) has been observed. The mechanism behind this may be different from that of the endometrium. In animal systems, tamoxifen has shown to induce DNA damage, with formation of DNA adducts. The risk of secondary gastrointestinal cancer needs to be further evaluated. The adverse effects of tamoxifen have led to the development of new anti-estrogenic drugs and other estrogen reducing agents (e.g. aromatase inhibitors).

Use of transvaginal sonography in the evaluation of endometrial hyperplasia and carcinoma

Published studies relating to the use of transvaginal sonography (TVS) in the evaluation of endometrial hyperplasia and carcinoma were reviewed. Approximately 80 percent of all curettage procedures performed for postmenopausal bleeding result in benign diagnoses, therefore, if a noninvasive modality such as TVS can be accurately used to determine endometrial thickness measurements below which pathology is less likely, sampling may be avoided. The largest study evaluating endometrial measurements in postmenopausal women with bleeding, known as the Nordic trial, found that for a cut-off value of < or = 4 mm, 96 percent sensitivity and 68 percent specificity was achieved. Another study evaluated endometrial echomorphology in addition to measurement and found that the combined approach improved the predictability of pathologic findings. TVS may also be used to assess the depth of myometrial invasion in patients already diagnosed histologically with endometrial carcinoma. Although MRI is considered the established tool in the presurgical evaluation of the patient with carcinoma, TVS was found to perform only slightly less accurately than MRI. The published studies regarding TVS and/or MRI are reviewed. Finally, the use of TVS in conjunction with a new modality, sonohysterography, in the evaluation of patients on tamoxifen therapy, is discussed. Although the published cut-off values for endometrial thickness measurements do not apply to this group of patients, a procedure whereby sterile saline is injected into the uterine cavity, via a thin catheter, provides additional information regarding endometrial contours. Once the procedure is performed, the supposed complex endometrial echo seen on TVS is often found to actually be located in the subendometrial myometrium and the endometrium itself is thin and regular. The role for TVS is well established in the search for endometrial hyperplasia and carcinoma, as well as evaluating the presence of myometrial invasion once the diagnosis is made.

Evaluation of patients with abnormal uterine bleeding

An accurate, efficient diagnosis of disorders responsible for abnormal uterine bleeding depends on a systematic consideration of all the possible causes. Careful history and physical and pelvic examinations provide the framework for evaluation. Many adjunctive diagnostic aids can be used to evaluate women with abnormal uterine bleeding. These tests include complete blood cell count, pregnancy test, hormone levels (estradiol, progesterone, follicle-stimulating hormone, luteinizing hormone, prolactin, testosterone, dehydroepiandrosterone sulfate), thyroid function tests, liver function tests, and coagulation profile. The need for these tests are individualized and based primarily on the patient's presentation. In women of reproductive age a complication of pregnancy should always be ruled out. Ectopic pregnancies can be life threatening. The prognosis in women with trophoblastic disease can be altered by a delay in establishing the correct diagnosis. Ultrasonographic studies, particularly transvaginal ultrasonography and hysteroscopy, have played an increasing role in the evaluation of patients with abnormal uterine bleeding over the past decade, especially for cases of intrauterine space-occupying lesions, including endometrial polyps, submucosal myomas, and retained placental fragments. Suspicion of reproductive tract malignancies is heightened in patients > 35 years old, those with a history of oligoovulation or anovulation suggestive of long-term unopposed estrogen exposure, those who are obese, and those who do not respond to first-line medical management. Diagnostic techniques available for the evaluation of these cases include endometrial biopsy, dilatation and curettage, transvaginal ultrasonography, and hysteroscopy. These procedures not only allow accurate diagnosis but may permit immediate therapeutic measures to be taken when organic causes are discovered. In summary, the key to the evaluation of abnormal uterine bleeding is a through history and physical and pelvic examinations governed by the differential diagnosis of excessive uterine bleeding and the selected use of adjunctive diagnostic tests and procedures only when absolutely necessary.

Effects of tamoxifen on uterus and ovaries of postmenopausal women in a randomised breast cancer prevention trial

Randomised, double-blind controlled trials have been started to determine whether tamoxifen can prevent or delay development of breast cancer in healthy women with a family history of the disease. We recruited a randomised cohort of 111 postmenopausal women (aged 46-71 years) from the Pilot Breast Cancer Prevention Trial at the Royal Marsden Hospital to study the effect of tamoxifen on the uterus and ovaries. The main outcome measures were obtained by transvaginal ultrasonography with colour doppler imaging and microscopic examination of endometrial biopsies removed at the time of the scan. There was no significant difference between tamoxifen (20 mg/day) and placebo groups in the age of the women, or the time of the scan (and sampling) after randomisation. Women taking tamoxifen had a significantly larger uterus and a lower impedance to blood flow in the uterine arteries. 39% of women taking tamoxifen had histological evidence of an abnormal endometrium compared with 10% in the control group. 10 patients in the tamoxifen group (16%) had atypical hyperplasia and another 5 (8%) had a polyp. Women with a histological abnormality had a significantly thicker endometrium and a decreased impedance to blood flow in the uterine arteries. There was no correlation between the presence of uterine abnormalities and the age of the women, or the concentrations of tamoxifen or desmethyl tamoxifen in the peripheral blood. These findings confirm that tamoxifen can cause potentially malignant changes in the endometrium of postmenopausal women. Transvaginal ultrasonography can be used to identify those women who should have endometrial samples removed for microscopic analysis.