- Email: [email protected]
The malignant potential of endometrial polyps
European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 206–210
The malignant potential of endometrial polyps Alon Ben-Arie*, Chen Goldchmit, Yosef Laviv, Roni Levy, Benjamin Caspi, Monica Huszar, Ram Dgani, Zion Hagay Department of Obstetrics and Gynecology, Kaplan Medical Center (affiliated to the Hadassah Medical School and the Hebrew University, Jerusalem), P.O. Box 1, Rehovot 76100, Israel Received 6 October 2003; received in revised form 22 December 2003; accepted 2 February 2004
Abstract Objectives: To determine the pre-malignant and malignant potential of endometrial polyps and to assess whether different clinical parameters are associated with malignancy in the polyps. Study design: Four hundred and thirty consecutive cases of hysteroscopic diagnosis of endometrial polyp were retrieved. The medical records, preoperative vaginal sonography results and histopathology findings were reviewed. Statistical analysis was performed. Results: Hysteroscopy truly identified endometrial polyps in 95.7% of the cases. In 11.4% cases, hyperplasia without atypia was found in the endometrial polyp. In 3.3 and 3.0% of women pre-malignant or malignant conditions were found in the polyp. Older age, menopause status and polyps larger than 1.5 cm were associated with significant pre-malignant or malignant changes, although the positive predictive value for malignancy was low. All the malignant polyps were diagnosed only in postmenopausal women. The presence of postmenopausal or irregular vaginal bleeding, was not a predictor of malignancy in the polyp. Conclusions: Postmenopausal women with endometrial polyps are at increased risk of malignancy in the polyp. Those patients, whether symptomatic or not should be evaluated by hysteroscopic resection of the polyps. Asymptomatic premenopausal patients with polyps smaller than 1.5 cm can be observed. # 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Endometrial polyp; Malignancy; Hysteroscopy
1. Introduction The endometrial polyp is a pedunculated or sessile excretion of the endometrium containing variable amounts of glands, stroma and blood vessels. Endometrial polyps are common pathological lesions of the uterine corpus and are usually found in perimenopausal women [1–6]. The prevalence of endometrial polyps in the general female population is estimated to be approximately 25% [6]. Abnormal uterine bleeding is frequently the presenting symptom but polyps are often asymptomatic and found during routine ultrasound or infertility investigations [1,5,7]. Polyps are in general benign growths [1], although malignancy confined to a polyp has also been identified [8–10]. The incidence of carcinoma confined to endometrial polyps varies between 0 and 4.8% depending on the selection of patients and the methods used in making the diagnosis. Polyps have been also described to be associated with premalignant changes of the endometrium, but their role as precursors for endometrial carcinoma has not been established [10–15].
* Corresponding author. Tel.: þ972-8-9441231; fax: þ972-8-9441776. E-mail address: [email protected] (A. Ben-Arie).
The objective of this study was to evaluate the risk of premalignant and malignant changes in endometrial polyps in symptomatic and asymptomatic women. Secondly to investigate whether clinical and ultrasonic parameters may predict malignancy in a polyp and by thus decide which polyp should be removed or observed.
2. Patients and method A retrospective study was conducted dating from January 1996 to July 2001. All consecutive cases where endometrial polyp was suspected by diagnostic hysteroscopy were included. The charts of 430 women aged 26–89 (mean 54.6 years) were reviewed. The study group included both pre- and postmenopausal women. Women were considered postmenopausal if they reported a period of at least 12 months of amenorrhea after the age of 45 years. The study group included women being treated with hormonal replacement therapy (HRT) and postmenopausal women being treated with tamoxifen as an adjuvant treatment for breast cancer (Table 1). Diagnostic hysteroscopy was performed in women with abnormal uterine bleeding, or in asymptomatic women in
0301-2115/$ – see front matter # 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejogrb.2004.02.002
A. Ben-Arie et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 206–210 Table 1 Menopause status, hormonal treatment and symptoms in the study group Clinical parameters
No. of patients (%)
Pre-menopause Menopause HRT Tamoxifen Asymptomatic Abnormal uterine bleeding
206 224 49 18 236 194
(47.9) (52.1) (11.7) (4.3) (54.9) (45.1)
HRT: hormone replacement therapy.
whom abnormal endometrial finding was suspected on routine pelvic sonography performed as a complementary mode during the annual gynecological examination. Abnormal uterine bleeding was defined as any vaginal bleeding in postmenopausal women not receiving HRT or as irregular vaginal bleeding in women still actively menstruating or being treated with HRT. All women included in the study underwent endometrial evaluation at the gynecologic ultrasound unit. Scans were performed using a real-time ultrasound scanners (Advanced Laboratories 3000, Bothell, Washington, USA) equipped with 5–9 MHz transvaginal probe and Vuluson 730 (Kretztechnik AG, Austria) equipped with 5–9 MHz transvaginal probe. The ultrasonographic diagnosis of endometrial polyp was based on the presence of the ‘‘bright edge of the polyps’’ as described elsewhere [7]. The largest polyp diameter as was measured during the ultrasound scan was recorded. Diagnostic hysteroscopy was performed, using saline infusion as a distention medium, and a Storz Endoscope (Tuttlingen, Germany), with a 5 mm diagnostic sheath. In patients, where a polyp was found, at the same session, immediate cervical dilatation, endometrial polypectomy and uterine curettage were performed. The operative procedure was done under general anesthesia or a paracervical block with 20 ml of lidocaine injection. The specimens were placed in 10% formaldehyde for histological examination. Histopathologic diagnosis distinguished between non-polypoid lesions that were mistakenly diagnosed as polyps (leiomyoma, atrophic, proliferative or secretory endometrium) and endometrial polyps, which were classified as benign, hyperplasic (simple or complex hyperplasia), premalignant (simple or complex hyperplasia with atypia), and those harboring carcinoma. The histopathalogic definitions of endometrial hyperplasia and adenocarcinoma were according to the following definitions. Endometrial simple hyperplasia was defined by the endometrial architecture that is moderately distorted with crowding of glands, cystic dilatation and noncomplex budding. The lining epithelium of the glands is pseudostratified showing mitotic activity with no atypia of cells. Endometrial complex hyperplasia was defined by crowding of branched complex glands lined by stratified mitotically active cigar shaped cells. Atypical simple hyperplasia was defined by architecture similar to simple hyperplasia, but the glands are more irregular. The
207
glands are lined by atypical cells with round hyperchromatic nuclei and granular chromatin. In atypical complex hyperplasia the architecture is similar to endometrial complex hyperplasia, but the glands are crowded and lined by atypical cells with round pleomorphic nuclei. Endometrial carcinoma was defined by crowded malignant tubular glands varying in size and invading the stroma. The glands are lined by cells showing marked atypia and mitotic activity. Ten patients were excluded from the study due to insufficient material for histologic diagnosis. The prevalence of the premalignant lesions that included endometrial hyperplasia with atypia and the risk of adenocarcinoma confined to the endometrial polyps in different subgroups of women were evaluated. The results are presented as mean S:D: Statistical analysis was performed using the Student’s t-test, chi-square test, Fisher’s-exact test and the positive and negative predictive (PPV/NPV) values were calculated where appropriate; P < 0:05 was considered significant.
3. Results The histopathalogic evaluation of 420 cases where an endometrial polyp was seen on diagnostic hysteroscopy is shown in Table 2. Hysteroscopic diagnosis of endometrial polyp compared to the histological findings, showed that hysteroscopy truly identified endometrial polyps in 95.7%
Table 2 Histological results of 420 endometrial polyps found on hysteroscopy Histology Benign Endometrial polyps Endometrial polyps with simple hyperplasia (no atypia) Endometrial polyps with complex hyperplasia (no atypia) Subtotal Pre-malignant or malignant Endometrial polyps with simple hyperplasia and atypia Endometrial polyps with complex hyperplasia and atypia Invasive endometrial carcinoma Subtotal Other Myoma Complex hyperplasia without atypia Atrophic endometrium Proliferate endometrium Secretory endometrium Subtotal Total
No.
%
327 36
77.9 8.6
12
2.9
375
89.3
11
2.6
3
0.7
13
3.0
27
6.3
9 2 1 5 1
2.1 0.5 0.2 1.2 0.2
18
4.3
420
100
208
A. Ben-Arie et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 206–210
Table 3 Benign vs. pre-malignant or malignant in subgroups of women with endometrial polyp
Menopause Abnormal bleeding Polyp > 1.5 cm HRT Tamoxifen
Benign—375 cases (%)
Pre-malignant or malignant—27 cases (%)
Statistical significance
196 176 150 47 16
22 12 21 2 2
P < 0.001 N.S P < 0.0001 N.S N.S
(49.9) (44.8) (37.3) (12.0) (4.1)
(81.5) (44.4) (77.8) (7.4) (7.4)
HRT: hormone replacement therapy.
(402/420) of the cases. In 18 cases other benign histologies were reported, mainly uterine myomas. In 11.4% cases, hyperplasia without atypia was found in the endometrial polyp. In 3.3 and 3.0% of women premalignant or malignant histologies, respectively were found to be confined to the polyp. Pre-malignant and malignant changes confined to the polyps were more common in older age (61:1 9:8 versus 54:2 11:6 years, P ¼ 0:03) and in larger polyps (3:2 1:9 versus 1:7 1:1 cm, P < 0:0001). Endometrial polyps larger than 1.5 cm were significantly associated with pre-malignant and malignant states compared to smaller polyps (Table 3). The PPV and NPV for malignancy in polyps larger than 1.5 cm were 0.1295 and 0.9733, respectively. Premenopausal women had significantly more abnormal uterine bleeding than women at menopause (56.8 versus 34.4%, P < 0:0001). On the other hand, there were significantly more women at menopause, whose endometrial polyp was found to be malignant or pre-malignant with a PPV of 0.1161 and a NPV of 0.7835. All 13 cases of malignant polyps and 9 of 14 cases with atypical hyperplasia were diagnosed in postmenopausal women. The presence of symptoms, whether in menopause or not, was not a predictor of premalignancy or malignancy (Table 3) as reflected in the PPV of 0.0689, with a NPV of 0.9299. Hormone replacement therapy and tamoxifen treatment were not found to be statistically significant factors associated with malignancy within an endometrial polyp (Table 3). Vaginal sonography had a sensitivity of 86% compared to diagnostic hysteroscopy for the identification of endometrial polyp. Of 27 pre-malignant and malignant endometrial polyps found in the study, only one polyp was not imaged by vaginal sonography (P < 0:0001).
4. Discussion In this study, one of the largest of those published concerning the significance of endometrial polyps, we have found a high rate (6.3%) of malignant and pre-malignant conditions confined to the polyps. In 3.3% of the polyps atypical hyperplasia was found, a known precursor of endometrial cancer and in 3.0% of the cases invasive endometrial adenocarcinoma was identified within the polyp.
Previous studies reported the malignancy rate in endometrial polyps to vary between 0 and 4.8% [5,9,10,12,14,15]. Orvieto et al. [12] examined 146 endometrial polyps and found 10% cases of hyperplasia (2.5% with atypia) without even one case of carcinoma. Bakour et al. [14] in a prospective cohort study of women with abnormal uterine bleeding, compared the histological findings in 62 endometrial polyps, to 186 cases where no polyp was seen. They found two (3.2%) cases of endometrial carcinoma in the polyps and 11.4% of hyperplasia (6.5% with atypia). Recently Savelli et al. [15], in the largest series to date, found that 16 of 509 polyps (3.1%) had hyperplasia with atypia, and only 4 polyps (0.8%) were cancerous. In an older study, Armenia showed that 3.5% of the women who were diagnosed with endometrial polyps, eventually had endometrial cancer in their follow-up [11]. Reasons for the discrepancies in the rate of hyperplasia and carcinoma in endometrial polyps may lie in biases originating in the small size of some of the published series and in the different diagnostic tools used. Currently, isolated endometrial polyps are diagnosed by hysteroscopy. In older studies, where uterine curettage rather than hysteroscopy was used, it was difficult to asses whether uterine pathologies were associated with an endometrial polyp, or were truly confined to a polyp. Another possible reason for the varying malignant rate found in endometrial polyps may be associated with different rates of high-risk patients for endometrial malignancy in the cohort studied. Age, race, menopausal status and HRT or tamoxifen treatments are known risk factors for endometrial cancer. Our study showed a relatively high rate of malignancy in endometrial polyps, comparable to the 3% malignancy rate found by Cohen et al. [16] in a high-risk cohort of women treated with tamoxifen. High rates of endometrial polyps are asymptomatic and are found incidentally on routine vaginal sonography [2,17,18]. Goldstein et al. [5] found that 19 of 61 polyps (31.1%) were incidental findings. In the current series, 236 (54.9%) of the resected polyps were asymptomatic. Most importantly, symptoms in our study could not differentiate benign from malignant or pre-malignant states and 15/232 (6.5%) of the asymptomatic polyps harbored atypical hyperlasia or carcinoma, an equal rate of malignancy that was found in the symptomatic polyps. Likewise, Savelli et al. [15] have also shown a high rate of polyps to be asymptomatic incidental finding with an equal rate of premalignant
A. Ben-Arie et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 206–210
or malignant conditions compared to polyps that were associated with abnormal uterine bleeding. Transvaginal sonography or saline infusion sonohysterography can be used for accurate diagnosis of endometrial polyps. The detection rate of endometrial polyps is strongly influenced by the diagnostic method used and by the experience of the sonographer [2,17,19]. In this study, vaginal sonography had a sensitivity of 86% in identification of endometrial polyps, while other studies report a sensitivity of up to 96% [6,19]. Therefore, this tool can be accurately used for the detection of asymptomatic cases, which according to our results harbor a malignant potential equal to that of polyps found during investigation of abnormal uterine bleeding. Vaginal sonography in our study detected 96% (26/27) of the pre-malignant and malignant cases. To date, ultrasound screening for uterine pathologies such as endometrial polyps or endometrial cancer is not recommended for large populations of women. Usually, even early endometrial cancer is associated with abnormal vaginal bleeding, leading women to seek immediate early advice. Moreover, as 75% of patients with endometrial cancer are diagnosed at stage I with a favorable outcome, mass screening is unlikely to be cost effective or to improve survival [20,21]. However, some authors advocate endometrial screening in selected group of high-risk patients. These include women treated with estrogen replacement therapy or tamoxifen, diabetic and obese menopausal women or those with genetic susceptibility for endometrial cancer, for example women with hereditary non-polyposis colon cancer (HNPCC) in their family [22,23]. Furthermore, because more women are having imaging studies done in the pelvis for a variety of reasons, more incidental findings like polyps will be diagnosed. In the current study, differently than in other studies, HRT and tamoxifen treatment were not associated with malignancy in the endometrial polyps probably due to the small number of women being treated with these drugs. Older age, menopause status and polyp size were significant factors predicting malignancy in the polyp, although the PPV for malignancy was low. Polyps greater than 1.5 cm were significantly more prone to be associated with malignancy than smaller ones. These results are in accordance with a recent study showing that small endometrial polyps can regress and disappear, while polyps larger than 1 cm were more likely to persist [2]. In our study, in all the 13 cases of malignancy, the histology of the polyp was of a well to moderate endometroid type adenocarcinoma. These cases are considered to be associated with a very good prognosis and one might wonder whether early diagnosis and treatment would contribute to better survival. However, several reports found aggressive uterine neoplasms confined to an endometrial polyp [24– 27]. Moreover, a recent report showed that tamoxifen treatment, a known etiology for endometrial polyps, was associated with aggressive histologic type and grade of uterine cancers [28,29]. These cases with their possible unfavorable
209
outcome, although rare, substantiate the argument in favor of early detection and resection of endometrial polyps, especially in high-risk cohort of women. Our results showing a high malignancy rate in both symptomatic and asymptomatic women without any risk factors are concerning. Finding a non-invasive method to reliably distinguish benign endometrial polyps from polyps that are malignant or pre-malignant would be advantageous and could possibly save surgical removal. Goldstein et al. [5] trying to predict the malignant potential of endometrial polyps by a non invasive methods found that objective assessment of blood flow impedance in endometrial polyps, size of such polyps, or the presence or absence of bleeding cannot predict their malignant potential. Therefore, they recommended that all endometrial polyps should be surgically resected and sent for pathological evaluation. These findings together with our results, contribute to an approach arguing in favor of endometrial screening in high-risk patients. We suggest that in menopause all polyps whether symptomatic or not should be removed and in premenopause active management of hysteroscopic endometrial polypectomy should be offered to symptomatic women, or those with polyps larger than 1.5 cm. Moreover, although screening for endometrial cancer is not recommended, we are aware of many women who are undergoing ‘‘routine’’ pelvic sonography. Evidence based management of incidental ultrasound findings like endometrial polyps, has become an important role of the modern gynecologist. Therefore understanding the significance of both symptomatic and asymptomatic endometreial polyps and their proposed management is important.
References [1] Schlaen I, Bergeron C, Ferenczy A, Wong P, Naves A. Endometrial polyps: a study of 204 cases. Surg Pathol 1988;1:375–82. [2] DeWaay DJ, Syrop CH, Nygaard IE, Davis WA, Van Voorhis BJ. Natural history of uterine polyps and leiomyomata. Obstet Gynecol 2002;100:3–7. [3] De Vries L, Dijkhuizen F, Mol B, Brolmann H, Moret E, Heintz A. Comparison of transvaginal sonography, saline infusion sonography, and hysteroscopy in premenopausal women with abnormal uterine bleeding. J Clin Ultrasound 2000;28:217–23. [4] Widrich T, Bradley L, Mitchinson A, Collins R. Comparison of saline infused sonography with office hysteroscopy for the evaluation of the endometrium. Am J Obstet Gynecol 1996;174:1327–34. [5] Goldstein SR, Monteagudo A, Popiolek D, Mayberry P, TimorTritsch I. Evaluation of endometrial polyps. Am J Obstet Gynecol 2002;186:669–74. [6] Sherman ME, Mazur MT, Kurman RJ. Benign diseases of the endometrium. In: Kurman RJ, editor. Blaunstein’s pathology of the female genital tract. 3rd ed. New York: Springer; 2002. p. 421–66. [7] Caspi B, Appelman Z, Goldchmit R, Ashkenazi M, Haruvy Y, Hagay Z. The bright edge of the endometrial polyp. Ultrasound Obstet Gynecol 2000;15:327–30. [8] Salm R. The incidence and significance of early carcinomas in endometrial polyps. Pathology 1972;108:47–53. [9] Wolfe A, Mackles A. Malignant lesions arising from benign endometrial polyps. Obstet Gynecol 1962;20:542–50.
210
A. Ben-Arie et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 206–210
[10] Anastasiadis PG, Koutlaki NG, Skaphida PG, Galazios GC, Tsikouras PN, Liberis VA. Endometrial polyps: prevalence, detection, and malignant potential in women with abnormal uterine bleeding. Eur J Gynaecol Oncol 2000;21:180–3. [11] Armenia CS. Sequential relationship between endometrial polyps and carcinoma of endometrium. Obstet Gynecol 1967;30: 524–9. [12] Orvieto R, Bar-Hava I, Dicker D, Bar J, Ben-Rafael Z, Neri A. Endometrial polyps during menopause: characterization and significance. Acta Obstet Gynecol Scand 1999;78:883–6. [13] Gray LA, Robertson RW, Christopherson WM. Atypical endometrial changes associated with carcinoma. Gynecol Oncol 1974; 2:93–100. [14] Bakour SH, Khan KS, Gupta JK. The risk of premalignant and malignant pathology in endometrial polyps. Acta Obstet Gynecol Scand 2000;79:317–20. [15] Savelli L, De Iaco P, Santini D, Rosati F, Ghi T, Pignotti E, et al. Histopathologic features and risk factors for benignity, hyperplasia, and cancer in endometrial polyps. Am J Obstet Gynecol 2003;188:927–31. [16] Cohen I, Azaria R, Bernheim J, Shapira J, Beyth Y. Risk factors of endometrial polyps resected from postmenopausal patients with breast carcinoma treated with tamoxifen. Cancer 2001;92:1151–5. [17] Clevenger-Hoeft M, Syrop CH, Stovall DW, Van Voorhis BJ. Sonohysterography in premenopausal women with and without abnormal bleeding. Obstet Gynecol 1999;94:516–20. [18] Tjarks M, Van Voorhis B. Treatment of endometrial polyps. Obstet Gynecol 2000;96:886–9. [19] Emanual MH, Aknum WM, Verdel MJC, Hart AAM. The reproducibility of the results of transvaginal sonography of the
[20] [21] [22]
[23]
[24] [25]
[26] [27]
[28]
[29]
uterus in patients with abnormal uterine bleeding. Ultrasound Obstet Gynecol 1996;8:346–9. Disaia PJ, Creasman WT. Clinical gynecologic oncology. 6th ed. St. Louis: Mosby; 2002. Pritchard KI. Screening for endometrial cancer: is it effective. Ann Intern Med 1989;110:177–9. Varner RE, Sparks JM, Cameron CD, Roberts LL, Soong SJ. Transvaginal sonography of the endometrium in postmenopausal women. Obstet Gynecol 1991;78:195–9. Watson P, Vasen HF, Mecklin JP, Jarvinen H, Lynch HT. The risk of endometrial cancer in hereditary nonpolyposis colorectal cancer. Am J Med 1994;96:516–20. Silva EG, Jenkins R. Serous carcinoma in endometrial polyps. Mod Pathol 1990;3:120–8. Kahner S, Ferenczy A, Richart RM. Homologous mixed mullerian tumours (carcinosarcoma) confined to endometrial polyps. Am J Obstet Gynecol 1975;121:278–9. Barwick KW, Livolsi VA. Heterologous mixed mullerian tumour confined to an endometrial polyp. Obstet Gynecol 1979;53:512–4. Patsner B. Late recurrence of a uterine clear cell adenocarcinoma confined to an endometrial polyp: case report. Eur J Gynaecol Oncol 1998;19:522–3. Magriples U, Naftolin F, Schwartz PE, Carcangiu ML. High grade endometrial carcinoma in tamoxifen treated breast cancer patients. J Clin Oncol 1993;11:485–90. Bergman L, Beelen ML, Gallee MP, Hollema H, Benraadt J, van Leeuwen FEComprehensive Cancer Centres ALERT Group. Risk and prognosis of endometrial cancer after tamoxifen for breast cancer. Assessment of liver and endometrial cancer risk following tamoxifen. Lancet 2000;356:881–7.
The malignant potential of endometrial polyps Alon Ben-Arie*, Chen Goldchmit, Yosef Laviv, Roni Levy, Benjamin Caspi, Monica Huszar, Ram Dgani, Zion Hagay Department of Obstetrics and Gynecology, Kaplan Medical Center (affiliated to the Hadassah Medical School and the Hebrew University, Jerusalem), P.O. Box 1, Rehovot 76100, Israel Received 6 October 2003; received in revised form 22 December 2003; accepted 2 February 2004
Abstract Objectives: To determine the pre-malignant and malignant potential of endometrial polyps and to assess whether different clinical parameters are associated with malignancy in the polyps. Study design: Four hundred and thirty consecutive cases of hysteroscopic diagnosis of endometrial polyp were retrieved. The medical records, preoperative vaginal sonography results and histopathology findings were reviewed. Statistical analysis was performed. Results: Hysteroscopy truly identified endometrial polyps in 95.7% of the cases. In 11.4% cases, hyperplasia without atypia was found in the endometrial polyp. In 3.3 and 3.0% of women pre-malignant or malignant conditions were found in the polyp. Older age, menopause status and polyps larger than 1.5 cm were associated with significant pre-malignant or malignant changes, although the positive predictive value for malignancy was low. All the malignant polyps were diagnosed only in postmenopausal women. The presence of postmenopausal or irregular vaginal bleeding, was not a predictor of malignancy in the polyp. Conclusions: Postmenopausal women with endometrial polyps are at increased risk of malignancy in the polyp. Those patients, whether symptomatic or not should be evaluated by hysteroscopic resection of the polyps. Asymptomatic premenopausal patients with polyps smaller than 1.5 cm can be observed. # 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Endometrial polyp; Malignancy; Hysteroscopy
1. Introduction The endometrial polyp is a pedunculated or sessile excretion of the endometrium containing variable amounts of glands, stroma and blood vessels. Endometrial polyps are common pathological lesions of the uterine corpus and are usually found in perimenopausal women [1–6]. The prevalence of endometrial polyps in the general female population is estimated to be approximately 25% [6]. Abnormal uterine bleeding is frequently the presenting symptom but polyps are often asymptomatic and found during routine ultrasound or infertility investigations [1,5,7]. Polyps are in general benign growths [1], although malignancy confined to a polyp has also been identified [8–10]. The incidence of carcinoma confined to endometrial polyps varies between 0 and 4.8% depending on the selection of patients and the methods used in making the diagnosis. Polyps have been also described to be associated with premalignant changes of the endometrium, but their role as precursors for endometrial carcinoma has not been established [10–15].
* Corresponding author. Tel.: þ972-8-9441231; fax: þ972-8-9441776. E-mail address: [email protected] (A. Ben-Arie).
The objective of this study was to evaluate the risk of premalignant and malignant changes in endometrial polyps in symptomatic and asymptomatic women. Secondly to investigate whether clinical and ultrasonic parameters may predict malignancy in a polyp and by thus decide which polyp should be removed or observed.
2. Patients and method A retrospective study was conducted dating from January 1996 to July 2001. All consecutive cases where endometrial polyp was suspected by diagnostic hysteroscopy were included. The charts of 430 women aged 26–89 (mean 54.6 years) were reviewed. The study group included both pre- and postmenopausal women. Women were considered postmenopausal if they reported a period of at least 12 months of amenorrhea after the age of 45 years. The study group included women being treated with hormonal replacement therapy (HRT) and postmenopausal women being treated with tamoxifen as an adjuvant treatment for breast cancer (Table 1). Diagnostic hysteroscopy was performed in women with abnormal uterine bleeding, or in asymptomatic women in
0301-2115/$ – see front matter # 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejogrb.2004.02.002
A. Ben-Arie et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 206–210 Table 1 Menopause status, hormonal treatment and symptoms in the study group Clinical parameters
No. of patients (%)
Pre-menopause Menopause HRT Tamoxifen Asymptomatic Abnormal uterine bleeding
206 224 49 18 236 194
(47.9) (52.1) (11.7) (4.3) (54.9) (45.1)
HRT: hormone replacement therapy.
whom abnormal endometrial finding was suspected on routine pelvic sonography performed as a complementary mode during the annual gynecological examination. Abnormal uterine bleeding was defined as any vaginal bleeding in postmenopausal women not receiving HRT or as irregular vaginal bleeding in women still actively menstruating or being treated with HRT. All women included in the study underwent endometrial evaluation at the gynecologic ultrasound unit. Scans were performed using a real-time ultrasound scanners (Advanced Laboratories 3000, Bothell, Washington, USA) equipped with 5–9 MHz transvaginal probe and Vuluson 730 (Kretztechnik AG, Austria) equipped with 5–9 MHz transvaginal probe. The ultrasonographic diagnosis of endometrial polyp was based on the presence of the ‘‘bright edge of the polyps’’ as described elsewhere [7]. The largest polyp diameter as was measured during the ultrasound scan was recorded. Diagnostic hysteroscopy was performed, using saline infusion as a distention medium, and a Storz Endoscope (Tuttlingen, Germany), with a 5 mm diagnostic sheath. In patients, where a polyp was found, at the same session, immediate cervical dilatation, endometrial polypectomy and uterine curettage were performed. The operative procedure was done under general anesthesia or a paracervical block with 20 ml of lidocaine injection. The specimens were placed in 10% formaldehyde for histological examination. Histopathologic diagnosis distinguished between non-polypoid lesions that were mistakenly diagnosed as polyps (leiomyoma, atrophic, proliferative or secretory endometrium) and endometrial polyps, which were classified as benign, hyperplasic (simple or complex hyperplasia), premalignant (simple or complex hyperplasia with atypia), and those harboring carcinoma. The histopathalogic definitions of endometrial hyperplasia and adenocarcinoma were according to the following definitions. Endometrial simple hyperplasia was defined by the endometrial architecture that is moderately distorted with crowding of glands, cystic dilatation and noncomplex budding. The lining epithelium of the glands is pseudostratified showing mitotic activity with no atypia of cells. Endometrial complex hyperplasia was defined by crowding of branched complex glands lined by stratified mitotically active cigar shaped cells. Atypical simple hyperplasia was defined by architecture similar to simple hyperplasia, but the glands are more irregular. The
207
glands are lined by atypical cells with round hyperchromatic nuclei and granular chromatin. In atypical complex hyperplasia the architecture is similar to endometrial complex hyperplasia, but the glands are crowded and lined by atypical cells with round pleomorphic nuclei. Endometrial carcinoma was defined by crowded malignant tubular glands varying in size and invading the stroma. The glands are lined by cells showing marked atypia and mitotic activity. Ten patients were excluded from the study due to insufficient material for histologic diagnosis. The prevalence of the premalignant lesions that included endometrial hyperplasia with atypia and the risk of adenocarcinoma confined to the endometrial polyps in different subgroups of women were evaluated. The results are presented as mean S:D: Statistical analysis was performed using the Student’s t-test, chi-square test, Fisher’s-exact test and the positive and negative predictive (PPV/NPV) values were calculated where appropriate; P < 0:05 was considered significant.
3. Results The histopathalogic evaluation of 420 cases where an endometrial polyp was seen on diagnostic hysteroscopy is shown in Table 2. Hysteroscopic diagnosis of endometrial polyp compared to the histological findings, showed that hysteroscopy truly identified endometrial polyps in 95.7%
Table 2 Histological results of 420 endometrial polyps found on hysteroscopy Histology Benign Endometrial polyps Endometrial polyps with simple hyperplasia (no atypia) Endometrial polyps with complex hyperplasia (no atypia) Subtotal Pre-malignant or malignant Endometrial polyps with simple hyperplasia and atypia Endometrial polyps with complex hyperplasia and atypia Invasive endometrial carcinoma Subtotal Other Myoma Complex hyperplasia without atypia Atrophic endometrium Proliferate endometrium Secretory endometrium Subtotal Total
No.
%
327 36
77.9 8.6
12
2.9
375
89.3
11
2.6
3
0.7
13
3.0
27
6.3
9 2 1 5 1
2.1 0.5 0.2 1.2 0.2
18
4.3
420
100
208
A. Ben-Arie et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 206–210
Table 3 Benign vs. pre-malignant or malignant in subgroups of women with endometrial polyp
Menopause Abnormal bleeding Polyp > 1.5 cm HRT Tamoxifen
Benign—375 cases (%)
Pre-malignant or malignant—27 cases (%)
Statistical significance
196 176 150 47 16
22 12 21 2 2
P < 0.001 N.S P < 0.0001 N.S N.S
(49.9) (44.8) (37.3) (12.0) (4.1)
(81.5) (44.4) (77.8) (7.4) (7.4)
HRT: hormone replacement therapy.
(402/420) of the cases. In 18 cases other benign histologies were reported, mainly uterine myomas. In 11.4% cases, hyperplasia without atypia was found in the endometrial polyp. In 3.3 and 3.0% of women premalignant or malignant histologies, respectively were found to be confined to the polyp. Pre-malignant and malignant changes confined to the polyps were more common in older age (61:1 9:8 versus 54:2 11:6 years, P ¼ 0:03) and in larger polyps (3:2 1:9 versus 1:7 1:1 cm, P < 0:0001). Endometrial polyps larger than 1.5 cm were significantly associated with pre-malignant and malignant states compared to smaller polyps (Table 3). The PPV and NPV for malignancy in polyps larger than 1.5 cm were 0.1295 and 0.9733, respectively. Premenopausal women had significantly more abnormal uterine bleeding than women at menopause (56.8 versus 34.4%, P < 0:0001). On the other hand, there were significantly more women at menopause, whose endometrial polyp was found to be malignant or pre-malignant with a PPV of 0.1161 and a NPV of 0.7835. All 13 cases of malignant polyps and 9 of 14 cases with atypical hyperplasia were diagnosed in postmenopausal women. The presence of symptoms, whether in menopause or not, was not a predictor of premalignancy or malignancy (Table 3) as reflected in the PPV of 0.0689, with a NPV of 0.9299. Hormone replacement therapy and tamoxifen treatment were not found to be statistically significant factors associated with malignancy within an endometrial polyp (Table 3). Vaginal sonography had a sensitivity of 86% compared to diagnostic hysteroscopy for the identification of endometrial polyp. Of 27 pre-malignant and malignant endometrial polyps found in the study, only one polyp was not imaged by vaginal sonography (P < 0:0001).
4. Discussion In this study, one of the largest of those published concerning the significance of endometrial polyps, we have found a high rate (6.3%) of malignant and pre-malignant conditions confined to the polyps. In 3.3% of the polyps atypical hyperplasia was found, a known precursor of endometrial cancer and in 3.0% of the cases invasive endometrial adenocarcinoma was identified within the polyp.
Previous studies reported the malignancy rate in endometrial polyps to vary between 0 and 4.8% [5,9,10,12,14,15]. Orvieto et al. [12] examined 146 endometrial polyps and found 10% cases of hyperplasia (2.5% with atypia) without even one case of carcinoma. Bakour et al. [14] in a prospective cohort study of women with abnormal uterine bleeding, compared the histological findings in 62 endometrial polyps, to 186 cases where no polyp was seen. They found two (3.2%) cases of endometrial carcinoma in the polyps and 11.4% of hyperplasia (6.5% with atypia). Recently Savelli et al. [15], in the largest series to date, found that 16 of 509 polyps (3.1%) had hyperplasia with atypia, and only 4 polyps (0.8%) were cancerous. In an older study, Armenia showed that 3.5% of the women who were diagnosed with endometrial polyps, eventually had endometrial cancer in their follow-up [11]. Reasons for the discrepancies in the rate of hyperplasia and carcinoma in endometrial polyps may lie in biases originating in the small size of some of the published series and in the different diagnostic tools used. Currently, isolated endometrial polyps are diagnosed by hysteroscopy. In older studies, where uterine curettage rather than hysteroscopy was used, it was difficult to asses whether uterine pathologies were associated with an endometrial polyp, or were truly confined to a polyp. Another possible reason for the varying malignant rate found in endometrial polyps may be associated with different rates of high-risk patients for endometrial malignancy in the cohort studied. Age, race, menopausal status and HRT or tamoxifen treatments are known risk factors for endometrial cancer. Our study showed a relatively high rate of malignancy in endometrial polyps, comparable to the 3% malignancy rate found by Cohen et al. [16] in a high-risk cohort of women treated with tamoxifen. High rates of endometrial polyps are asymptomatic and are found incidentally on routine vaginal sonography [2,17,18]. Goldstein et al. [5] found that 19 of 61 polyps (31.1%) were incidental findings. In the current series, 236 (54.9%) of the resected polyps were asymptomatic. Most importantly, symptoms in our study could not differentiate benign from malignant or pre-malignant states and 15/232 (6.5%) of the asymptomatic polyps harbored atypical hyperlasia or carcinoma, an equal rate of malignancy that was found in the symptomatic polyps. Likewise, Savelli et al. [15] have also shown a high rate of polyps to be asymptomatic incidental finding with an equal rate of premalignant
A. Ben-Arie et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 206–210
or malignant conditions compared to polyps that were associated with abnormal uterine bleeding. Transvaginal sonography or saline infusion sonohysterography can be used for accurate diagnosis of endometrial polyps. The detection rate of endometrial polyps is strongly influenced by the diagnostic method used and by the experience of the sonographer [2,17,19]. In this study, vaginal sonography had a sensitivity of 86% in identification of endometrial polyps, while other studies report a sensitivity of up to 96% [6,19]. Therefore, this tool can be accurately used for the detection of asymptomatic cases, which according to our results harbor a malignant potential equal to that of polyps found during investigation of abnormal uterine bleeding. Vaginal sonography in our study detected 96% (26/27) of the pre-malignant and malignant cases. To date, ultrasound screening for uterine pathologies such as endometrial polyps or endometrial cancer is not recommended for large populations of women. Usually, even early endometrial cancer is associated with abnormal vaginal bleeding, leading women to seek immediate early advice. Moreover, as 75% of patients with endometrial cancer are diagnosed at stage I with a favorable outcome, mass screening is unlikely to be cost effective or to improve survival [20,21]. However, some authors advocate endometrial screening in selected group of high-risk patients. These include women treated with estrogen replacement therapy or tamoxifen, diabetic and obese menopausal women or those with genetic susceptibility for endometrial cancer, for example women with hereditary non-polyposis colon cancer (HNPCC) in their family [22,23]. Furthermore, because more women are having imaging studies done in the pelvis for a variety of reasons, more incidental findings like polyps will be diagnosed. In the current study, differently than in other studies, HRT and tamoxifen treatment were not associated with malignancy in the endometrial polyps probably due to the small number of women being treated with these drugs. Older age, menopause status and polyp size were significant factors predicting malignancy in the polyp, although the PPV for malignancy was low. Polyps greater than 1.5 cm were significantly more prone to be associated with malignancy than smaller ones. These results are in accordance with a recent study showing that small endometrial polyps can regress and disappear, while polyps larger than 1 cm were more likely to persist [2]. In our study, in all the 13 cases of malignancy, the histology of the polyp was of a well to moderate endometroid type adenocarcinoma. These cases are considered to be associated with a very good prognosis and one might wonder whether early diagnosis and treatment would contribute to better survival. However, several reports found aggressive uterine neoplasms confined to an endometrial polyp [24– 27]. Moreover, a recent report showed that tamoxifen treatment, a known etiology for endometrial polyps, was associated with aggressive histologic type and grade of uterine cancers [28,29]. These cases with their possible unfavorable
209
outcome, although rare, substantiate the argument in favor of early detection and resection of endometrial polyps, especially in high-risk cohort of women. Our results showing a high malignancy rate in both symptomatic and asymptomatic women without any risk factors are concerning. Finding a non-invasive method to reliably distinguish benign endometrial polyps from polyps that are malignant or pre-malignant would be advantageous and could possibly save surgical removal. Goldstein et al. [5] trying to predict the malignant potential of endometrial polyps by a non invasive methods found that objective assessment of blood flow impedance in endometrial polyps, size of such polyps, or the presence or absence of bleeding cannot predict their malignant potential. Therefore, they recommended that all endometrial polyps should be surgically resected and sent for pathological evaluation. These findings together with our results, contribute to an approach arguing in favor of endometrial screening in high-risk patients. We suggest that in menopause all polyps whether symptomatic or not should be removed and in premenopause active management of hysteroscopic endometrial polypectomy should be offered to symptomatic women, or those with polyps larger than 1.5 cm. Moreover, although screening for endometrial cancer is not recommended, we are aware of many women who are undergoing ‘‘routine’’ pelvic sonography. Evidence based management of incidental ultrasound findings like endometrial polyps, has become an important role of the modern gynecologist. Therefore understanding the significance of both symptomatic and asymptomatic endometreial polyps and their proposed management is important.
References [1] Schlaen I, Bergeron C, Ferenczy A, Wong P, Naves A. Endometrial polyps: a study of 204 cases. Surg Pathol 1988;1:375–82. [2] DeWaay DJ, Syrop CH, Nygaard IE, Davis WA, Van Voorhis BJ. Natural history of uterine polyps and leiomyomata. Obstet Gynecol 2002;100:3–7. [3] De Vries L, Dijkhuizen F, Mol B, Brolmann H, Moret E, Heintz A. Comparison of transvaginal sonography, saline infusion sonography, and hysteroscopy in premenopausal women with abnormal uterine bleeding. J Clin Ultrasound 2000;28:217–23. [4] Widrich T, Bradley L, Mitchinson A, Collins R. Comparison of saline infused sonography with office hysteroscopy for the evaluation of the endometrium. Am J Obstet Gynecol 1996;174:1327–34. [5] Goldstein SR, Monteagudo A, Popiolek D, Mayberry P, TimorTritsch I. Evaluation of endometrial polyps. Am J Obstet Gynecol 2002;186:669–74. [6] Sherman ME, Mazur MT, Kurman RJ. Benign diseases of the endometrium. In: Kurman RJ, editor. Blaunstein’s pathology of the female genital tract. 3rd ed. New York: Springer; 2002. p. 421–66. [7] Caspi B, Appelman Z, Goldchmit R, Ashkenazi M, Haruvy Y, Hagay Z. The bright edge of the endometrial polyp. Ultrasound Obstet Gynecol 2000;15:327–30. [8] Salm R. The incidence and significance of early carcinomas in endometrial polyps. Pathology 1972;108:47–53. [9] Wolfe A, Mackles A. Malignant lesions arising from benign endometrial polyps. Obstet Gynecol 1962;20:542–50.
210
A. Ben-Arie et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 206–210
[10] Anastasiadis PG, Koutlaki NG, Skaphida PG, Galazios GC, Tsikouras PN, Liberis VA. Endometrial polyps: prevalence, detection, and malignant potential in women with abnormal uterine bleeding. Eur J Gynaecol Oncol 2000;21:180–3. [11] Armenia CS. Sequential relationship between endometrial polyps and carcinoma of endometrium. Obstet Gynecol 1967;30: 524–9. [12] Orvieto R, Bar-Hava I, Dicker D, Bar J, Ben-Rafael Z, Neri A. Endometrial polyps during menopause: characterization and significance. Acta Obstet Gynecol Scand 1999;78:883–6. [13] Gray LA, Robertson RW, Christopherson WM. Atypical endometrial changes associated with carcinoma. Gynecol Oncol 1974; 2:93–100. [14] Bakour SH, Khan KS, Gupta JK. The risk of premalignant and malignant pathology in endometrial polyps. Acta Obstet Gynecol Scand 2000;79:317–20. [15] Savelli L, De Iaco P, Santini D, Rosati F, Ghi T, Pignotti E, et al. Histopathologic features and risk factors for benignity, hyperplasia, and cancer in endometrial polyps. Am J Obstet Gynecol 2003;188:927–31. [16] Cohen I, Azaria R, Bernheim J, Shapira J, Beyth Y. Risk factors of endometrial polyps resected from postmenopausal patients with breast carcinoma treated with tamoxifen. Cancer 2001;92:1151–5. [17] Clevenger-Hoeft M, Syrop CH, Stovall DW, Van Voorhis BJ. Sonohysterography in premenopausal women with and without abnormal bleeding. Obstet Gynecol 1999;94:516–20. [18] Tjarks M, Van Voorhis B. Treatment of endometrial polyps. Obstet Gynecol 2000;96:886–9. [19] Emanual MH, Aknum WM, Verdel MJC, Hart AAM. The reproducibility of the results of transvaginal sonography of the
[20] [21] [22]
[23]
[24] [25]
[26] [27]
[28]
[29]
uterus in patients with abnormal uterine bleeding. Ultrasound Obstet Gynecol 1996;8:346–9. Disaia PJ, Creasman WT. Clinical gynecologic oncology. 6th ed. St. Louis: Mosby; 2002. Pritchard KI. Screening for endometrial cancer: is it effective. Ann Intern Med 1989;110:177–9. Varner RE, Sparks JM, Cameron CD, Roberts LL, Soong SJ. Transvaginal sonography of the endometrium in postmenopausal women. Obstet Gynecol 1991;78:195–9. Watson P, Vasen HF, Mecklin JP, Jarvinen H, Lynch HT. The risk of endometrial cancer in hereditary nonpolyposis colorectal cancer. Am J Med 1994;96:516–20. Silva EG, Jenkins R. Serous carcinoma in endometrial polyps. Mod Pathol 1990;3:120–8. Kahner S, Ferenczy A, Richart RM. Homologous mixed mullerian tumours (carcinosarcoma) confined to endometrial polyps. Am J Obstet Gynecol 1975;121:278–9. Barwick KW, Livolsi VA. Heterologous mixed mullerian tumour confined to an endometrial polyp. Obstet Gynecol 1979;53:512–4. Patsner B. Late recurrence of a uterine clear cell adenocarcinoma confined to an endometrial polyp: case report. Eur J Gynaecol Oncol 1998;19:522–3. Magriples U, Naftolin F, Schwartz PE, Carcangiu ML. High grade endometrial carcinoma in tamoxifen treated breast cancer patients. J Clin Oncol 1993;11:485–90. Bergman L, Beelen ML, Gallee MP, Hollema H, Benraadt J, van Leeuwen FEComprehensive Cancer Centres ALERT Group. Risk and prognosis of endometrial cancer after tamoxifen for breast cancer. Assessment of liver and endometrial cancer risk following tamoxifen. Lancet 2000;356:881–7.