Hysteroscopic surgery does not adversely affect the long-term prognosis of women with endometrial adenocarcinoma

Journal of Minimally Invasive Gynecology (2007) 14, 205-210

Hysteroscopic surgery does not adversely affect the long-term prognosis of women with endometrial adenocarcinoma George A. Vilos, MD, Fawaz Edris, MD, Awatif Al-Mubarak, MD, Helen C. Ettler, MB ChB, Jackie Hollett-Caines, MD, and Basim Abu-Rafea, MD From the Department of Obstetrics and Gynecology (Drs. Vilos, Edris, Al-Mubarak, Hollett-Caines, and Abu-Rafea), and Department of Pathology (Dr. Ettler), The University of Western Ontario, London, Ontario, Canada. KEYWORDS: Uterine cancer; Endometrial adenocarcinoma; Resectoscopic surgery; Hysteroscopy

Abstract STUDY OBJECTIVE: To determine the effect of hysteroscopic surgery on the long-term clinical outcome of women diagnosed with endometrial cancer. DESIGN: Prospective cohort study (Canadian Task Force classification II-3). SETTING: University-affiliated teaching hospital. PATIENTS: From January 1990 through December 2005, the principal author (GAV) performed primary hysteroscopic surgery in 3401 women with abnormal uterine bleeding. Among these women, there were 16 occult and 3 known endometrial cancers. INTERVENTIONS: All women underwent hysteroscopic evaluation and partial (n ⫽ 8) or complete (n ⫽ 11) rollerball electrocoagulation and/or endomyometrial resection. After diagnosis of endometrial malignancy, women were counseled regarding their disease and management, in accordance with established clinical practice guidelines. Follow-up ranged from 1 to 14 years and was conducted by office visits and telephone interviews. MEASUREMENTS AND MAIN RESULTS: Among the 3401 women, there were 19 women with endometrial adenocarcinoma, 3 of whom were known to harbor cancer before hysteroscopic surgery. One woman refused hysterectomy and remains alive and well 5 years after total hysteroscopic endomyometrial resection. Two women wished to maintain fertility; 1 consented to hysterectomy after incomplete resection of her lesion. The other was treated with progestins. Her cancer reverted to complex hyperplasia, and she requested hysterectomy 4 years later. No residual cancer was found. After 5 years of follow-up, 1 patient died from carcinoma of the gallbladder (2 years), and 2 died at 4 years; 1 at the age of 87 years of natural causes and the other at the age of 86 years from acute renal failure unrelated to her cancer. Fourteen women remain alive and well at 5 to 14 years of follow-up. Two additional women remain alive and well at 1 and 4 years of follow-up. CONCLUSION: Resectoscopic surgery did not adversely affect the 5-year survival and the long-term prognosis in 14 women with endometrial cancer. © 2007 AAGL. All rights reserved.

The authors have no commercial, proprietary, or financial interest in the products or companies described in this article. Corresponding author: George A. Vilos, MD, Professor, Department of Obstetrics and Gynecology, St. Joseph’s Health Care, 268 Grosvenor Street, London, Ontario, Canada, N6A 4V2. E-mail: [email protected] Submitted May 31, 2006. Accepted for publication October 13, 2006.

1553-4650/$ -see front matter © 2007 AAGL. All rights reserved. doi:10.1016/j.jmig.2006.10.010

206

Journal of Minimally Invasive Gynecology, Vol 14, No 2, March/April 2007

Abnormal uterine bleeding (AUB) is experienced by 10% to 30% of reproductive-age women.1,2 In menopausal and premenopausal women (with certain risk factors), AUB is associated with an approximately 10% risk of endometrial cancer.3–5 Traditionally, investigation of AUB included fractional dilatation and curettage (D&C) in the operating room under general anesthesia. It has been reported that D&C provides adequate sampling in 75% of women and misses up to 10% of pathology.6 In the presence of positive diagnosis or in the absence of improved symptomatology, hysterectomy was carried out. Hysterectomy was particularly attractive to physicians and women not only because it eliminated the patients’ bleeding, but because it could detect occult uterine malignancies missed by D&C. Hysteroscopic endometrial ablation was introduced in the 1980s, while nonhysteroscopic or global endometrial ablation (GEA) was introduced in the 1990s. Both were designed as alternatives to hysterectomy for treatment of women with AUB of benign etiology.7 Before endometrial ablation, it is recommended that AUB be investigated in accordance with established clinical practice guidelines.8 Such guidelines recommend that office endometrial biopsy or D&C be performed in all women with postmenopausal bleeding and in premenopausal women with certain risk factors for endometrial hyperplasia and cancer such as irregular bleeding; age greater than 45 years; obesity (weight ⬎ 90 kg or BMI ⬎ 27 kg/m2); personal history of polycystic ovarian syndrome, infertility, or nulliparity; or family history of endometrial or colon cancer.4,5,8 Office endometrial biopsy is occasionally impossible to perform or provides an inadequate sample due to technical issues, cervical stenosis, or other patient conditions such as morbid obesity and discomfort. Under such circumstances, we have adopted the philosophy of performing hysteroscopic evaluation with directed biopsies or endomyometrial resection in the operating room under appropriate anesthesia and optimal medical conditions. In the present study, we identified 19 women with AUB and endometrial cancer. In 16 of these women, prehysteroscopy office endometrial biopsy was not done, impossible to perform, provided inadequate sample, or was reported as proliferative endometrium or endometrial hyperplasia. The diagnosis of endometrial cancer was known preoperatively in 3 women. The purpose of the study is to determine the effect of hysteroscopic surgery on the long-term (more than 5 years) clinical outcome in women with endometrial cancer.

Material and methods From January 1990 through December 2005, the principal author (GAV) performed primary hysteroscopic endometrial ablation using electrocoagulation with rollerball, resection with a loop electrode, or a combination of both in

3399 women with AUB. Two women with known endometrial cancer wished to retain fertility and had only a biopsy with the loop electrode with no electrical power applied to it. A 26F (⬃9 mm) diameter resectoscope and 3-to 5-mm rollerballs or 8-mm diameter loop electrodes were used to coagulate or cut tissue at 100 ⫾ 20 W of power. The uterus was distended/irrigated with 1.5% glycine solution at 100 cm H2O (⬃75 mm Hg) pressure with 100 ⫾ 20 mm Hg suction to evacuate bubbles, clots, and debris from the uterus. As a rule, resection rather than rollerball ablation of the entire endometrium was performed in the absence of a recent (⬍ 6 months) negative endometrial biopsy; in the presence of intrauterine polyp(s), myoma(s), or suspicious lesion(s); and in women with any of the risk factors for endometrial hyperplasia or cancer, as described in the introduction. Having adopted the above principles, we identified 16 endometrial adenocarcinomas (Tables 1 and 2). The characteristics of the patients, preoperative endometrial biopsy, hysteroscopic findings and treatment, posthysteroscopy pathology and treatment, and clinical outcomes are also listed in the corresponding tables. Details of the first 8 patients in Table 1 and first 5 patients in Table 2 have been previously published.9

Results Table 1 includes 10 women with undiagnosed and 1 woman with known endometrial adenocarcinoma before hysteroscopy. All these women had attempted complete endometrial ablation/resection. Patient number 9 was diagnosed by preoperative endometrial biopsy to have endometrioid adenocarcinoma (grade I). This patient declined traditional hysterectomy and bilateral salpingo-oophorectomy (BSO) and agreed to only hysteroscopic evaluation and possible endomyometrial resection. At hysteroscopy, a solitary 1.5-cm exophytic lesion was identified against a background of normal-appearing endometrium. After total endomyometrial resection and refusal of adjuvant therapy, she remains amenorrheic and alive and well 5-years postoperatively, with no evidence of recurrent disease by physical examination, transvaginal ultrasound, or magnetic resonance imaging evaluation of the pelvis. This case has been accepted for publication.10 After traditional hysterectomy and BSO, no residual cancer was found in 8 out of 10 hysterectomy specimens. Adenomyosis was found in 3 women. Table 2 lists 8 women who had incomplete hysteroscopic endometrial ablation. Two women with known endometrioid adenocarcinoma wished further evaluation and exploration of alternatives to hysterectomy to retain their fertility. After hysteroscopic and pathologic confirmation of adenocarcinoma, in a background of diffuse atypical complex hyperplasia, 1 patient (No. 4) requested hysterectomy and BSO within 6 weeks of hysteroscopic confirmation of her cancer, while the other (No. 8) underwent treatment with

Vilos et al

Hysteroscopy and uterine malignancy

207

Table 1 Patient characteristics, preoperative endometrial biopsy, hysteroscopic findings and procedure, and clinical outcomes in women with endometrial cancer in whom complete endometrial ablation/resection was intended Preoperative endometrial Age (yrs)/ BMI Patient parity kg/m2 biopsy/grade

Hysteroscopic Hysteroscopic findings treatment

1

38/1

26.4

Not done

Polypoid

2

68/4

36.8

3

56/1

32.1

4

51/0

30.8

5

57/2

30.0

6

54/3

29.9

7

64/2

25.5

Simple hyperplasia Complex hyperplasia Atypical complex hyperplasia Atypical simple hyperplasia Inadequate sample Unable to do

8

53/2

37.6

9

53/2

10 11

*Hysteroscopic *Hysterectomy ⫹ BSO Clinical pathology/grade pathology/grade outcome/yrs Single focus Adeno-Ca/II No residual Ca

A&W/14

Polypoid

†Rollerball and Adeno-Ca/I resection Total resection Adeno-Ca/I

Polypoid

Total resection Adeno-Ca/I

‡No residual Ca

A&W/8

Polypoid

Total resection Adeno-Ca/I

No residual Ca

A&W/7

Polyps

Adeno-Ca/II

A&W/6

No residual Ca Adenomyosis No residual Ca Adenomyosis

A&W/6

Polypoid

Total resection Adeno-Ca/II and adenomyosis †Rollerball and Adeno-Ca/I and resection adenomyosis Total resection Adeno-Ca/I

Unable to do

Polypoid

Total resection Adeno-Ca/I

20.0

Adeno-Ca/I

Total resection Adeno-Ca/I

62/2

26.8

Solitary cancerous lesion Polypoid

No residual Ca Adenomyosis Refused surgery

45/2

22.7

Inadequate sample Proliferative Polypoid endometrium

Polypoid

Total resection Adeno-Ca/I ‡No residual Ca (mucinous) D&C ⫹ Adeno-Ca/I and No residual Ca rollerball atypical simple hyperplasia

A&W/12

Died from gallbladder carcinoma/2 A&W/6 A&W/5

A&W/5 A&W/1

A&W ⫽ alive and well; Adeno-Ca ⫽ adenocarcinoma; BMI ⫽ body mass index; BSO ⫽ bilateral salpingo-oophorectomy; Ca ⫽ cancer; D&C ⫽ dilatation and curettage. *Endometrioid adenocarcinoma unless specified. †Rollerball of the cornua and fundus and resection of the remaining endometrium. ‡Negative peritoneal washings.

oral Provera (Pfizer, New York, NY) 300 mg daily. Office endometrial biopsy was reported as complex endometrial hyperplasia with atypia at 12 months and without atypia at 24 months. This patient consented to hysterectomy and BSO 4 years after the initial diagnosis of cancer. The pathology of the uterus was reported as adenomyosis with no residual cancer or hyperplasia. In the remaining 6 women in Table 2, the endometrium appeared highly suspicious for cancer and several strips of tissue were taken with the loop electrode (incomplete resection) mostly for diagnostic purposes. Except for 2 women, all women found to have endometrial cancer at the time of their resectoscopic surgery underwent definitive surgical management within 4 to 8 weeks, in the form of hysterectomy and BSO with or without lymphadenectomy and adjuvant therapy (as was recommended by the local tumor board). All women were followed regularly as per local oncology protocols. In Tables 1 and 2, we list 14 women with endometrial cancer and a median follow-up of 7 years (range 5–14

years). All of these women are alive and well with no evidence of recurrent disease. Two additional women are alive and well at 1- and 4-year follow-up.

Discussion Endometrial cancer is the most common malignancy of the female genital tract.5 Fortunately, 90% of women with endometrial cancer develop AUB, which prompts investigation and leads to diagnosis of most cancers (72%) in stage I. Following early diagnosis and treatment by hysterectomy and BSO, with or without lymphadenectomy and adjuvant therapy, the overall 5-year survival rate is approximately 90%.5 To our knowledge, this is the first study to examine the effects of inadvertent or deliberate resectoscopic surgery on the 5-year survival of women with endometrial cancer. The study also highlights several issues encountered by clinicians during diagnosis and treatment of women with AUB.

208

Journal of Minimally Invasive Gynecology, Vol 14, No 2, March/April 2007

Table 2 Patient characteristics, preoperative endometrial biopsy, hysteroscopic findings and procedure, and clinical outcomes in women with endometrial cancer in whom complete endometrial ablation/resection was not intended Preoperative endometrial Age (yrs)/ BMI Patient parity kg/m2 biopsy/grade

*Hysteroscopic †Hysteroscopy findings treatment

1

45/2

30.7

Simple hyperplasia

Polypoid

Partial resection Adeno-Ca/I

2 3

46/0 63/0

54.5 40.0

Partial resection Adeno-Ca/I Partial resection Adeno-Ca/I

4

35/0

36.8

Unable to do Polypoid Proliferative Polypoid endometrium Adeno-Ca/I Polypoid

5

49/2

27.8

Proliferative Polypoid endometrium

6

83/1

51.8

Unable to do

Polypoid

Partial resection Adeno-Ca/I

7

82/3

26.4

Polypoid

8

32/0

35

Inadequate sample Adeno-Ca/I, complex atypical hyperplasia

Partial resection Adeno-Ca/I (mucinous) D&C ⫹ Adeno-Ca/I, polypectomy atypical complex hyperplasia

Polypoid

‡Hysteroscopic ‡Hysterectomy ⫹ BSO Clinical pathology/grade pathology/grade outcome/yrs

Biopsy

Adeno-Ca/I, atypical complex hyperplasia Partial resection Adeno-Ca/I, atypical complex hyperplasia

Adeno-Ca/I (mucinous), complex atypical hyperplasia, adenomyosis Adeno-Ca/I Adeno-Ca/I

A&W/11

A&W/8 A&W/8

§Adeno-Ca/I, atypical A&W/7 complex hyperplasia, adenomyosis A&W/6 §No malignancy, focal atypical complex hyperplasia, adenomyosis Adeno-Ca/I Died of natural causes/4 Adeno-Ca/I Died from ARF/4 No residual Ca, A&W/4 adenomyosis

A&W ⫽ alive and well; Adeno-Ca ⫽ adenocarcinoma; ARF ⫽ acute renal failure; BMI ⫽ body mass index; BSO ⫽ bilateral salpingo-oophorectomy; D&C ⫽ dilatation and curettage. *Polypoid highly suspicious for malignancy. †Partial resection. ‡Endometrioid adenocarcinoma unless specified. §Negative peritoneal cytology.

Many surgeons advocate routine use of thorough curettage before rollerball electrocoagulation or routinely perform resection of the entire endometrium for histologic evaluation when treating AUB in the presence of benign disease. Indeed, routine endometrial resection has identified several unsuspected uterine malignancies in premenopausal and postmenopausal women.9 This issue has become particularly significant since the introduction and use of second-generation endometrial ablation technologies, or GEA. These ablations are usually performed without the use of hysteroscopy. It is expected that cases of undiagnosed uterine malignancies will be inadvertently ablated using GEA technologies. Therefore, when using these technologies, many surgeons recommend thorough pre-ablation evaluation of the endometrial cavity in all women with AUB. Several concerns have been raised by oncologists regarding routine and indiscriminate use of diagnostic, and especially resectoscopic, hysteroscopy in women suspected of harboring uterine malignancies.11 The first concern is transfallopian tube dissemination of malignant cells into the

peritoneal cavity, which may upstage the disease and alter the prognosis. In the present study, hysterectomy was performed by different surgeons at different sites. We obtained peritoneal washings in 4 women during hysterectomy (Table 1, Nos. 3 and 10; Table 2, Nos. 4 and 5). Although none had a previous tubal occlusion, the peritoneal washings were negative for malignancy. Others, however, have reported retrograde transfallopian tube dissemination of malignant cells during hysteroscopy,12–16 and sonohysterography.17 The latter study reported on 32 patients with endometrial carcinoma who underwent sonohysterography at the time of abdominal hysterectomy. The researchers collected fluid from the fallopian tubes and identified cancer cells in 2 patients (6.25%) and suspicious cells in 6 (18.75%). They concluded that sonohysterography should not be performed in women with suspicious diagnosis of endometrial carcinoma. A case of peritoneal seeding was reported after D&C at hysteroscopic diagnosis of cervical adenocarcinoma.18 In another case, peritoneal and hepatic metastases occurred 5 months after operative hysteroscopy

Vilos et al

Hysteroscopy and uterine malignancy

of an unsuspected mixed mullerian tumor (carcinosarcoma) of the uterus.19 A prospective, randomized, crossover study concluded that transfallopian tube dissemination of benign endometrial cells occurs in 25% of women, regardless of whether saline or CO2 is used to distend the uterus.20 However, in a retrospective review of 284 women with endometrial cancer diagnosed by endometrial biopsy (n ⫽ 173) or D&C (n ⫽ 111), the frequency of positive peritoneal washings at the time of hysterectomy for those who had hysteroscopy and those who did not were 9.6% and 8.3%, respectively (p ⫽ .85).21 The value of peritoneal cytology as an independent adverse prognostic factor was retrospectively evaluated in 381 women with endometrial cancer.22 Positive cytology was found in 24 women (6.3%). Five-year survival was significantly lower for those women with positive versus negative cytology (50% vs 81.2%). However for women with stage I disease, there was no significant difference (80% vs 86.3%).22 In another study of 697 women, disease recurred in 10.5% (64 out of 611) of women with negative peritoneal cytology versus 31.2% (25 out of 80) with positive cytology. However, in 17 patients in the latter group, recurrence developed outside the peritoneal cavity.23 A study of 269 women with stage I or II disease concluded that if the disease was confined to the uterus, positive peritoneal cytology did not affect survival. If the disease had spread to the adnexa, lymph nodes, or peritoneum, then positive peritoneal cytology decreased the 5-year survival from 73% to 13%.24 A review of patients with clinical stage I endometrial cancer also concluded that poor prognosis associated with positive peritoneal washings was largely a reflection of other adverse prognostic factors, such as deep myometrial invasion and high histologic grade.25 Another study of 534 patients with endometrial cancer concluded that positive peritoneal cytology was not an independent risk factor. Instead, positive peritoneal cytology potentiated other prognostic indicators, and it was recommended that such patients should not be upstaged in the absence of other prognostic factors.26 The same group of investigators suggested that endometrial cancer cells found in the peritoneal cavity have low malignant potential and usually disappear within a short period of time.27 A casecontrol study of 27 women with endometrial cancer, followed for 2 years after hysterectomy, concluded that preceding diagnostic hysteroscopy did not increase the risk of early recurrence when the disease was confined to the uterus. The frequency of positive peritoneal cytology at hysterectomy in the study group (10%) was comparable with that in women with endometrial cancer who had not had fluid hysteroscopy.28 The second concern raised by oncologists is dissemination of malignant cells into deeper myometrial tissue or even direct infusion into vascular and lymphatic channels, which might facilitate distant metastases. Every resectoscopic surgeon has witnessed large vascular channels

209 created by the use of loop electrodes resulting in rapid intravasation of irrigating fluid, gas bubbles, and even intrauterine tissue. These observations would suggest that intravasation of malignant cells may occur more often than we think. However, the significance of this event and the fate of the intravasated cells remain speculative. One can only infer from published observations that the likelihood of such cells causing distant metastases is negligible. Studies on the fate of a variety of cancer cells, including breast, after they were injected into the circulation of animals, indicated that few isolated cancer cells detected in secondary sites remain viable and grow, and not all micrometastases that began growing persisted to form macrometastases.29 –31 Furthermore, it is possible that the intense radiofrequency-energy effect on tumor cells and surrounding tissue, including vessels, further reduces the viability of malignant cells and their potential for intravasation and subsequent metastases. The third concern raised by pathologists and oncologists is distortion of the resected tissue and the inability of proper orientation by the pathologist to accurately determine the depth of myometrial tumor invasion and stage of the disease. In our hospital, we have addressed this issue by resecting the endomyometrium in long sequential strips of uniform depth that can be oriented. Using this technique in the patient who refused hysterectomy (Table 1, No. 9), we were able to take additional strips of myometrium, deeper and around the tumor, after the tumor was resected and determined that the margins were clear of disease. Although the concerns listed above are legitimate based on speculation and theories, our present study clearly indicates that hysteroscopic surgery did not result in distant metastases because the 5-year prognosis for these 14 women was not affected.

Conclusion Endometrial cancer is occasionally found in women undergoing hysteroscopic endometrial ablation for AUB. Routinely performing endomyometrial resection, rather than rollerball ablation, in women at high risk for neoplasia may be the ultimate diagnostic tool for detecting uterine malignancies. Resectoscopic surgery does not appear to adversely affect the 5-year survival and prognosis of women with endometrial cancer.

References 1. Hallberg L, Hogdahl AM, Nilsson L, Rybo G. Menstrual blood loss: a population study—variation at different ages and attempts to define normality. Acta Obstet Gynecol Scand. 1966;45:320 –351. 2. Shapley M, Jordan K, Croft PR. An epidemiological survey of symptoms of menstrual loss in the community. Br J Gen Pract. 2004;54: 359 –363. 3. Rose PG. Endometrial carcinoma. N Engl J Med. 1996;335:640 – 649.

210

Journal of Minimally Invasive Gynecology, Vol 14, No 2, March/April 2007

4. Farquhar CM, Lethaby A, Sowter M, Verry J, Baranyai J. An evaluation of risk factors for endometrial hyperplasia in premenopausal women with abnormal menstrual bleeding. Am J Obstet Gynecol. 1999;181:525–529. 5. Orr JJ, Chamberlain D, Kilgone L, Naumann W. Management of endometrial cancer. American College of Obstetricians and Gynecologists (ACOG) Practice Bulletins. Number 65. August 2005. Obstet Gynecol. 2005;106:413– 425. 6. Spencer CP, Whitehead MI. Endometrial assessment re-visited (Review). Br J Obstet Gynaecol. 1999;106:623– 632. 7. Vilos GA. Hysteroscopic and nonhysteroscopic endometrial ablation. Obstet Gynecol Clin North Am. 2004;31:687–704, xi. 8. Vilos GA, Lefebvre G, Graves GR. Guidelines for the management of abnormal uterine bleeding. Society of Obstetrics and Gynaecology of Canada (SOGC) Clinical Practice Guidelines. J Obstet Gynaecol Can. 2001;23:704 –709. 9. Vilos GA, Harding PG, Silcox JA, Sugimoto AK, Carey M, Ettler HC. Endometrial adenocarcinoma encountered at the time of hysteroscopic endometrial ablation. J Am Assoc Gynecol Laparosc. 2002;9:40 – 48. 10. Vilos AG, Ettler HC, Edris F, Hollett-Caines J, Abu-Rafea B. Endometrioid adenocarcinoma treated by hysteroscopic endomyometrial resection: a case report. J Minim Invasive Gynecol. 2007;14:119 –122. 11. Brand A, Dubuc-Lissoir J, Ehlen TG, Plante M. Diagnosis of endometrial cancer in women with abnormal vaginal bleeding: SOGC clinical practice guidelines. J Soc Obstet Gynaecol Can. 2000;22: 102–104. 12. Romano S, Shimoni Y, Muralee D, Shalev E. Retrograde seeding of endometrial carcinoma during hysteroscopy. Gynecol Oncol. 1992;44: 116 –118. 13. Schmitz MJ, Nahhas WA. Hysteroscopy may transport malignant cells into the peritoneal cavity: case report. Eur J Gynaecol Oncol. 1994; 15:121–124. 14. Egarter C, Krestan C, Kurz C. Abdominal dissemination of malignant cells with hysteroscopy. Gynecol Oncol. 1996;63:143–144. 15. Zerbe MJ, Zhang J, Bristow RE, Grumbine FC, Abularach S, Montz FJ. Retrograde seeding of malignant cells during hysteroscopy in presumed early endometrial cancer. Gynecol Oncol. 2000;79:55–58. 16. Obermair A, Geramou M, Gucer F, et al. Does hysteroscopy facilitate tumor cell dissemination? Incidence of peritoneal cytology from patients with early stage endometrial carcinoma following dilatation and curettage (D & C) versus hysteroscopy and D & C. Cancer. 2000;88: 139 –143. 17. Dessole S, Rubattu G, Farina M, et al. Risks and usefulness of sonohysterography in patients with endometrial carcinoma. Am J Obstet Gynecol. 2006;194:362–368.

18. Wilson A, Duncan A. Peritoneal seeding after hysteroscopic diagnosis of cervical adenocarcinoma. Gynaecol Endosc. 1999;8:179 –181. 19. Benifla JL, Filippini F, Darai E, et al. Operative hysteroscopy procedure on an unsuspected mixed mullerian tumor of the uterus. Gynaecol Endosc. 1997;6:147–149. 20. Nagele F, Wieser F, Deery A, Hart R, Magos A. Endometrial cell dissemination at diagnostic hysteroscopy: a prospective randomized cross-over comparison of normal saline and carbon dioxide uterine distension. Hum Reprod. 1999;14:2739 –2742. 21. Gu M, Shi W, Huang J, Barakat RR, Thaler HT, Saigo PE. Association between initial diagnostic procedure and hysteroscopy and abnormal peritoneal washings in patients with endometrial carcinoma. Cancer. 2000;90:143–147. 22. Grimshaw RN, Tupper WC, Fraser RC, Tompkins MG, Jeffrey JF. Prognostic value of peritoneal cytology in endometrial carcinoma. Gynecol Oncol. 1990;36:97–100. 23. Morrow CP, Bundy BN, Kurman RJ, et al. Relationship between surgical-pathological risk factors and outcome in clinical stage I and II carcinoma of the endometrium: a Gynecologic Oncology Group study. Gynecol Oncol. 1991;40:55– 65. 24. Kadar N, Homesley HD, Malfetano JH. Positive peritoneal cytology is an adverse factor in endometrial carcinoma only if there is other evidence of extrauterine disease. Gynecol Oncol. 1992;46:145–149. 25. Milosevic MF, Dembo AJ, Thomas GM. The clinical significance of malignant peritoneal cytology in stage I endometrial carcinoma. Int J Gynecol Cancer. 1992;2:225–235. 26. Takeshima N, Nishida H, Tabata T, Hirai Y, Hasumi K. Positive peritoneal cytology in endometrial cancer: enhancement of other prognostic indicators. Gynecol Oncol. 2001;82:470 – 473. 27. Hirai Y, Takeshima N, Kato T, Hasumi K. Malignant potential of positive peritoneal cytology in endometrial cancer. Obstet Gynecol. 2001;97:725–728. 28. Gucer F, Tamussine K, Moser F, et al. Two-year followup of patients with endometrial carcinoma after preoperative fluid hysteroscopy. Int J Gynecol Cancer. 1998;8:476 – 480. 29. Swartz MA, Kristensen CA, Melder RJ, et al. Cells shed from tumours show reduced clonogenicity, resistance to apoptosis, and in vivo tumorigenicity. Br J Cancer. 1999;81:756 –759. 30. Chambers AF, Naumov GN, Vantyghen SA, et al. Breast cancer metastasis: clinical implications of experimental studies on metastatic inefficiency. Breast Cancer Res. 2000;2:400 – 407. 31. Wyckoff JB, Jones JG, Condeelis JS, Segall JE. A critical step in metastasis: in vivo analysis of intravasation at the primary tumor. Cancer Res. 2000;60:2504 –2511.