139 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 Andreas Obermair, M.D.1 Magda Geramou, M.D.1 Fatih Gucer, M.D.2 Ursula Denison, M.D.3 Anton H. Graf, M.D., Ph.D.4 Elisabeth Kapshammer, M.D.5 Walter Neunteufel, M.D.6 Inge Frech, M.D.7 Alexandra Kaider, M.Sc.8 Christian Kainz, M.D.1 for the Austrian Gynecologic Oncology Study Group 1 Division of Gynecology, Department of Gynecology and Obstetrics, University Hospital of Vienna, Vienna, Austria. 2 Department of Gynecology and Obstetrics, University Hospital of Graz, Graz, Austria. 3 Department of Gynecology and Obstetrics, General Hopital Wilhelminenspital, Vienna, Austria. 4 Department of Gynecology and Obstetrics, General Hospital Salzburg, Salzburg, Austria. 5 Department of Gynecology, Hospital Barmherzige Schwestern, Linz, Austria. 6 Department of Gynecology and Obstetrics, General Hospital Dornbirn, Dornbirn, Austria. 7 Department of Gynecology and Obstetrics, Kaiser Franz Josef Hospital, Vienna, Austria. BACKGROUND. In several case reports, distension and irrigation of the uterine cavity during fluid hysteroscopy was suspected to cause tumor cell dissemination into the abdominal cavity in patients with endometrial carcinoma. It was the aim of this study to compare the incidence of positive peritoneal cytology in patients who underwent dilatation and curettage (D & C) with or without previous hysteroscopy. METHODS. The authors conducted a multicentric, retrospective cohort analysis. One hundred thirteen consecutive patients with endometrial carcinoma treated between 1996 and 1997 were included. Endometrial carcinoma had to be limited to the inner half or less than the inner half of the myometrium (pathologic Stage IA,B). Positive peritoneal cytology was obtained during staging laparotomy. Patients underwent D & C either with or without prior diagnostic fluid hysteroscopy. No selection or randomization was applied to the two groups. Positive peritoneal cytology, defined as malignant or suspicious, was considered the primary statistical endpoint. RESULTS. Peritoneal cytology was suspicious or positive in 10 of 113 patients (9%). The presence of suspicious or positive peritoneal cytology was associated with a history of hysteroscopy (P 5 0.04) but not with myometrial invasion (P 5 0.57), histologic subtype (P 5 1.00) or grade (r 5 0.16, P 5 0.10), or the time between D & C and staging laparotomy (r 5 0.04, P 5 0.66). CONCLUSIONS. Based on the limited extent of endometrial carcinoma in the current analysis, our data strongly suggest dissemination of endometrial carcinoma cells after fluid hysteroscopy. Determining whether a positive peritoneal cytology affects the prognoses of patients without further evidence of extrauterine disease will require longer follow-up. Cancer 2000;88:139 – 43. © 2000 American Cancer Society. KEYWORDS: endometrial carcinoma, peritoneal cytology, tumor cell dissemination, hysteroscopy, dilatation and curettage. 8 Department of Computer Sciences/Biometrics, University Hospital of Vienna, Vienna, Austria. The authors acknowledge the contributions of the following institutions and pathologists/cytologists: Department of Clinical Pathology/Gynecopathology, University Hospital of Vienna, Vienna, Austria (G. Breitenecker, M.D.); Department of Gynecology and Obstetrics, University Hospital of Graz, Graz, Austria (H. Pickel, M.D.); Department of Clinical Pathology, General Hospital Wilhelminenspital Vienna, Vienna, Austria (K. Kofler, M.D., and G. Rona-Selnic, M.D.); Department of Clinical Pathology, General Hospital © 2000 American Cancer Society Salzburg, Salzburg, Austria (O. Dietze, M.D.); Hospital Barmherzige Schwestern Linz, Linz, Austria (W. Sega, M.D.); Department of Clinical Pathology, General Hospital Dornbirn, Dornbirn, Austria (G. Breitfellner, M.D.); Department of Clinical Pathology, Kaiser Franz Josef Hospital, Vienna, Austria (M. Klimpfinger, M.D.). Address for reprints: Andreas Obermair, M.D., Department of Gynecology and Obstetrics, University Hospital of Vienna, Waehringer Guertel 18-20; A-1090 Vienna, Austria. Received April 8, 1999; revision received August 2, 1999; accepted August 2, 1999. 140 CANCER January 1, 2000 / Volume 88 / Number 1 P atients with endometrial carcinoma commonly present with abnormal uterine bleeding as a leading symptom.1,2 Endometrial carcinoma is usually diagnosed as an organ-confined disease, with 5-year disease free survival rates of 90% and 80% for patients with Stage IA and B disease, respectively.3 Although there is some debate regarding whether a positive peritoneal cytology worsens the prognosis in the absence of further extrauterine disease, patients with organ-confined endometrial carcinoma who present with malignant cells in the peritoneal washing have to be classified as Stage IIIA according to the current International Federation of Gynecology and Obstretrics (FIGO) staging classification.4 In order to improve the sensitivity of dilatation and curettage (D & C), hysteroscopy has been established as a feasible and widely accepted procedure for the assessment of abnormal uterine bleeding within the last decade.5–7 Though it is a common practice, the role of hysteroscopy in diagnosing endometrial carcinoma is not yet established. An increasing number of case reports assume abdominal dissemination of malignant cells during hysteroscopy in patients with endometrial carcinoma.8 –12 Until now, systematic investigation of tumor cell dissemination by hysteroscopy has failed because of ethical reservations. Several authors have expressed the concern that distension and irrigation of the uterine cavity with fluid media may increase intrauterine pressure and thereby facilitate the dispersion of malignant cells through the fallopian tubes into the abdominal cavity. However, it has not been clarified whether tumor cell dissemination is more frequent when D & C is combined with hysteroscopy compared with D & C alone. A search of the WINSPIRS MEDLINE database (1985–1999) found no study that compared the incidence of positive peritoneal cytology in endometrial carcinoma patients with or without previous hysteroscopy. Therefore, the aim of the current study was to evaluate the incidence of positive peritoneal washings in patients with early stage endometrial carcinoma in the inner half or less than the inner half of the myometrium, and to compare these findings for patients who underwent hysteroscopy plus D & C with those for patients who only had D & C prior to surgical staging. MATERIALS AND METHODS Patients The authors conducted a multicentric, retrospective chart review of all patients who underwent D & C with or without prior diagnostic hysteroscopy as a single procedure for abnormal uterine bleeding. Data for this analysis were collected from 113 patients who had undergone surgery for endometrial carcinoma between 1996 and 1997 at 7 Austrian hospitals. At all participating institutions, patients with abnormal uterine bleeding were diagnosed by D & C as well as by means of a diagnostic hysteroscopy. However, due to organizational circumstances (scheduling, equipment problems), hysteroscopy was not performed in 39 of 113 cases. Patients were included if they had histologically proven endometrial carcinoma invading half or less than half of the myometrium and evaluable peritoneal cytology. World Health Organization (WHO) criteria were used for histologic classification. Surgery records and histologic findings were used to determine the stage of disease according to the FIGO 1988 criteria.4 Treatment All procedures were carried out under general anesthesia. The cervix was dilated sufficiently to Hegar size 6 in order to allow insertion of a standard 4-mm telescope with a 5-mm diagnostic sheath. The uterine cavity was generally distended with normal saline up to a pressure of 150 mmHg, in order to allow full inspection. All pathologic specimens were obtained with a blind D & C, which was carried out after hysteroscopy. In cases of histologically verified endometrial carcinoma, the standard surgical procedure consisted of a total abdominal hysterectomy (TAH) with bilateral oophorectomy (BSO). In patients with myometrial invasion of up to one-half of the inner myometrium, we do not perform lymph node dissection on a routine basis, because positive pelvic lymph nodes are present in 3– 6% of the patients.13 Nevertheless, in 3 patients with pathologic Stage IA and 18 patients with pathologic Stage IB, a pelvic lymph node sampling was performed. Peritoneal cytology was obtained before hysterectomy by peritoneal washing of the pouch of Douglas, using 100 –200 mL of sterile saline. After centrifugation at 3000 g for 10 minutes, the pellets were stained with May-Grünwald-Giemsa and Papanicolaou stains. Peritoneal cytology was considered positive or suspicious regardless of the degree of positive or suspicious cytology. Statistical Methods Peritoneal cytology was considered the primary statistical endpoint of this analysis. Frequency distributions of ordered categoric variables were compared by means of exact Wilcoxon rank sum tests. Correlation between dichotomous variables were tested using Fisher exact tests, and correlations between continous or ordered categoric variables were estimated by the Spearman rank correlation coefficient. Reported P values are two-sided and a P value , 0.05 was considered Hysteroscopy and Endometrial Carcinoma/Obermair et al. TABLE 1 Patients Characteristics Peritoneal cytology Negative Suspicious Malignant Myometrial invasion None ,50% Histologic subtype Endometrioid Others Grading G1 G2 G3 Time between D & C and TAH/BSO (days) (Median; quartiles) All patients HSK 1 D & C n 5 74 D&C n 5 39 P value 103 5 5 65 (87.8%) 4 (5.4%) 5 (6.8%) 38 (97.5%) 1 (2.5%) 0 0.046a 20 93 12 (16.2%) 62 (83.8%) 8 (20.5%) 31 (79.5%) 0.609b 102 11 67 (90.5%) 7 (9.4%) 35 (89.7%) 4 (10.3%) 1.000c 56 41 11 37 (52.8%) 27 (38.6%) 6 (8.6%) 19 (50.0%) 14 (36.8%) 5 (13.2%) 0.666c 16 (8,27) 13 (8,27) 20 (11,26) 0.146 HSK: hysteroscopy; D & C: dilatation and curettage. a Exact Wilcoxon rank sum test (one-sided). b Exact Wilcoxon rank sum test (two-sided). c Fisher exact test. significant. A one-sided Wilcoxon rank sum test was performed to compare peritoneal cytology with the application of hysteroscopy, because there is no rationale for increased tumor cell dissemination by D & C alone compared with hysteroscopy followed by D & C. RESULTS Analysis was based on the data of 113 patients with endometrial carcinoma who had D & C prior to TAH/ BSO. Seventy-four patients underwent a hysteroscopy before D & C, and in 39 patients D & C was performed without prior hysteroscopy. The mean age of the patients was 64 years (range, 43– 86 years). Twenty patients were regarded as Stage IA (tumor confined to the endometrium) and 93 as Stage IB (tumor confined to the inner half of the myometrium). Overall, we found peritoneal cytology to be suspicious or positive in 10 of 113 cases (8.8%). Patient characteristics showed that the patients were similar with regard to pathologic stage, histologic type and grade, and the time period between D & C and TAH/BSO (Table 1). The presence of suspicious or carcinoma cells within the peritoneal fluid was associated with a history of hysteroscopy (P 5 0.04) but not with myometrial invasion (P 5 50.57), histologic subtype (P 5 1.00), grade (r 5 0.16, P 5 0.10), or the time between D & C and staging laparotomy (r5 0.04, P 5 0.66). 141 Strong evidence for tumor cell dissemination was also suspected in 1 case of a patient age 65 years with an undifferentiated Stage IA endometrioid-type adenocarcinoma. During staging laparotomy, which was performed 16 days (quartiles: 8, 27) after hysteroscopy and D & C, we found not only positive peritoneal cytology but also a single islet of carcinomatous tissue on the surface of the uterine corpus. No other cases with positive peritoneal cytology presented with visible carcinomatous tissue within the abdomen. The clinical and histopathologic features of all patients with positive peritoneal cytology are given in Table 2. DISCUSSION These data strongly suggest hysteroscopic tumor cell dissemination of endometrial carcinoma cells in the course of fluid hysteroscopy. In a retrospective chart analysis of patients with endometrial carcinoma invading half or less than half of the myometrium (pathologic Stage IA,B), we found positive peritoneal washings to be significantly more frequent for patients who underwent hysteroscopy before surgical staging than for those who underwent D & C alone. Five of 113 patients with pathologic Stage IA,B endometrial carcinoma had suspicious cytology and an additional 5 patients had positive cytology. Four of the five patients with suspicious cytology had had hysteroscopy prior to surgical staging, whereas only one of those patients was diagnosed by D & C alone. All of the five patients with positive peritoneal cytology had had hysteroscopy prior to staging laparotomy. Analysis was restricted to patients with involvement of the inner half of the myometrium, because the incidence of positive peritoneal cytology is much higher in patients with involvement of the outer half of the myometrium.14 In patients in whom cancer was restricted to the inner half of the myometrium, the finding of positive peritoneal cytology would therefore be much more indicative of tumor cell dissemination, compared with patients with involvement of the outer half of the myometrium. Due to the multicentric design of this study, the position of the tumor within the uterine cavity and the presence of vascular space involvement were not evaluated. Although diagnostic hysteroscopy is widely used for the diagnosis of endometrial hyperplasia or carcinoma, a clear statement of acceptance from a medical point of view has not been issued regarding tumor cell dissemination into the peritoneal cavity in association with endometrial carcinoma. The authors of various case reports discussed an association between hysteroscopy and abdominal dissemination of malignant cells into the peritoneal cavity. Romano et al.8 reported a case of clinical Stage IA, 142 CANCER January 1, 2000 / Volume 88 / Number 1 TABLE 2 Clinical and Histologic Features of Patients with Positive Peritoneal Cytology Patient Initials Age (yrs) HSK Cytology Myometrial invasion Histologic type Grading Time betw D & C 6 H and laparoto (days) 1 2 3 4 5 6 7 8 9 10 SC RO BA HO DO WU DE KA BR BE 70 56 60 61 78 67 77 46 72 75 Y Y Y Y Y Y Y Y Y N S S M S M M M M S S ,50% NONE ,50% ,50% ,50% ,50% ,50% ,50% ,50% ,50% MMT END END END END END END END END END 1 2 1 2 3 2 2 2 2 24 35 23 3 13 19 13 19 10 29 HSK: hysteroscopy; D & C: dilatation and curettage; Y: yes; N: no; S: suspicious; M: malignant; MMT: malignant mixed mesodermal tumor (Müller); END: endometrioid. Grade 2 endometrial adenocarcinoma diagnosed by hysteroscopy and endometrial biopsy. Positive cytology was found during surgical staging. Schmitz and Nahhas9 reported a case of a patient age 63 years who underwent hysteroscopy followed by D & C. Three weeks later, malignant cells were found in pelvic washings during laparotomy. Another case with Stage IA Grade 1 adenocarcinoma of the endometrium has been reported from our hospital.10 A hysteroscopy was performed after peritoneal cytology had been obtained by peritoneal washing with saline. In contrast to the first washing, a second peritoneal lavage that was carried out immediately after hysteroscopy demonstrated positive cytology. Rose et al.12 reported a case of a patient age 39 years with endometrial carcinoma that had been resected by hysteroscopy and treated conservatively. Leveque et al.15 found a relatively high incidence of positive peritoneal cytology (37%) in 19 patients with Stage IA–C endometrial carcinoma; they assumed that hysteroscopy, which had been carried out in all 19 cases preoperatively, might have been the reason for this. All these authors speculate that saline irrigation of the uterine cavity might cause dissemination of endometrial carcinoma cells through the fallopian tubes into the peritoneal cavity. Although transtubal spill seems likely, it is not fully clarified whether other factors might play an additional role. Recently, Baker and Adamson determined that 100 mmHg was the median intrauterine perfusion pressure for spill of dye from both tubes; no spill occurred at pressures ,70 mmHg.16 Although endometrial cells could not be found before hysteroscopy, the incidence of endometrial cells in the peritoneal fluid was 65% after chromopertubation. Creasman et al.17 suggested that spill might occur at a different anatomic site than the fallopian tube, as malignant cells were discovered in the peritoneal cavity in 3 of 44 patients with endometrial carcinoma. One of these three patients had had a previous bilateral salpingectomy and two patients a tubal ligation before hysterectomy.17 From all these reports, it has not been proven that tumor cell dissemination is more common in cases in which D & C plus hysteroscopy is performed compared with cases in which D & C alone is performed. Whether positive peritoneal cytology affects the prognosis of patients with Stage I endometrial carcinoma is currently a subject of animated scientific discussion. Although some authors14,18 –20 have shown that increasing concentrations of carcinoma cells in peritoneal washings significantly shorten the time to recurrence of disease, others have suggested that positive peritoneal cytology is prognostically irrelevant unless extrauterine disease is present.21–23 Gucer et al.24 report that previous hysteroscopy did not seem to increase the risk of early recurrence in 25 patients with and 55 patients without fluid hysteroscopy. In both groups, only 1 recurrence could be observed after median follow-up periods of 29 and 30 months, respectively. In the current investigation, we chose the presence of malignant cells as the primary statistical endpoint because it represented “advanced disease” according to the current FIGO classification.4 For this reason, and because of the short observation period, the question of whether hysteroscopy affects the patients’ prognosis has not been addressed in this study but will be a topic of further investigation. In response to case reports and also our own observations, we intended to compare the rate of positive peritoneal washings of patients with hysteroscopy before laparotomy with those of patients who did not have hysteroscopy. Therefore, we initiated a retrospective chart analysis with peritoneal cytology as the statistical endpoint. All hysteroscopies were per- Hysteroscopy and Endometrial Carcinoma/Obermair et al. formed with saline irrigation. ITo avoid positive results due to involvement of the uterine serosa, only cases with invasion of the inner half (or less than the inner half) of the myometrium were selected. Patients with involvement of the outer half of the myometrium, as well as those with invasion of the uterine cervix or the fallopian tube, were excluded. In summary, our data strongly suggest hysteroscopic tumor cell dissemination of carcinoma cells via fluid hysteroscopy. We found positive peritoneal cytology significantly more often for patients who underwent hysteroscopy before surgical staging than for those who did not undergo hysteroscopy but only D & C. An association between hysteroscopy and dispersion of tumor cells is therefore highly possible. Established prognostic factors, such as depth of myometrial invasion, histologic subtype, grading, and the time between D & C and TAH/BSO, were similar for patients with and without previous hysteroscopy. Nevertheless, the retrospective design of the study did not permit a conclusive statement in this regard. For ethical reasons, it is difficult to tackle this question from a prospective approach. Prospectively, it is suggested that investigators pay close attention to peritoneal cytology findings. Moreover, it would be of considerable interest to collect a much larger sample size to verify the interesting phenomenon that we have reported in this article, so that final conclusions may be drawn. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. Hacker NF. Uterine cancer. In: Hacker NF, Bereck JS, editors. Practical gynecologic oncology. 2nd edition. Baltimore, MD: Williams and Wilkins; 1994:285–326. Barakat RR, Park RC, Grigsby PW, Muss HD, Norris HJ. Corpus: epithelial tumors. In: Hoskins WJ, Perez CA, Young RC, editors. Principles and practice of gynecologic oncology. 2nd edition. Philadelphia: Lippincott-Raven; 1997:859 –96. Wingo PA, Ries LA, Rosenberg HM, Miller DS, Edwards BK. Cancer incidence and mortality, 1973–1995: a report card for the U.S. Cancer 1998;82:1197–1207. International Federation of Gynecology and Obstetrics. Corpus cancer staging. Int J Gynecol Obstet 1989;28:190. Nagele F, O’Connor H, Davies A, Badawy A, Mohamed H, Magos A. 2500 outpatient diagnostic hysteroscopies. Obstet Gynecol 1996;88:87–92. Schwärzler P, Concin H, Bösch H, et al. An evaluation of sonohysterography and diagnostic hysteroscopy for the assessment of intrauterine pathology. Ultrasound Obstet Gynecol 1998;11:337– 42. Toki T, Oka K, Nakayama K, Oguchi O, Fujii S. A comparative study of preoperative procedures to assess cervical invasion by endometrial carcinoma. Br J Obstet Gynecol 1998; 105:512– 6. Romano S, Shimoni Y, Muralee D, Shalev E. Retrograde seeding of endometrial carcinoma during hysteroscopy. Gynecol Oncol 1992;44:116 – 8. 20. 21. 22. 23. 24. 143 Schmitz MJ, Nahhas WA. Hysteroscopy may transport malignant cells into the peritoneal cavity . Eur J Gynecol Oncol 1994;2:121– 4. Egarter C, Krestan C, Kurz C. Abdominal dissemination of malignant cells with hysteroscopy. Case Report. Gynecol Oncol 1996;63:143– 4. Bettocchi S, Di Vagno G, Cormio G, Selvaggi L. Intraabdominal spread of malignant cells following hysteroscopy. Gynecol Oncol 1997;66:165– 6. Rose PG, Mendelsohn G, Kornbluth I. Hysteroscopic dissemination of endometrial cancer. Gynecol Oncol 1998;71: 145– 6. Creasman WT, Morrow CP, Bundy BN, Homesley HD, Graham JE, Heller PB. Surgical pathologic spread patterns of endometrial cancer. Cancer 1987;60:2035– 41. Kashimura M, Sugihara K, Toki N, Matsuura Y, Kawagoe T, Kamura T, et al. The significance of peritoneal cytology in uterine cervix and endometrial cancer. Gynecol Oncol 1997; 67:285–90. Leveque J, Goyat F, Dugast J, Loiellet L, Grall JY, Le Bars S. Value of peritoneal cytology after hysteroscopy in surgical stage I adenocarcinoma of the endometrium. Oncol Rep 1998;5:713–5. Baker VL, Adamson GD. Threshold intrauterine perfusion pressures for intraperitoneal spill during hydrotubation and correlation with tubal adhesive disease. Fertil Steril 1995;64: 1066 –9. Creasman WT, Lukeman J. Role of the Fallopian Tube in dissemination of malignant cells in corpus cancer. Cancer 1972;29:456 –7. Creasman WT, Rutledge F. The prognostic value of peritoneal cytology in gynecologic malignant disease. Am J Obstet Gynecol 1971;110:773– 81. Creasman WT, Disaia PJ, Blessing J, Wilkinson RH, Johnston W, Weed JC. Prognostic significance of peritoneal cytology in patients with endometrial cancer: preliminary data concerning therapy with intraperitoneal radiopharmaceuticals. Am J Obstet Gynecol 1981;141:921–9. Morrow CP, Bundy BN, Kurmann 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. 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;46:145–9. Kennedy AW, Webster KD, Nunez C, Bauer LJ. Pelvic washings for cytologic analysis in endometrial adenocarcinoma. J Reprod Med 1993;38:637– 42. Eltabbakh GH, Piver S, Hempling RE, Shin KH. Excellent long-term survival and absence of vaginal recurrence in 332 patients with low-risk stage I endometrial adenocarcinoma treated with hysterectomy and vaginal brachytherapy without formal staging lymph node sampling: report of a prospective trial. Int J Radiat Oncol Biol Phys 1997;38:373– 80. Gucer F, Tamussino K, Reich O, Moser F, Arikan G, Winter R. Two-year follow-up of patients with endometrial carcinoma after preoperative fluid hysteroscopy. Int J Gynecol Cancer 1998;8:476 – 80.