ANTICANCER RESEARCH 31: 243-248 (2011)

Detection and Significance of Parametrial Micrometastases in Early-stage Cervical Cancer

PAULA C. BRADY1, XIAOWEI CHEN2,4, WILLIAM M. BURKE1, ISRAEL DEUTSCH3,4, XUMING SUN1, THOMAS J. HERZOG1,4 and JASON D. WRIGHT1,4

1Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, 2Department of Pathology, 3Department of Radiation Oncology, Columbia University College of Physicians and Surgeons, New York, NY 10032, U.S.A.; 4Herbert Irving Comprehensive Cancer Center, New York, NY 10032, U.S.A.

Abstract. Background: Parametrial metastases are rare in and parametrium (29). While associated with excellent women with small, node-negative cervical tumors. We examined oncologic outcomes, the operation is also accompanied by the incidence of micrometastases in women with pathologically significant morbidity (15, 19, 21, 28, 30). Nearly 40% of negative parametria. Patients and Methods: Patients with IA-IB patients who underwent radical in a prospective cervical cancer who underwent radical hysterectomy and had trial had late morbidity associated with the procedure, no evidence of parametrial metastases were examined. including urologic complications in a third of patients (19). Immunohistochemistry (IHC) using a cytokeratin antibody was Much of the morbidity associated with the procedure stems performed in parametrial sections. Results: Among 46 patients, from removal of the parametrial soft tissue (11). The a parametrial micrometastasis was identified in 1 (2.2%). autonomic nervous fibers, which play a critical role in normal Micrometastases were seen in 5% of women with lymphvascular voiding and sexual function, course through the parametrium space invasion (LVSI) and in 0 (0%) patients without LVSI and are susceptible to damage during parametrial resection (p=0.43). The patient with a parametrial micrometastasis performed at the time of radical hysterectomy. experienced recurrence and died from her disease. In a previous The goal of parametrectomy is the resection of occult report, we defined patients with tumors <2 cm, negative pelvic parametrial disease. However, the utility of radical lymph nodes, and no LVSI as being at low-risk for parametrial parametrial resection has been questioned by a number of disease. There were no micrometastases in this group of women investigators (10, 22). The risk of parametrial metastasis in the current series. Conclusion: Parametrial micrometastases appears to be small in women with early, small volume are uncommon in early-stage cervical cancer. cervical tumors with favorable pathologic features (6, 16, 33). While parametrial metastases are rare in some In 2008, it is estimated that 11,070 women were diagnosed subgroups of patients, the risk of occult micrometastasis is with cervical cancer and that 3,870 died from their disease poorly understood. To date, relatively little work has been (13). Worldwide, nearly 500,000 cases of cervical cancer are performed to examine the frequency of parametrial diagnosed annually (27). Treatment for stage IA-IIA cervical micrometastases in women with early-stage cervical cancer. cancer typically consists of either radical hysterectomy or The goal of this study was to identify the frequency of radiation. A randomized clinical trial demonstrated parametrial micrometastases in women with early-stage equivalent survival for the two modalities (18). cervical who had undergone radical hysterectomy and had Type III Wertheim radical hysterectomy consists of en bloc no evidence of parametrial disease by standard pathologic resection of the , upper , uterosacral analysis.

Patients and Methods

Correspondence to: Jason D. Wright, MD, Division of Gynecologic Study approval was obtained from the Columbia University Oncology, Department of Obstetrics and Gynecology, Columbia Institutional Review Board. Patients with stage IA2-IIA cervical University College of Physicians and Surgeons, 161 Fort cancer who underwent type III radical hysterectomy as described by Washington Ave, 8th Floor, New York, NY 10032, U.S.A. Tel: +1 Piver et al. (29) between 1998 and 2007 were then examined. Only 2123053410, Fax: +1 2123053412, e-mail: [email protected] patients with negative parametria were included in the analysis. Demographic, clinical and pathologic data was abstracted from Key Words: Cervical cancer, parametrium, micrometastasis, medical records. Disease status and vital status were obtained from immunohistochemistry. a prospectively maintained hospital database.

0250-7005/2011 $2.00+.40 243 ANTICANCER RESEARCH 31: 243-248 (2011)

All radical hysterectomy specimens were collected and processed Microscopic parametrial spread was seen in 6% of women in a routine clinical manner. Briefly, at the time of hysterectomy the with adenocarcinomas, and in no patients with squamous cell parametrium and margins were inked for processing. The carcinomas (p=0.54). There was no apparent association parametrial tissue was removed and processed separately. For the between pelvic or para-aortic nodal spread and microscopic present study archival blocks of the parametrium were obtained and representative sections re-cut. Pathologic data reviewed included parametrial disease. histology, grade, depth of tumor invasion, lymphovascular space Recurrent or progressive disease was identified in 7% of invasion, vaginal and uterine involvement, pelvic and para-aortic patients without parametrial micrometastases and in the one lymph node status and status of the margins of excision. patient with microscopic parametrial involvement (p=0.17). Immunohistochemistry was performed on the re-cut parametrial At the last follow-up, 4% of the women with negative sections of each patient. Two 5 μm slides were examined for each immunohistochemistry and the one patient with parametrial patient. Standard immunohistochemistry was performed using a micrometastases had died from their tumors (p=0.10). mouse monoclonal anti-AE1/AE3 cytokeratin antibody at a dilution of 1:50 (Dako Corporation, Carpinteria, CA, USA). Each slide was examined by a gynecologic pathologist blinded to all patient data. Discussion A micrometastatic lesion was defined as a tumor implant between 0.2 mm and 2 mm. Our study demonstrates that parametrial micrometastases are Statistical analysis was performed using SAS version 9.2 (SAS uncommon in women with negative parametria by standard Institute, Cary, N.C., USA). Patients who were alive were censored pathologic analysis. We were able to identify micrometastases at the date of last contact. Overall survival was defined as the interval to the parametrium in only 2% of our cohort. Given the rarity from surgery to death from any cause, and disease-free survival was defined as the time interval from surgery to the first incidence of of parametrial micrometastases, it is difficult to determine the recurrence or death, whichever occurred first. Those patients alive prognostic significance of the finding. without recurrence were censored at the date of last contact. The The identification of micrometastatic tumor deposits by association between parametrial micrometastases and other immunohistochemistry has been successfully demonstrated for pathologic factors was evaluated using Fisher’s exact test. A p-value a number of solid tumors (1, 5, 7, 14, 17, 20, 23, 24, 26, 31). In less than 0.05 was taken to indicate statistical significance and all cervical cancer, the concept has been most widely applied to the tests were two-sided. detection of occult lymph nodal metastases in women without Results nodal disease by standard analysis (20, 31). Among 132 patients with early-stage cervical cancer, Lentz and colleagues found Forty-six women with pathologically negative parametria at micrometastases in 15% of the sentinel lymph nodes that were the time of radical hysterectomy were identified. The negative by standard hematoxylin and eosin analysis. The demographic and pathologic characteristics of the cohort are investigators proposed that this rate of micrometastases displayed in Table I. The median age of the patients was 42 approximated the recurrence rate of node-negative cervical years. Stage IB tumors were noted in 87% of patients, while cancer, but they were unable to demonstrate an association 13% had stage IA lesions. Squamous histology predominated between micrometastatic disease and tumor size or and the majority of tumors were <4 cm in diameter (65%). lymphvascular space invasion (20). In contrast, in a study of 49 Lymphvascular space invasion was seen in 43% of the early stage cervical cancer patients, Juretzka and co-workers tumors. Eleven percent of women had pelvic nodal found micrometastases in the pelvic lymph nodes of 8% of metastases, and 2% had para-aortic spread. patients, and correlated micrometastasis to lymphovascular Parametrial micrometastasis was identified in 1 (2%) space invasion and tumor size (14). Despite the documentation patient by immunohistochemistry. The micrometastasis was of nodal micrometastases, the prognostic significance of this detected in a lymphvascular channel of the parametrium. The finding is uncertain. Marchiole et al. found that in 292 patients patient with a parametrial micrometastasis had a 3 cm with early-stage cervical cancer, while immunohistochemistry adenocarcinoma with lymphvascular space invasion and no was not significantly more effective in determining evidence of lymph node metastases. After initial treatment micrometastasis to lymph nodes, the presence of these tumors the experienced recurrence and died of disease. was correlated to recurrence (23). A literature review examining The pathologic characteristics and outcomes between micrometastases in gynecologic cancers concluded that the patients with and without parametrial micrometastases were prognostic significance of microscopic metastases could not be then compared (Table II). None of the patients with stage IA determined from the body of literature currently available (7). tumors had parametrial micrometastases, while 3% of The present study focused specifically on the examination patients with IB tumors had microscopic parametrial disease of the parametrial soft tissue for occult micrometastases. A (p=0.87). Likewise, parametrial micrometastases were not recent report of 18 patients with negative parametria by identified in any patients without lymphvascular space standard histology noted parametrial micrometastases in 17% invasion but were found in 5% of those women with of patients. The researchers were unable to correlate lymphvascular space invasion in the (p=0.43). microscopic parametrial involvement with outcome (4). The

244 Brady et al: Parametrial Micrometastases in Early-stage Cervical Cancer

Table I. Demographic characteristics of the study cohort. Table II. Comparison of women patients with and without parametrial micrometastases. Patients (n=46) Parametrium Parametrium p-Value Age (years) negative positive Median 41.5 (n=45) (n=1) Race White 24 (52) Age (years) 0.24 Black 6 (13) Median 42 32 Hispanic 14 (30) Mean±SD 44.0±9.93 32±0 Other 2 (4) Race 0.17 Stage White 24 (100) 0 (–) IA 6 (13) Black 5 (83) 1 (17) IB 40 (87) Hispanic 14 (100) 0 (–) Lymphvascular space invasion Other 2 (100) 0 (–) Yes 20 (43) Stage 0.87 No 26 (57) IA 6 (100) 0 (–) Histology IB 39 (97) 1 (3) Squamous 21 (47) Lymphvascular Adenocarcinoma 17 (37) space invasion 0.43 Adenosquamous 8 (17) Yes 19 (95) 1 (5) Grade No 26 (100) 0 (–) 1 20 (43) Histology 0.54 2 17 (37) Squamous 21 (100) 0 (–) 3 7 (15) Adenocarcinoma 16 (94) 1 (6) Unknown 2 (4) Adenosquamous 8 (100) 0 (–) Pelvic nodes Grade 0.20 Positive 5 (11) 1 20 (100) 0 Negative 41 (89) 2 17 (100) 0 Paraaortic lymph nodes 3 6 (86) 1 (14) Positive 1 (2) Unknown 2 (100) 0 Negative 11 (24) Pelvic nodes 0.89 Not performed 34 (74) Positive 5 (100) 0 (–) Tumor size (cm) Negative 40 (98) 1 (2) <1 4 (9) Paraaortic lymph nodes 1.00 1-2 13 (28) Positive 1 (100) 0 (–) 2-4 13 (28) Negative 11 (100) 0 (–) >4 2 (4) Not performed 33 (97) 1 (3) Unknown 14 (30) Tumor size (cm) 0.70 Recurrence <1 4 (100) 0 (–) None 38 (83) 1-2 13 (100) 0 (–) Recurrent/progressive disease 4 (9) 2-4 12 (92) 1 (8) Unknown 4 (9) >4 2 (100) 0 (–) Vital status Unknown 14 (100) 0 (–) No evidence of disease 37 (81) Recurrence 0.17 Alive with disease 1 (2) None 38 (100) 0 Dead of disease 3 (7) Recurrent/progressive disease 3 (75) 1 (25) Dead of other causes 1 (2) Unknown 4 (100) 0 Unknown 4 (9) Vital status 0.11 No evidence of disease 37 (100) 0 Alive with disease 1 (100) 0 Dead of disease 2 (67) 1 (33) Dead of other causes 1 (100) 0 rate of microscopic parametrial infiltration was lower in our Unknown 4 (100) 0 study, only 2%. Similarly, however, we were unable to Follow up (months) 0.65 correlate parametrial micrometastatic disease with outcome Median 42 27 Mean±SD 42.7±33.9 27±0 or any specific pathologic characteristic. It is notable that in our series, microscopic parametrial spread was not documented in any patients who did not have lymphvascular space invasion within the cervix. Given that only one patient The necessity of parametrectomy for women with small, had parametrial micrometastasis, however, it is difficult to early-stage cervical cancer remains a subject of active debate. interpret the prognostic significance of this finding. Previous reports have documented parametrial disease in 4-39%

245 ANTICANCER RESEARCH 31: 243-248 (2011) of patients (2, 3, 6, 8, 9, 12, 16, 25, 32-35). Despite the node biopsy in colon cancer: a prospective multicenter trial. Ann frequency of parametrial tumor spread, several series have Surg 245: 858-863, 2007. suggested that women with small volume tumors and other 2 Benedetti-Panici P, Maneschi F, D’Andrea G, Cutillo G, Rabitti favorable pathologic characteristics are at a relatively low risk C, Congiu M, Coronetta F and Capelli A: Early cervical carcinoma: the natural history of lymph node involvement for parametrial metastasis (6, 16, 33). In a series of over 800 redefined on the basis of thorough parametrectomy and giant women with stage I cervical cancer who underwent radical section study. Cancer 88: 2267-2274, 2000. hysterectomy, those with tumors less than 2 cm in diameter, with 3 Bleker OP, Ketting BW, van Wayjen-Eecen B and Kloosterman negative pelvic lymph nodes and superficial stromal invasion had GJ: The significance of microscopic involvement of the a rate of parametrial involvement of less than 1% (6). parametrium and/or pelvic lymph nodes in cervical cancer stages The goal of our study was to determine the rate of occult IB and IIA. Gynecol Oncol 16: 56-62, 1983. parametrial micrometastasis in women with early-stage 4 Candido EB, Silva-Filho AL, Triginelli SA, Pedrosa MS, Macarenco R, Porto L and Traiman P: Histopathological and cervical cancer. We sought to identify a group of patients at immunohistochemical (cytokeratins AE1/AE3) study of the extremely low risk for parametrial spread who may be parametrium of patients with early-stage cervical cancer. Eur J candidates for parametrium-sparing surgery. Our group Obstet Gynecol Reprod Biol 141: 58-63, 2008. previously demonstrated that women with tumors <2 cm in 5 Cote RJ, Peterson HF, Chaiwun B, Gelber RD, Goldhirsch A, diameter without lymphvascular space invasion and negative Castiglione-Gertsch M, Gusterson B and Neville AM: Role of pelvic nodes are at extremely low risk for parametrial immunohistochemical detection of lymph-node metastases in metastasis. In our retrospective report, parametrial management of breast cancer. International Breast Cancer Study Group. Lancet 354: 896-900, 1999. involvement was noted in only 0.4% of women whose 6 Covens A, Rosen B, Murphy J, Laframboise S, DePetrillo AD, tumors met the above pathologic criteria (33). In our current Lickrish G, Colgan T, Chapman W and Shaw P: How important series, none of the patients who fell into this group had is removal of the parametrium at surgery for carcinoma of the parametrial micrometastasis. cervix? Gynecol Oncol 84: 145-149, 2002. While it is encouraging that the rate of parametrial 7 Darai E, Rouzier R, Ballester M, Barranger E and Coutant C: micrometastasis in our study was low, it is also a limitation Sentinel lymph node biopsy in gynaecological cancers: the in so as much as the study is underpowered to determine the importance of micrometastases in cervical cancer. Surg Oncol 17 prognostic significance of parametrial micrometastasis, as : 227-235, 2008. 8 Delgado G, Bundy BN, Fowler WC Jr., Stehman FB, Sevin B, well as factors associated with micrometastatic disease. To Creasman WT, Major F, DiSaia P and Zaino R: A prospective identify the true rate of parametrial micrometastasis we surgical pathological study of stage I squamous carcinoma of the examined multiple parametrial sections from each patient. cervix: a Gynecologic Oncology Group Study. Gynecol Oncol However, the entire parametrium was not examined, and as 35: 314-320, 1989. such we cannot exclude the possibility that additional 9 Girardi F, Lichtenegger W, Tamussino K and Haas J: The parametrial micrometastases were not detected. importance of parametrial lymph nodes in the treatment of Our findings are reassuring in that it appears that occult cervical cancer. Gynecol Oncol 34: 206-211, 1989. 10 Hoffman MS: Extent of radical hysterectomy: evolving parametrial micrometastases are rare in women with stage IA emphasis. Gynecol Oncol 94: 1-9, 2004. and IB cervical cancer. As such, the routine performance of 11 Hoffman MS and Cardosi RJ: Intraoperative measurements to immunohistochemistry to identify parametrial metastases is determine the extent of radical hysterectomy. Gynecol Oncol 87: probably of little clinical utility. Our data provide further support 281-286, 2002. to the concept that a low-risk subset of patients with early-stage 12 Inoue T and Okumura M: Prognostic significance of parametrial cervical cancer exists and that these women may be candidates extension in patients with cervical carcinoma stages IB, IIA, and for less radical, parametrium-sparing surgical techniques. IIB. A study of 628 cases treated by radical hysterectomy and lymphadenectomy with or without postoperative irradiation. Conization or simple hysterectomy in combination with pelvic Cancer 54: 1714-1719, 1984. lymphadenectomy may be adequate treatment for these patients. 13 Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T and Thun As more conservative surgical approaches gain acceptance for MJ: Cancer statistics, 2008. CA Cancer J Clin 58: 71-96, 2008. other solid tumors, it is time to re-evaluate the necessity of 14 Juretzka MM, Jensen KC, Longacre TA, Teng NN and Husain radical en bloc resection of early cervical carcinomas. A: Detection of pelvic lymph node micrometastasis in stage IA2- IB2 cervical cancer by immunohistochemical analysis. Gynecol References Oncol 93: 107-111, 2004. 15 Kadar N, Saliba N and Nelson JH: The frequency, causes and 1 Bembenek AE, Rosenberg R, Wagler E, Gretschel S, Sendler A, prevention of severe urinary dysfunction after radical Siewert JR, Nahrig J, Witzigmann H, Hauss J, Knorr C, hysterectomy. Br J Obstet Gynaecol 90: 858-863, 1983. Dimmler A, Grone J, Buhr HJ, Haier J, Herbst H, Tepel J, 16 Kinney WK, Hodge DO, Egorshin EV, Ballard DJ and Podratz Siphos B, Kleespies A, Koenigsrainer A, Stoecklein NH, KC: Identification of a low-risk subset of patients with stage IB Horstmann O, Grutzmann R, Imdahl A, Svoboda D, Wittekind invasive squamous cancer of the cervix possibly suited to less C, Schneider W, Wernecke KD and Schlag PM: Sentinel lymph radical surgical treatment. Gynecol Oncol 57: 3-6, 1995.

246 Brady et al: Parametrial Micrometastases in Early-stage Cervical Cancer

17 Knopp S, Holm R, Trope C and Nesland JM: Occult lymph node 27 Parkin DM, Bray F, Ferlay J and Pisani P: Global cancer metastases in early-stage vulvar carcinoma patients. Gynecol statistics, 2002. CA Cancer J Clin 55: 74-108, 2005. Oncol 99: 383-387, 2005. 28 Photopulos GJ and Zwaag RV: Class II radical hysterectomy 18 Landoni F, Maneo A, Colombo A, Placa F, Milani R, Perego P, shows less morbidity and good treatment efficacy compared to Favini G, Ferri L and Mangioni C: Randomised study of radical class III. Gynecol Oncol 40: 21-24, 1991. surgery versus radiotherapy for stage Ib-IIa cervical cancer. 29 Piver MS, Rutledge F and Smith JP: Five classes of extended Lancet 350: 535-540, 1997. hysterectomy for women with cervical cancer. Obstet Gynecol 19 Landoni F, Maneo A, Cormio G, Perego P, Milani R, Caruso O 44: 265-272, 1974. and Mangioni C: Class II versus class III radical hysterectomy in 30 Sartori E, Fallo L, La Face B, Bianchi UA and Pecorelli S: stage IB-IIA cervical cancer: a prospective randomized study. Extended radical hysterectomy in early-stage carcinoma of the Gynecol Oncol 80: 3-12, 2001. uterine cervix: tailoring the radicality. Int J Gynecol Cancer 5: 20 Lentz SE, Muderspach LI, Felix JC, Ye W, Groshen S and 143-147, 1995. Amezcua CA: Identification of micrometastases in histologically 31 Wang HY, Sun JM, Lu HF, Shi DR, Ou ZL, Ren YL and Fu SQ: negative lymph nodes of early-stage cervical cancer patients. Micrometastases detected by cytokeratin 19 expression in Obstet Gynecol 103: 1204-1210, 2004. sentinel lymph nodes of patients with early-stage cervical cancer. 21 Low JA, Mauger GM and Carmichael JA: The effect of Int J Gynecol Cancer 16: 643-648, 2006. Wertheim hysterectomy upon bladder and urethral function. Am 32 Winter R, Haas J, Reich O, Koemetter R, Tamussino K, J Obstet Gynecol 139: 826-834, 1981. Lahousen M, Petru E and Pickel H: Parametrial spread of 22 Magrina JF, Goodrich MA, Lidner TK, Weaver AL, Cornella JL cervical cancer in patients with negative pelvic lymph nodes. and Podratz KC: Modified radical hysterectomy in the treatment Gynecol Oncol 84: 252-257, 2002. of early squamous cervical cancer. Gynecol Oncol 72: 183-186, 33 Wright JD, Grigsby PW, Brooks R, Powell MA, Gibb RK, Gao 1999. F, Rader JS and Mutch DG: Utility of parametrectomy for early- 23 Marchiole P, Buenerd A, Benchaib M, Nezhat K, Dargent D and stage cervical cancer treated with radical hysterectomy. Cancer Mathevet P: Clinical significance of lympho-vascular space 110: 1281-1286, 2007. involvement and lymph node micrometastases in early-stage 34 Yang YC and Chang CL: Modified radical hysterectomy for cervical cancer: a retrospective case–control surgico-pathological early Ib cervical cancer. Gynecol Oncol 74: 241-244, 1999. study. Gynecol Oncol 97: 727-732, 2005. 35 Yen MS, Yuan CC, Wang PH, Ng HT, Twu NF and Juang CM: 24 Marinho VF, Zagury MS, Caldeira LG and Gobbi H: Parametrial tumor spreading patterns in cervix cancer patients Relationship between histologic features of primary breast treated by radical hysterectomy. Int J Gynaecol Obstet 80: 145- carcinomas and axillary lymph node micrometastases: Detection 151, 2003. and prognostic significance. Appl Immunohistochem Mol Morphol 14: 426-431, 2006. 25 Nakanishi T, Ishikawa H, Nawa A, Suzuki Y, Inoue T, Nakamura S and Kuzuya K: The significance of tumor size in clinical stage IB cervical cancer: Can a cut-off figure be determined? Int J Gynecol Cancer 10: 397-401, 2000. 26 Narayansingh GV, Miller ID, Sharma M, Welch CJ, Sharp L, Parkin DE and Cruickshank ME: The prognostic significance of Received November 3, 2010 micrometastases in node-negative squamous cell carcinoma of Revised December 10, 2010 the . Br J Cancer 92: 222-224, 2005. Accepted December 13, 2010

247