03 (2006) 942–947www.elsevier.com/locate/ygyno

Gynecologic Oncology 1

Detection of disseminated tumor cells in patients withgynecological cancers

T. Fehm a,⁎, S. Becker a, C. Bachmann a, V. Beck a, G. Gebauer b,M. Banys a, D. Wallwiener a, E.F. Solomayer a

a Department of Obstetrics and Gynecology, University of Tuebingen, Calwerstrasse 7, D- 72076 Tuebingen, Germanyb Department of Gynecology and Obstetrics, University of Heidelberg, Vossstrasse 9, 69115 Heidelberg, Germany

Received 15 February 2006Available online 4 August 2006

Abstract

Objectives. The presence of disseminated tumor cells (DTC) in breast cancer patients is associated with poor prognosis. However, there arelimited data about the prevalence and prognostic impact of DTC in patients with gynecological tumors. The aim of this study was to evaluate thepresence of DTC in the bone marrow (BM) of patients with gynecological cancers and to correlate their presence with established prognosticfactors.

Methods. BM aspirates of 201 patients with primary ovarian (n=69), cervical (n=54) and endometrial cancer (n=78), undergoing surgery atthe Department of Gynecology and Obstetrics, University Hospital, Tuebingen, Germany between 1/2002 and 01/2006, were included into thestudy. Cytokeratin (CK)-positive cells were identified by immunocytochemistry using the pancytokeratin antibody A45B/B3.

Results. The bone marrow positivity rate was 36% in ovarian, 26% in cervical and 17% in endometrial cancer, respectively. Presence of DTCwas significantly correlated with FIGO (International Federation of Gynecology and Obstetrics) tumor stage (p<0.05). The recurrence rate was14% in patients with CK-positive cells compared to 8% in CK-negative patients (p=0.2). There was no correlation between DTC and otherestablished prognostic factors including nodal status or grading except for cervical cancer. Patients with positive lymph node status were morelikely to be bone marrow positive compared to those with negative lymph node status (p<0.05).

Conclusions. Disseminated tumor cells seem to be a general phenomenon in epithelial tumors even though their clinical impact remains to beevaluated. The hypothesis that bone marrow is the homing site of disseminated tumor cells is further supported by these data since gynecologicaltumors only rarely metastasize to the skeletal system.© 2006 Elsevier Inc. All rights reserved.

Keywords: Tumor cell dissemination; Ovarian cancer; Endometrial cancer; Cervical cancer; Bone marrow

Introduction

Disseminated tumor cells (DTC) are detectable in 10–40%of bone marrow aspirates of breast cancer patients. Theirpresence is associated with poor prognosis. Breast cancerpatients with DTC in bone marrow have a shorter disease-freeand overall survival compared to those patients with negativebone marrow status [1–6]. Disseminated tumor cells are thoughtto be derived from the primary tumor and are considered to bethe potential precursors of subsequent metastatic disease [7].

⁎ Corresponding author. Fax: +49 7071 295286.E-mail address: tanja.fehm@med.uni-tuebingen.de (T. Fehm).

0090-8258/$ - see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.ygyno.2006.05.049

Over the past years, evidence has been accumulating thattumor cell dissemination is a common phenomenon in solidtumors and not limited to breast cancer [8–10]. Based onexperimental animal studies with tumors of the mammarygland, it was hypothesized that solid cancers are regularlyshedding neoplastic cells into the blood and subsequently intothe bone marrow and other organs as well [10]. Several studieshave been already supporting the theory that shedding oftumor cells occurs in other types of solid cancers includingcolorectal, pancreatic or gynecological cancers [11–13].However, only limited data are available regarding thepresence of disseminated tumor cells in these tumor entities.Therefore, the aim of this prospective study was (a) to

Table 1Clinical data of the 201 patients with gynecological cancers

Clinical data n=201 CK-pos % p value a

Total 201 52 26 –Ovarian cancer 69 25 36 <0.05Endometrial cancer 78 13 17Cervical cancer 54 14 26FIGOI 107 19 18 <0.05II 25 9 36III–IV 68 24 35

Nodal status b

No 125 30 24 n.s.N1 57 17 30

GradingG1 12 1 8 n.s.G2 121 30 25G3 57 20 35

Lymphangiosis carcinomatosaLo 123 29 24 n.s.L1 40 13 33

Vascular invasionVo 155 38 25 n.s.V1 7 3 43

a Chi-squared-test: p<0.05.b Lymphonodectomy was not performed in patients with M1 or small tumors.

943T. Fehm et al. / Gynecologic Oncology 103 (2006) 942–947

determine the preoperative bone marrow status in patients withnewly diagnosed ovarian, endometrial and cervical cancer and(b) to evaluate its correlation with established prognosticfactors.

Patients and methods

Patients with primary gynecological cancers such as ovarian, cervical andendometrial cancer were eligible for this study. All patients underwent surgery atthe Department of Oncology and Obstetrics, University Hospital, Tuebingen,Germany, from 1/2002 until 1/2006. Exclusion criteria from this analysis wereprevious history of cancer or secondary malignancy. After surgery, patients weretreated with chemotherapy and/or radiotherapy based on type of cancer andaccording to current treatment guidelines of the comprehensive cancer center(CCC Tuebingen, Germany). Relapse was confirmed – depending on tumorentity and site of recurrence – by examination, CT scan, X-ray, ultrasound,tumor marker and/or relaparotomy. The median follow-up was 5 months(ranging between 0.1 and 43 months).

Collection and analysis of BM

Between 10 and 20 ml of bone marrow was obtainedintraoperatively from the anterior iliac crest prior to laparotomyas a routine procedure and processed within 24 hours.

All specimens were obtained after written informedconsent. The analysis was approved by the local ethicalcommittee (114/2006A). Tumor cell isolation and detectionwas performed based on the recommendations for standardizedtumor cell detection [15]. Bone marrow samples wereseparated by density centrifugation using Bicoll (density1077 g/ml, Biochrom, Germany). Mononuclear cells werecollected from the interphase layer. If necessary, lysis of redblood cells was performed with lysis buffer (155 mM NH4Cl,10 mM KHC03, 0.1 mM EDTA pH 7.2). Cells (106 MNC/spot) were spun down onto a glass slide (Hettich cytocen-trifuge, Germany). Slides were air-dried overnight at roomtemperature. For detection of cytokeratin-positive (CK) tumorcells, slides were fixed in 4% neutral buffered formalin for 10min and rinsed in PBS. Automatic immunostaining wasperformed on the DAKO Autostainer using the monoclonalmouse A45-B/B3 antibody (Micromet, Germany) and theDAKO-APAAP detection kit (DakoCytomation, Denmark)according to the manufacturer's instructions. The A45-B/B3antibody is directed against common cytokeratin epitopesincluding the CK heterodimers 8/18 and 8/19. The malignantbreast cell line MCF-7 was used as a positive control.Leukocytes of a healthy volunteer served as negative control.In addition, isotype matched myeloma protein conjugated toFITC was included as negative staining controls (Sigma,Deisenhofen). For each patient, 2×106 cells were analyzed ontwo slides. Slides were automatically scanned using theACIS™ imaging system (ChromaVision, Medical SystemsInc., San Juan, Capistrano, CA) which is a computerizedmicroscope with an imaging processing system optimized fordetection of rare cells in specimens. Details of this systemhave been described in detail elsewhere [14]. Criteria forevaluation of immunostained cells were based on the criteriaof the European ISHAGE Working group for standardizationof tumor cell detection [16].

Statistical analysis

Chi-squared test was used to examine the associationbetween positive bone marrow status and clinicopathologicalfactors. Survival analysis was performed by Kaplan–Meier-method. Statistical analysis was performed using SPSS (Version11.5) considering p values less than 0.05 to be statisticallysignificant.

Results

201 patients with primary ovarian, cervical and endometrialcancer were included into the analysis (Table 1). 53% of thesepatients were at FIGO I stage, 13% at FIGO II and 34% atFIGO stage III/IV. Clinical data are summarized in Table 1.The overall incidence rate of disseminated tumor cells inpatients with gynecological cancer was 26% (Table 1).Disseminated tumor cells were identified by cytokeratin-positivity and cytomorphology (Fig. 1). 52 of 201 patientswith gynecological cancer showed disseminated tumor cells inthe bone marrow. The number of CK (cytokeratin)-positivecells ranged from 1 to 20 per 2×106 mononuclear cells.Including all cancer types, a significant correlation could beseen between bone marrow positivity and FIGO stage (Table1). Only 19% of patients at FIGO I stage were bone marrowpositive compared to 36% and 35% at FIGO II and III–IVstage, respectively (p<0.01).

69 of 201 patients were diagnosed with primary ovariancancer, 54 patients with cervical and 78 with primaryendometrial cancer. Table 2 shows the positivity rates sub-divided by cancer diagnosis. The highest incidence ofdisseminated tumor cells was seen in patients with ovarian

Table 3Recurrence rates subdivided after bone marrow positivity

Cancer n Recurrence p value

Total 199 18 (9%) 0.2CK negative 147 11 (8%)CK positive 52 7 (14%)Ovarian cancer 69 12 (17%) 0.3CK negative 44 6 (14%)CK positive 25 6 (26%)Endometrial cancer 75 3 (4%) 0.4CK negative 62 3 (5%)CK positive 13 0Cervical cancer 54 2 (4%) 0.4CK negative 40 1 (3%)CK positive 14 1 (7%)

Fig. 1. Cytokeratin-positive tumor cells of an ovarian cancer patient.

944 T. Fehm et al. / Gynecologic Oncology 103 (2006) 942–947

cancer with 36% followed by 26% in cervical cancer and 17% inendometrial cancer. No correlation was observed between bonemarrow positivity and other clinico-pathological factors such asFIGO tumor stage, lymph node status, grading, lymphangiosiscarcinomatosa or vascular invasion except for cervical cancer.Cervical cancer patients with positive lymph nodes were morelikely to be bone marrow positive compared to those withnegative lymph node status (p<0.01). These results aresummarized in detail in Table 2.

The correlation between bone marrow positivity andclinical outcome was analyzed. Eighteen of 199 patientsrelapsed during follow-up. Twelve of these patients had localrecurrence. Recurrence rates were 17%, 4% and 4% inovarian, endometrial and cervical cancer, respectively. Resultsare shown in Table 3. Presence of disseminated tumor cells

Table 2Incidence of disseminated tumor cells (CK+) in patients with gynecologicalcancer based on diagnosis and clinico-pathological factors

Factor Ovarian cancer;n=69; CK+/total(%)

Endometrial cancer;n=78; CK+/total(%)

Cervical cancer;n=54; CK+/total(%)

FIGOI 2/8 (25) 12/63 (19) 6/37 (16)II 1/2 (50) 1/10 (10) 7/13 (54)III–IV 22/59 (37) 1/5 (20) 1/4 (25)

Nodal statusNo 12/25 (48) 13/65 (20) 6/36 (17) a

N1 10/35 (29) 0/6 (0) 7/16 (44)GradingG1 0/3 (0) 1/9 (11) –G2 11/31 (36) 12/56 (21) 8/35 (23)G3 13/29 (45) 1/11 (9) 6/17 (35)

LymphangiosisLo 9/26 (35) 14/67 (21) 7/31 (23)L1 7/20 (35) 0/7 (0) 6/13 (46)

Vascular invasionVo 14/41 (34) 14/73 (20) 11/42 (26)V1 1/4 (25) – 2/3 (67)a p<0.05.

was not associated with higher recurrence rates (p=0.2) (Table3) or with shortened disease-survival (p=0.03). The recurrencerate was 14% in patients with CK-positive cells compared to8% in CK-negative patients.

Discussion

The aim of this prospective study was to evaluate theincidence of DTC in gynecological cancer and to determine itsassociation with established prognostic factors.

Disseminated tumor cells could be detected in 52 of 201patients with gynecological cancer (26%). The highestpositivity rate was observed in ovarian cancer (36%) followedby cervical cancer (26%) and endometrial cancer (17%). Thedifferences in positivity rates may be not related to the tumorlocation but rather to the fact that the FIGO tumor stages werenot equally distributed in the different tumor entities. In ovariancancer most of the patients were FIGO III (86%) whereas incervical cancer and endometrial cancer most patients wereFIGO I (76%). The detection rates reported by other studies forgynecological cancers were between 21% and 30% for ovariancancer [24–26] and 5% to 87% for cervical cancer [17–23],respectively. For endometrial cancer, the presence of DTC hasonly been investigated in peripheral blood: 13% to 51% ofendometrial cancer patients are reported to be positive for DTC[27,28]. Table 4 summarizes the positivity rates reported byother studies.

Based on these results, the positivity rates for gynecologicalcancers do not significantly differ from those observed in breastcancer [1–6]. These findings are surprising since in these cancerentities bone metastases are relatively rare events compared tobreast cancer [29–32]. This would support the theory that theDTC are not necessarily precursors of bone metastases. Rather,it appears that BM is an ideal homing site for DTC, irrespectiveof their oncologic origin [7,33,34].

Correlation with other prognostic factors

Presence of disseminated tumor cells was correlated withFIGO tumor stages when all patients were included in theanalysis regardless of tumor location. However, when

Table 4Incidence of disseminated tumor cells in bone marrow of patients with gynecological cancer (breast cancer (BC), cervical cancer (CC), ovarian cancer (OVC),endometrial cancer (EMC) and vulva cancer (VC)

Author Specimen Method Marker Cancer n BMa positive In %

Sehouli '03 [13] PBb ICC c CK 8–18, A45-B/B3 BCd, OC, CC, EMC, VC 167 47 28Scheungraber '02 [16] PB, BMa PCR HPV e CC 24 6 25Janni '03 [17] BM ICC CK8/18/19 CC 130 38 29Dong '02 [18] PB PCR HPV CC 232 12 5Yuan '02 [19] PB RT-PCR CK 19 CC 84 18 21Tseng '99 [20] PB RT-PCR HPV CC 35 18 51Stenman '97 [21] PB RT-PCR SSC f CC 15 6 40Pao '97 [22] PB (RT-)PCR HPV CC 15 13 87Braun ´01 [23] BM ICC CK8/18/19 OVC 108 32 30Roggel '01 [24] BM ICC, IF CK8/18/19 OVC 60 18 30Marth '02 [25] PB, BM ICC MOC-31 OVC 90 14 21Ji '05 [26] PB PCR, RT-PCR CK 19 EMC 30 4 13Klein '00 [27] PB RT-PCR CK 20 EMC 59 30 51Ji' 06 [37] PB RT-PCR CK 19 EMC 30 3 10a Bone marrow, b Peripheral blood, c Immunocytochemistry (ICC), d Breast cancer, e Human papilloma virus, f Squamous-cell-carcinoma antigen (SCC).

945T. Fehm et al. / Gynecologic Oncology 103 (2006) 942–947

subdivided by tumor entity, no correlation could be observed.Moreover, no association was seen for established conventionalprognostic factors including lymph node status or grading forovarian and endometrial cancer. Only in patients with cervicalcancer an association between tumor cell dissemination andpositive lymph node status could be detected. In concordancewith Tseng et al. [21] cervical cancer patients with lymph nodemetastases were more likely to be tumor cell positive comparedto those who were lymph node negative.

Detection of DTC based on other markers

Detection of epithelial cells in blood or bone marrow isusually performed by immunocytochemistry using antibodiesagainst cytokeratins [1,5,13,18,24]. An alternative method isthe amplification of (m)RNA of epithelial markers. RT-PCRprotocols for detection of cytokeratin 20 in blood/bonemarrow of endometrial cancer patients [28] and for SCC(squamous cancer cell) antigen in cervical cancer patientshave been developed [22]. Furthermore, in cervical cancer,the detection of human papilloma virus (HPV) RNA/DNAmay be an interesting marker for detection of minimalresidual disease, since nearly 99% of cervical cancers areassociated with an HPV-high-risk infection [17]. So far, fourstudies used this strategy for the detection of disseminatedtumor cells in the peripheral blood and bone marrow based onthe detection of HPV-DNA and RNA [17,19,21,23]. Positivityrates were between 5 and 87%. However, currently there areno data available whether HPV is a better marker thancytokeratin for detection of disseminated tumor cells.

DTC and prognostic significance

The prognostic significance of disseminated tumor cells inbone marrow of breast cancer has been widely studied. Basedon a pooled analysis including more than 4700 patients, bonemarrow positivity is an independent prognostic factor [5]. Incontrast to breast cancer, the prognostic significance of

disseminated tumor cells in gynecological cancer is controver-sial [14,17,18,24,28]. One study including more than 100ovarian cancer patients could demonstrate the unfavorableprognosis associated with the presence of occult metastaticdisease [24]. For cervical cancer [17,18], only data from studieswith small numbers of patients are available. These investiga-tions support the prognostic relevance of DTC and the potentialorigin of systemic tumor cell spread during follow-up.However, in our analysis, no significant correlation betweenbone marrow positivity and clinical outcome was observed forany of the three tumor entities studied. This may be explainedby short follow-up and low number of recurrences in eachtumor entity. Moreover, most of the patients had only localrecurrence. The relevance of DTC in endometrial cancer forprognosis has not been investigated. Interestingly, in our subset,none of the cytokeratin-positive patients with endometrialcancer relapsed whereas 3 patients with CK-negative bonemarrow had a recurrence during follow-up. The prognosticrelevance and clinical impact of DTC may vary between thedifferent tumor entities. The initiation of prospective multi-center studies will be necessary to definitively assess theprognostic value of disseminated tumor cells in these tumorentities.

DTC for monitoring therapy response

The detection of DTCs in gynecological cancer might alsobecome a useful parameter for monitoring therapy response, inparticular in ovarian and cervical cancer. Tumor-associatedantigens such as CA125 in ovarian or SCC in cervical cancer arewell established as tumor markers in advanced disease fortherapy monitoring. However, in ovarian cancer patients, tumormarker levels usually drop below cut-off during first linechemotherapy even though most patients suffer relapse of thedisease over the following 5 years. In cervical cancer, elevatedtumor marker levels are rarely observed due to small tumorburden in most cases limiting the usefulness of serial tumormarker monitoring. The detection of DTCs in bone marrow and

946 T. Fehm et al. / Gynecologic Oncology 103 (2006) 942–947

possibly in peripheral blood, however, could indicate sub-clinical tumor load especially after completion of first linetreatment. Therefore, the detection of DTCs may be regarded asan indicator for suboptimal therapy response. For breast cancer,it has already been shown that tumor cell persistence aftersurgery and adjuvant chemotherapy is associated with poorprognosis [35,36]. Targeting such minimal residual disease bysecondary adjuvant therapy is currently under investigation andshould also be further explored in gynecological cancer.

In conclusion, disseminated tumor cells are present in 17% to36% of patients with gynecological cancer. The hypothesis thatbone marrow is the homing site of disseminated tumor cells isfurther supported by these data sine bone metastases are rareevents in these cancer entities. However, their clinical impactremains to be evaluated in further (multicenter) trials.

Acknowledgment

We thank Nancy Lane, Cancer Immunobiology Center, UTSouthwestern Medical School for reviewing the manuscript.

References

[1] Braun S, Pantel K, Muller P, Janni W, Hepp F, Kentenich CR, et al.Cytokeratin-positive cells in the bone marrow and survival of patients withstage I, II, or III breast cancer. N Engl J Med 2000;342:525–33.

[2] Mansi JL, Gogas H, Bliss JM, Gazet JC, Berger U, Coombes RC. Outcomeof primary breast cancer patients with micrometastases: a long-termfollow-up study. Lancet 1999;354:197–202.

[3] Wiedswang G, Borgen E, Karesen R, Kvalheim G, Nesland JM, QvistH, et al. Detection of isolated tumor cells in bone marrow is anindependent prognostic factor in breast cancer. J Clin Oncol 2003;21(18):3469–78.

[4] Gebauer G, Fehm T, Merkle E, Beck EP, Lang N, Jager W. Epithelial cellsin bone marrow of breast cancer patients at time of primary surgery:clinical outcome during long-term follow-up. J Clin Oncol 2001;19(16):3669–3674.

[5] Braun S, Vogl FD, Naume B, Janni W, Osborne MP, Coombes RC, et al. Apooled analysis of bone marrow micrometastasis in breast cancer. N Engl JMed 2005;353(8):793–802.

[6] Solomayer EF, Diel IJ, Salanti G, Hahn M, Gollan C, Schutz F, et al. Timeindependence of the prognostic impact of tumor cell detection in the bonemarrow of primary breast cancer patients. Clin Cancer Res 2001;7(12):4102–4108.

[7] Pantel K, Brakenhoff RH. Dissecting the metastatic cascade. Nat Rev,Cancer 2004;4:448–56.

[8] Fidler IJ. Metastasis: quantitative analysis of distribution and fate of tumorembolilabeled with 125 I-5-iodo-2′-deoxyuridine. N Engl J Med 2005;353(8):793–802.

[9] Liotta LA, Kleinerman J, Saidel GM. Quantitative relationships ofintravascular tumor cells, tumor vessels, and pulmonary metastasesfollowing tumor implantation. Cancer Res 1974;34(5):997–1004.

[10] Butler TP, Gullino PM. Quantitation of cell shedding into efferent blood ofmammary adenocarcinoma. Cancer Res 1975;35(3):512–6.

[11] Chausovsky G, Luchansky M, Figer A, Shapira J, Gottfried M, Novis B,et al. Expression of cytokeratin 20 in the blood of patients withdisseminated carcinoma of the pancreas, colon, stomach, and lung. Cancer1999;86(11):2398–405.

[12] Lukyanchuk VV, Friess H, Kleeff J, Osinsky SP, Ayuni E, Candinas D,et al. Detection of circulating tumor cells by cytokeratin 20 and prostatestem cell antigen RT-PCR in blood of patients with gastrointestinalcancers. Anticancer Res 2003;23(3B):2711–6.

[13] Sehouli J, Konsgen D, Nimpsch R, Stengel D, Oskay G, Mustea A, et al.

Prognostic significance of epithelial cells in the blood of patients withgynaecological malignancies. Anticancer Res 2003;3(5b):4133–40.

[14] Bauer KD, de la Torre-Bueno J, Diel IJ, Hawes D, Decker WJ, Priddy C,et al. Reliable and sensitive analysis of occult bone marrow metastasesusing automated cellular imaging. Clin Cancer Res 2000;6(9):3552–9.

[15] Fehm T, Braun S, Müller V, Janni W, Gebauer G, Marth C. et al. Aconcept for the standardized detection of disseminated tumor cells (DTC)in bone marrow of patients with primary breast cancer and its clinicalimplementation. Cancer (in press).

[16] Borgen E, Naume B, Nesland JM, Kvalheim G, Beiske K, Fodstad O, et al.Standardization of the immunocytochemical detection of cancer cells inBM and blood: I. Establishment of objective criteria for the evaluation ofimmunostained cells. Cytotherapy 1999;5:377–88.

[17] Scheungraber C, Muller B, Kohler C, Possover M, Leistritz S, SchneiderA, et al. Detection of disseminated tumor cells in patients with cervicalcancer. J Cancer Res Clin Oncol 2002;128(6):323–9.

[18] Janni W, Hepp F, Strobl B, Rack B, Rjosk D, Kentenich C, et al. Patterns ofdisease recurrence influenced by hematogenous tumor cell disseminationin patients with cervical carcinoma of the uterus. Cancer 2003;97(2):405–411.

[19] Dong SM, Pai SI, Rha SH, Hildesheim A, Kurman RJ, Schwartz PE, et al.Detection and quantitation of human papillomavirus DNA in the plasma ofpatients with cervical carcinoma. Cancer Epidemiol Biomarkers Prev2002;11(1):3–6.

[20] Yuan CC, Wang PH, Ng HT, Li YF, Huang TS, Chen CY, et al.Detecting cytokeratin 19 mRNA in the peripheral blood cells of cervicalcancer patients and its clinical–pathological correlation. Gynecol Oncol2002;85(1):148–53.

[21] Tseng CJ, Pao CC, Lin JD, Soong YK, Hong JH, Hsueh S. Detection ofhuman papillomavirus types 16 and 18 mRNA in peripheral blood ofadvanced cervical cancer patients and its association with prognosis. J ClinOncol 1999;17(5):1391–6.

[22] Stenman J, Lintula S, Hotakainen K, Vartiainen J, Lehvaslaiho H, StenmanUH. Detection of squamous-cell carcinoma antigen-expressing tumourcells in blood by reverse transcriptase-polymerase chain reaction in cancerof the uterine cervix. Int J Cancer 1997;74(1):75–80.

[23] Pao CC, Hor JJ, Yang FP, Lin CY, Tseng CJ. Detection of humanpapillomavirus mRNA and cervical cancer cells in peripheral blood ofcervical cancer patients with metastasis. J Clin Oncol 1997;15(3):1008–1012.

[24] Braun S, Schindlbeck C, Hepp F, Janni W, Kentenich C, Riethmuller G,et al. Occult tumor cells in bone marrow of patients with locoregionallyrestricted ovarian cancer predict early distant metastatic relapse. J ClinOncol 2001;19(2):368–75.

[25] Roggel F, Hocke S, Lindemann K, Sinz S, Welk A, Bosl M, et al. Minimalresidual disease in breast cancer and gynecological malignancies: phenotypeand clinical relevance. Recent Results Cancer Res 2003;162: 89–100.

[26] Marth C, Kisic J, Kaern J, Trope C, Fodstad O. Circulating tumor cells inthe peripheral blood and bone marrow of patients with ovarian carcinomado not predict prognosis. Cancer 2002;94(3):707–12.

[27] Ji XQ, Sato H, Tanaka H, Konishi Y, Fujimoto T, Takahashi O, et al. Real-time quantitative RT-PCR detection of disseminated endometrial tumorcells in peripheral blood and lymph nodes using the LightCycler System.Gynecol Oncol 2006;100(2):355–60.

[28] Klein A, Fishman A, Zemer R, Zimlichman S, Altaras MM. Detection oftumor circulating cells by cytokeratin 20 in the blood of patients withendometrial carcinoma. Gynecol Oncol 2000;78:352–5.

[29] Cormio G, Rossi C, Cazzolla A, Resta L, Loverro G, Greco P, et al. Distantmetastases in ovarian carcinoma. Int J Gynecol Cancer 2003; 13(2):125–9.

[30] Fagundes H, Perez CA, Grigsby PW, Lockett MA. Distant metastases afterirradiation alone in carcinoma of the uterine cervix. Int J Radiat Oncol BiolPhys 1992;24(2):377–80.

[31] Ratanatharathorn V, Powers WE, Steverson N, Han I, Ahmad K, Grimm J.Bone metastasis from cervical cancer. Cancer 1994;73(9):2372–9.

[32] Mariani A, Webb MJ, Keeney GL, Calori G, Podratz KC. Hematogenousdissemination in corpus cancer. Gynecol Oncol 2001;80(2):233–8.

[33] Puricelli L, Gomez DE, Vidal MC, Eijan AM, Spinelli O, Alonso DF,de Lustig ES, et al. Effect of host-organ environment on the in vivo and

947T. Fehm et al. / Gynecologic Oncology 103 (2006) 942–947

in vitro behaviour of a murine mammary adenocarcinoma. Tumour Biol1994;15(5):284–93.

[34] Sipkins DA, Wei X, Wu JW, Runnels JM, Cote D, Means TK, et al. In vivoimaging of specialized bone marrow endothelial microdomains for tumourengraftment. Nature 2005;435(7044):969–73.

[35] Janni W, Rack B, Schindlbeck C, Strobl B, Rjosk D, Braun S, et al. Thepersistence of isolated tumor cells in bone marrow from patients with breastcarcinoma predicts an increased risk for recurrence. Cancer 2005;103:884–91.

[36] Wiedswang G, Borgen E, Karesen R, Qvist H, Janbu J, Kvalheim G,Nesland JM, et al. Isolated tumor cells in bone marrow three years afterdiagnosis in disease-free breast cancer patients predict unfavorable clinicaloutcome. Clin Cancer Res 2004;10:5342–8.

[37] Ji XQ, Sato H, Tanaka H, Konishi Y, Fujimoto T, Takahashi O, et al. Real-time quantitative RT-PCR detection of disseminated endometrial tumorcells in peripheral blood and lymph nodes using the LightCycler System.Gynecol Oncol 2006;100(2):355–60.