8/12/2019 70118535

1/10

Pre-treatment diagnosis of endometrial cancer through acombination of CA125 and multiplication of neutrophiland monocyte

Bo Wook Kim1, Young Eun Jeon1, Hanbyoul Cho1, Eun Ji Nam2, Sang Wun Kim2,Sunghoon Kim2, Young Tae Kim2 and Jae-Hoon Kim1

1Department of Obstetrics and Gynecology, Gangnam Severance Hospital, Yonsei University College of Medicine, and2Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Yonsei Cancer Center, Yonsei UniversityCollege of Medicine, Seoul, Korea jog_1694 48..56

Abstract

Aim:This study aimed to investigate the clinical value of pre-treatment leukocyte differential counts and theprediction of endometrial cancer using leukocyte markers.

Material and Methods: Medical records of 238 women with pathologically confirmed endometrial cancerbetween March 2000 and June 2009 at two Korean hospitals were reviewed and compared to 596 healthypeople visiting the Health Promotion Center in Gangnam Severance Hospital. For all study subjects, leukocytedifferential counts and CA125 levels in serum obtained prior to operation were recorded. Multiplication ofneutrophil and monocyte (MNM) was determined by multiplying neutrophil and monocyte counts thendividing by 10 000. Differences between endometrial cancer patients and healthy controls were compared. Thesensitivity and specificity for each marker as well as the combined use of CA125 and other leukocyte markerswere assessed using receiver operating characteristic curves.

Results: Mean white blood cell (WBC) counts were 6676 (64406913) cells/mL in endometrial cancer patientscompared to 5663 (55425784) cells/mL in healthy controls (P < 0.001). The area under curve (AUC) for CA125was 0.689 with a sensitivity of 49.13% and specificity of 83.1% using an optimal cut-off value of 18.7 U/mL. TheAUC for MNM was 0.696 with a sensitivity of 62.9% and specificity of 69.1%. The combination of CA125 andMNM showed a higher AUC of 0.760 than use of CA125 or MNM alone.

Conclusion: The combination of MNM and CA125 is a simple and cost-effective method for predictingendometrial cancer.

Key words: CA 125, endometrial cancer, leukocyte, monocyte, neutrophil.

Introduction

Endometrial cancer is the fourth most common malig-

nancy in women and the most common in the femalegenital tract in the USA. 7780 deaths resulting fromendometrial cancer wereexpected, while42 160 womenwere estimated to be newly diagnosed in 2009.1

Vaginal bleeding commonly occurs in the early stageof endometrial cancer. Because patients usually visitmedical institutions during this stage, an early diagno-

sis can be made in most cases. In some cases of endome-trialcancer, however,asymptomatic metastasis is foundin the lymph node or intra-abdominal space, of whichprognosis is commonly poor. Regardless, in cases in

Received: March 23 2011.Accepted: April 27 2011.Reprint request to: Dr Jae-Hoon Kim, Department of Obstetrics and Gynecology, Gangnam Severance Hospital, Yonsei UniversityCollege of Medicine, 146-92 Dogok-Dong, Gangnam-Gu, Seoul 135-720, Korea. Email: jaehoonkim@yuhs.ac

doi:10.1111/j.1447-0756.2011.01694.x J. Obstet. Gynaecol. Res. Vol. 38, No. 1: 4856, January 2012

48 2011 The AuthorsJournal of Obstetrics and Gynaecology Research 2011 Japan Society of Obstetrics and Gynecology

8/12/2019 70118535

2/10

which the presence of endometrial cancer is suspecteddue to the presence of bleeding, it is necessary toperform a non-invasive test prior to invasive tests, suchas an endometrial biopsy.

In high-risk patients, such as breast cancer patientstaking tamoxifen, patients who are severely obese,patients with diabetes mellitus or those with heredi-tary non-polyposis colon cancer (HNPCC) syndrome,a cost-effective screening test may be helpful for earlydetection of endometrial cancer. Sonography andendometrial biopsy are recommended for diagnosis ofendometrial cancer. To date, however, effective sero-logical markers have not been introduced. Elevatedserum CA125 levels have been detected in 1020% ofpatients with early-stage disease and approximately25% of patients with asymptomatic recurrent dis-ease.2,3 Moreover, serum CA125 elevation greater than35 U/mL is not sufficiently accurate to stratify patient

risk and the prognostic cut-off value for CA125 levelsto guide other therapeutic plans is undefined.4,5 Other

biomarkers, including CA 199, M-CSF, insulin andleptin have been studied, but they are less clinicallyeffective.69

It is well established that in cases in which unop-posed estrogen is persistently present, the endome-trium can progress to endometrial cancer due toexcessive proliferation and resultant DNA damage.Recently, inflammation and elevated leukocyte counthave shown proven associations with various can-cers of the lung, breast, ovary, colon and liver.1015

Such cytokines associated with inflammation asCOX-2, PGE2 and NF-kB have been reported at ahigher concentration in endometrial cancer.16 Accord-ing to a prospective study, white blood cell (WBC)counts were relatively higher in the endometrialcancer of postmenopausal women.17 It can therefore

be inferred that the development of endometrialcancer is associated with the inflammatory andimmune responses. Recently, diagnostic and prognos-tic tumor markers using leukocytes subsets have beenreported, with the neutrophil to lymphocyte ratio(NLR) and multiplication of neutrophil and monocytecounts (MNM) being recently introduced in various

cancers.15,1820

The present study aimed to investigate the clinicalvalue of pre-treatment leukocyte differential countsand its diagnostic and prognostic importance inendometrial cancer. We hypothesized that leukocytecount differences between endometrial cancer pati-ents and healthy women could be found, and thesedifferences may be useful indices of diagnostic and

prognostic markers for endometrial cancer, especiallyin combination with CA-125.

Material and Methods

Study population

The present study was carried out in accordance withthe ethical standards of the Helsinki Declaration andwas approved by the Institute of Review Boards(IRBs # 3-2009-0161) of Severance and Gangnam Sev-erance Hospitals. The study involved a retrospectivemedical record review of 238 women with pathologi-cally confirmed endometrial cancer between March2000 and June 2009 at two Korean academic institu-tions. All endometrial cancer patients were surgicallystaged according to the International Federation ofGynecology and Obstetrics (FIGO) staging system.Recurrence status was determined using computed

tomography (CT), magnetic resonance imaging (MRI)and positron emission tomography (PET). Healthycontrols consisted of 596 women visiting the HealthPromotion Center in Gangnam Severance Hospital

between June 2007 and June 2009 with no history ofcancer or gynecological disease. WBC counts andCA125 level for healthy controls were collected fromretrospective medical records. There were no abnor-malities in the laboratory examinations or gynecologi-cal sonography, and participants with a history ofhematological abnormality were excluded.

Hematological analysis

For all study subjects, peripheral venous blood wascollected 3 days before operation for leukocyte differ-ential counts and CA125 levels. While the interval ofendometrial biopsy to staging surgery was less than2 weeks, WBC counts prior to endometrial biopsywere selected because endometrial biopsy can affectWBC count elevation. Leukocyte differential countswere analyzed by the ADIVA 120/2120 Hematologysystem (Bayer Health Care, Diagnostics Division,Tarrytown, NY, USA), and CA125 levels were mea-sured with CA125 II ECLIA (electrochemilumines-

cence immunoassay) on the Roche/Hitachi ModularAnalytics E170 (Roche Diagnostics, Tokyo, Japan).Calculated leukocyte markers were NLR and MNM.The NLR was defined as the absolute neutrophilcount divided by the absolute lymphocyte count.19

The MNM was defined as the multiplication of neu-trophil counts and monocyte counts then divided by10 000.18

Diagnosis of endometrial cancer

2011 The Authors 49Journal of Obstetrics and Gynaecology Research 2011 Japan Society of Obstetrics and Gynecology

8/12/2019 70118535

3/10

Statistical analysis

Data were summarized with number of patients,mean SD and 95% confidence intervals (CI) for themean. Comparisons of leukocyte counts between thetwo study groups were analyzed using independent

t-test and ANOVA post hoc test. To evaluate the oddsratios (OR) for development of endometrial cancer,

binary logistic regression was applied. For constructionof receiver operating characteristic (ROC) curves, apredicted probability for development of endometrialcancer was calculated for each patient using the logisticregression formula. Sensitivity and specificity for eachmarker were assessed using ROC curves, with theresulting area under the curve (AUC) indicatingaverage sensitivity of a marker over the entire ROCcurve. All analyses were performed using SPSS version12.0 (SPSS, Chicago, IL, USA). A P-value of

8/12/2019 70118535

4/10

healthy controls. In endometrial cancer patients,

CA125 was not elevated in 186 (78.1%) of 238 patients.Of these 186 patients who were false negative forCA125, 116 (62.3%) showed a positive MNM cut-offvalue of 100.1, which maximized the sum of sensitivityand specificity.

Discussion

Routine screening methods for endometrial cancer arenot yet proven. Transvaginal sonography and endome-trial biopsy are recommended to diagnose endome-trial cancer in postmenopausal bleeding or high-risk

patients. Hematological marker has not been intro-duced to diagnose endometrial cancer in such patients.CA125 has limitation to diagnose because of sensitivity.However, lessinvasiveand cost-effectivehematologicalmarker could be helpful to diagnose endometrial cancerin such patients. In the current study, we observed thatneutrophil and monocyte counts were elevated inendometrial cancer. The MNM, which was a combina-tion of two parameters, showed a similar AUC ofCA125 alone. However, the combination of CA125and MNM showed higher diagnostic accuracy thansingle markers, CA125 or MNM. CA125 is insufficientto diagnose endometrial cancer due to low sensitivity;

however, MNM was shown to complement CA125.Transvaginal sonography for diagnosis of endome-

trial cancer in postmenopausal women produced9096% sensitivity and 5461% specificity with a 5-mmcut-off value.2123 In the United Kingdom CollaborativeTrial of Ovarian Cancer Screening (UKCTOCS), trans-vaginal sonography to predict endometrial cancershowed 80.5% sensitivity and 85.7% in postmeno-

pausal women with a 5-mm cut-off value.24 Although

this study included endometrial hyperplasia and wasnot designed for endometrial cancer, it showed favor-able sensitivity in postmenopausal women. Biologicalmarker in our study showed lower diagnostic powerthan sonography as a single marker. However, biologi-cal marker could complement the low positive predic-tive value of sonography and be a useful marker fordetecting recurrence depending on markers.

Chronic inflammation has been reported to have arelationship with malignancy of epithelial origin.25

Chronic inflammation featuring elevated WBC isassociated with cancer development and progression

in bladder, cervical, gastric, intestinal, esophageal,ovarian, prostate and thyroid cancers.2630 In gynecologi-cal cancers, ovarian and cervical cancer have revealedelevated whole blood cells counts, including elevatedneutrophil and monocyte counts.15,18,31 Margolis et al.reported that elevated WBC counts were observed inendometrial cancer in postmenopausal women as wellas in breast, colorectal and lung cancer.17 In the currentstudy, meanleukocyte counts, including neutrophilandmonocyte counts, were more elevated in the endome-trial cancergroupthan in healthy controls, as they areinother gynecological cancers.

In cases of colorectal, ovarian and cervical cancer,

NLR and MNM have been reported to be high,15,18,19

as they were in our study. Although the mechanismsby which neutrophil and monocyte counts increasein cancer are not fully understood, tumor cells secretevarious cytokines and chemokines, and these cellsare eventually involved in the proliferation of leuko-cytes, including neutrophils and monocytes. The toxicgranule of neutrophilic cytoplasm attacks the cancer

Table 2 Mean counts of leukocyte subsets, NLR, MNM and combination markers in the study subjects

Mean (95% CI) P-valueControl(n = 596)

Endometrial cancer(n = 238)

Age 50.7 (50.151.4) 52.1 (50.553.2) 0.06

CA125 14.1 (13.314.9) 42.0 (31.152.8)