Sema Rang

Ž .International Journal of Gynecology & Obstetrics 65 1999 273]279

Article

Cellular immune response levels in gestationaltrophoblastic tumor patients after hydatidiform mole

M.T. SutotoU

Department of Obstetrics and Gynecology, Diponegoro Uni ersity, Semarang, Indonesia

Received 3 September 1998; received in revised form 18 January 1999; accepted 21 January 1999

Abstract

Ž .Objecti e: This study looked for an association of natural innate cellular immune responses with the developmentŽ . Ž .of gestational trophoblastic tumors GTT after hydatidiform moles HM . Method: An observational prospective

Ž . Ž . Ž .cohort study 73 patients was carried out to measure the serum level of Natural Killer NK cells, Monocytes Moand a Bio-Merieux skin test was performed on the day of HM evacuation and 4 weeks afterwards. The diagnosis of

Ž .GTT were after Soper and Hammond 1998 and followed up for at least 6 months after evacuation. Patients’characteristics and cellular immune responses among patients who developed GTT were analysed by multiple logistic

Ž . Ž .regression SPSS version 5 . Results: The results showed that a low NK cell percentage was associated Ps0.020Ž .with the development of GTT, also a low Mo Ps0.034 on day 0. While on day 28 a low NK cell percentage

Ž . Ž .Ps0.008 and Mo cell count Ps0.043 were associated with the development of GTT. The cut-off point betweenŽ .the sensitivity 60% and specificity of NK cell percentage on day 0 was 16, and also showed a negative predictive

Ž .value of 83%. Meanwhile on day 28, the cut-off point between the sensitivity and specificity 80% was 17, andŽ .showed a negative predictive value of NK cell percentage 85% . Conclusion: These figures indicated that HM-lowered

NK cell serum percentages were associated with the development of GTT. Q 1999 International Federation ofGynecology and Obstetrics.

Keywords: Hydatidiform mole; Gestational trophoblastic tumor

U Fax: q62 24 317650.

0020-7292r99r$20.00 Q 1999 International Federation of Gynecology and Obstetrics.Ž .PII: S 0 0 2 0 - 7 2 9 2 9 8 0 0 0 2 1 - 1

( )M.T. Sutoto r International Journal of Gynecology & Obstetrics 65 1999 273]279274

1. Introduction

The influence of body immune responses onthe development of trophoblastic diseases havebeen put forward by many experts, such as:

w x1. Fuller et al. 1 who reported that the chorio-carcinoma regression mechanism could notbe distinguished from the allograft responsemechanism;

w x2. Bagshawe 2 reported that resistant chorio-carcinoma sometimes regressed after skintransplantations from a womans’ husband;

w x3. Dent et al. 3 showed that lymphocyte infil-trations in choriocarcinoma tissues influencethe prognosis of this tumor;

w x4. Sugawa et al. 4 found that cellular immunecapacities and serum immune suppressioneffects decrease in trophoblastic diseasepatients, moreover, the changes are differentin each group, namely, in hydatidiform mole,invasive mole and choriocarcinoma.

All these phenomena support the concept ofimmuno-surveillance by Burnette-Thomas.

As Sugawa found that cellular immune capaci-ties are different in each of the ‘trophoblastic

w xdisease stages’ and Lavin 5 reported that cellu-lar immune responses influence the developmentof malignant tumors more than their humoralsides, it is presumed that cellular immune respon-ses will also influence the development of gesta-

w xtional trophoblastic tumors 5 .In 1993 Livingston concluded that although the

adaptive immunity mechanism concept had beenestablished, its practical use had not been com-

w x w xpletely proven 6 . Iaffaioli et al. 7 , in their studyof breast cancer, showed that the CD-4 and CD-8lymphocyte population did not significantly asso-ciate with the prognosis of breast cancer, how-

Ž .ever, they showed that Natural Killer NK cellsas natural components of immunity played more

w xof a role in breast cancer development 7 .This study will look for the association of natu-Ž .ral innate cellular immune responses with the

development of gestational trophoblastic tumorsafter hydatidiform moles.

2. Materials and methods

2.1. Study design

The study was carried out as an observationalprospective cohort investigation, measuring the

Ž .level of Natural Killer NK cells, monocytes inthe serum of patients who suffer from hydatidi-form mole, and a Bio-Merieux skin test on days 0and 28 after evacuation.

The mole patients were managed and moni-tored according to the protocol stated by Soper

w xand Hammond 8 , until at least 24 weeks afterevacuation.

The diagnosis of gestational trophoblastic tu-mors were determined if hCG production in-creased or plateaued on three consecutive mea-surements before a normal level was reachedŽ .normal level: -5 Url , the serum hCG was

Ž .determined by immunoradiometric assay IRMA .

2.2. Subjects

The subjects were hydatidiform mole patientstreated at Dr. Kariadi hospital in Semarang andeight surrounding hospitals, who met the inclu-sion criteria, between October 1 1995 and De-cember 1 1996. The sample size determined by a

w x ŽLemeshaw et al. 9 table was 90 with a power of.80%, and significance of 0.05 , however, because

of limited funds only 73 patients were observedw x9 .

The inclusion criteria were:

1. Complete hydatidiform mole patients, withdiagnosis confirmed by histopathology exami-nation.

2. Aged between 17 and 40 years old, with parityof less than 7.

3. The patients were only surgically treated bycurettage.

4. The patients received no chemotherapy norimmunotherapy, which were not scheduled inthis study.

5. Chemotherapy was only given to patients whodeveloped gestational trophoblastic tumors.

6. The patients were followed up according tothe designed schedule.

( )M.T. Sutoto r International Journal of Gynecology & Obstetrics 65 1999 273]279 275

7. The patients agreed to be admitted into thestudy as indicated by signing the informedconsent form after being informed extensivelyabout the study.

Patients were excluded from the study if:

1. They expressed their wish to be dropped fromthe study with or without reasonrs presented.

2. Became pregnant before the study finished.3. Several days before or during the study the

patient took corticosteroids or non-steroidalanti-inflammatory drugs.

During the study the patients used a simpleŽ .contraceptive method condoms, tissues to pre-

vent pregnancy.

2.2.1. Patient managementAll clinically suspected patients suffering from

hydatidiform mole were evacuated. The speci-mens were examined histopathologically and onlypatients with complete hydatidiform mole wereadmitted into this study. Ultrasound examinationof the abdominal cavity and chest X-ray werecarried out to exclude the possibility of metasta-sis. No patients suffering from distant metastasiswere admitted into the study.

After evacuation, patients were examined every2 weeks to determine whether there were compli-cations or if they had developed into gestationaltrophoblastic tumor.

After the study the patients were grouped intotrophoblastic tumor patients and non-trophoblas-tic ones. They were tabulated and analysed ac-cording to their characteristics and their cellularimmune responses.

2.2.2. Laboratory assaysThe b-hCG assay was carried out by the IRMA

method at the Nuclear Medicine laboratory ofPadjadjaran University, Bandung.

The determination of NK cell serum level wascarried out by indirect immunofluorescence assayfor CD-16 and the monocytes by direct im-munofluorescence assay for CD-14 in the labora-tory of Immunology, Telogorejo Hospital, Se-marang.

The skin test to determine skin delayed reac-tion was done with a Bio-Meriuex skin test, whichcontains: Tetanus 550 Urml; Diphtheria 1 100 000Urml; Streptococcus C 2000 Urml; Tuberculine300 000 Urml; Glycerine 70%; Candida albicans2000 Urml; Trichophyton mentagrophytes 150Urml and Proteus mirabilis 150 Urml as anti-gens. Induration skin diameters were read 48 hafterwards, with a skin calipers.

2.2.3. Statistical analysisPatient’s characteristics and cellular immune

responses between the trophoblastic tumor groupand the non-trophoblastic one were analysed bymultiple logistic regression to show their signifi-cant associations, and their sensitivity and speci-ficity levels as ‘diagnostic tools’ were calculatedwith the Statistical Program for Social Studies

Ž .version 5 SPSS , differences were regarded assignificant if P-0.05.

3. Results

There were 73 patients admitted into the studywho fulfilled the inclusion criteria.

After at least 6 months of follow-up, 17 patients

Table 1Ž . Ž .Characteristics among patients ns73 who developed gestational trophoblastic tumors GTT

Ž . Ž . Ž . Ž .Variable MLRGTT q ns17 GTT y ns56PMean S.D. Mean S.D.

Ž .Age years 27.29 5.47 25.39 5.55 0.045Parity 1.06 1.25 0.93 0.89 0.133

Ž .b-hCG IUrlrday 465 529.35 692 964.49 210 593.72 240 354.72 0.023

Ž . Ž .Notes. GTT q , group of patients who developed gestational trophoblastic tumor; GTT y , group of patients who did notdevelop gestational trophoblastic tumor; MLR, multiple logistic regresion.


developed a gestational trophoblastic tumorŽ .GTT , while 56 patients showed remission.

Among 17 patients who developed GTT, theirages were: 27.29"5.47 years old, while the other

Ž .group with GTT y were: 25.39"0.045, PsŽ .0.045 Table 1 .

Parity among patients who developed GTT wasŽ .1.06"1.25, compared to the GTT y group:

Ž .0.93"0.89, Ps0.133 Table 1 .b-hCG production among patients who devel-

oped GTT was 465 529.35"692 964.49 Url, com-pared to b-hCG production among patients from

Ž .the GTT y group which was 210 593.68"Ž .240 354.72, Ps0.023 Table 1 .

Cellular immune response parameters at day 0Ž .beginning of the study showed:

Ž .v Monocyte percentage among GTT qŽ .patients was 15.40"8.00; among GTT y

Žpatients it was 17.05"9.00, Ps0.034 Table.2 .

Ž .v NK cell percentage among GTT q patientsŽ .was 16.22"10.58; among GTT y patients it

Ž .was 19.89"8.60, Ps0.020 Table 2 .v Skin test induration diameters among GTT

Ž . Ž .q patients was 5.89"4.95 cm; GTT yŽ .patients was 6.17"5.70, Ps0.733 Table 2 .

Among 73 hydatidiform mole patients, thereŽwere 36 patients who were randomly by coin

Table 2Cellular immune responses at day 0 among patients whodeveloped GTT

Variable Tumor development

Ž . Ž .GTT q GTT y MRLŽ . Ž .ns17 ns56 P

Mean S.D. Mean S.D.

Mo 0 15.40 8.00 17.05 9.00 0.034NK 0 16.22 10.58 19.89 8.60 0.020ST 0 5.89 4.95 6.17 5.70 0.733

Abbre¨iations: Mo 0, monocyte percentage at day 0; NK 0,Natural Killer cells percentage at day 0; ST 0, skin test

Ž . Ž .induration diameter cm at day 0; GTT q , group of patientsŽ .who developed gestational trophoblastic tumor; GTT y :

group of patients who did not develop gestational trophoblas-tic tumor; MLR, multiple logistic regression.

Table 3Cellular immune responses at day 28 among patients whodeveloped GTT

Variable Tumor development

Ž . Ž .GTT q GTT y MRLŽ . Ž .ns17 ns56 P

M Sd M Sd

Mo 28 14.35 7.87 14.94 9.81 0.043NK 28 15.92 9.88 22.43 9.97 0.008ST 28 5.35 5.29 7.04 6.38 0.425

Mo 28: monocyte percentage at Day 28; NK 28: Naturalkiller cells percentage at Day 28; ST 28: skin test induration

Ž .diameter cm at Day 28; M: mean; Sd: standard deviation;Ž .GTT q : group of patients who developed gestational tro-

phoblastic tumor;Ž .GTT y : group of patients who did not develop into

gestational trophoblastic tumor; MLR: multiple logistic re-gression.

. Žflipping vaccinated with BCG on day 0 afterserum specimen was taken and a skin test was

.carried out , the other 37 patients were used ascontrols for further study.

Four weeks later serum cellular immune re-sponses were taken again. Cellular immune re-sponse parameters at day 28 showed:

Ž .v The monocyte percentage among GTT qŽ .patients was 14.35"7.87; among GTT y


Ž .v The NK cell percentage among GTT qŽ .patients was 15.92"9.88; among GTT y


v The skin test induration diameter among GTTŽ . Ž .q patients was 5.35"5.29; among GTT y


The above analysis showed that NK cell per-centage was associated more with the develop-ment of GTT after hydatidiform mole than withmonocyte percentage, but the induration skin di-ameter after Bio-Merieux skin test showed nochange in both groups.

The following analysis will describe the sensitiv-ity and specificity of NK cell percentages associ-


Fig. 1. Sensitivity and specificity of hydatiform mole NK cell level developing into GTT at day 0.

Fig. 2. Sensitivity and specificity of post-hydatiform mole NK cell level developing into GTT, 28 days after evacuation.


Žated with the development of GTT Figs. 1 and.2 .These same figures showed the association

between sensitivity and specificity of each post-hydatidiform mole NK level developing GTT at

Ž .days 0 and 28 Figs. 1 and 2 .On day 0, the cut-off point between the sensi-

tivity and specificity curve was at 16% NK cellswhich gave the sensitivity and specificity rate of60%, negative predictive value of 83% and rela-tive risk of 1.9, with confidence interval between0.7 and 3.6.

At day 28, the cut-off point between the sensi-tivity and specificity curve was at 17% of NK cellswhich gave the sensitivity and specificity rate of70%, negative predictive value of 85% and rela-tive risk of 2.7 with confidence interval between

Ž .1.2 and 6.3 Fig. 2 .These figures showed that low NK cell percent-Ž .age -16 on patients who suffered from hydatid-

iform mole had a negative predictive value of83% developing into gestational trophoblastic tu-mor.

4. Discussion

Our references showed that naturalrinnate cel-lular immunity played a main role in the develop-ment of cancer. In blood circulation, NK cells areavailable to do immunosurveillance against can-cer cells, meanwhile monocytes play a bigger rolein tissue immunosurveillance.

As gestational trophoblastic diseases are pre-sumed to be sensitive to the immunity process,this disease can be taken as a model of neoplasmwhich is amenable to the influence of immuno-surveillance, moreover, this disease is easily moni-tored, developed and resolved relatively fastercompared to other neoplasms.

Multiple logistic regression analysis on charac-teristics of hydatidiform mole between the GTT

Ž .group and GTT y group showed that the age ofŽ .patients among GTT q patients is significantly

Ž .higher compared with the GTT y patients, inagreement with the common understanding that

the older the patient the more risk they have fordeveloping gestational trophoblastic tumors.

Analysis on b-hCG production also indicatesthat the higher the b-hCG production the higherrisk that the hydatidiform mole patients developinto gestational trophoblastic tumors.

On day 0, the hydatidiform mole patient’s lowNK cell and monocyte percentage was signifi-cantly associated with the development of GTT.

Ž .At day 28 after hydatidiform mole evacuation ,Ž .the patients’ low NK cell significantly Ps0.08

showed this association. This finding is in agree-ment with the references described above. Theseresults led to a further study to look at the NKcell level in predicting the development of GTTafter hydatidiform mole through diagnostic tests.

Figs. 1 and 2 showed that a serum NK cellpercentage of -16 has a high negative predictive

Ž .value 83% .This study suggests that low hydatidiform mole

patient’s serum NK cell level is associated withthe development of gestational trophoblastic tu-mor, and an NK cell level of -16 also predictedthe development of gestational trophoblastic tu-mor.

References

w x1 Fuller AF, Schiff I, Knapp RS. Immunity, trophoblastand trophoblastic neoplasia. Clin Obstet Gynecol1977:681]690.

w x2 Bagshawe KD. Risk and prognostic factors in tro-phoblastic neoplasia. Cancer 1976;38:1373]1385.

w x3 Dent J, Bagshawe KD. Hydatidiform mole: data from aseries of 5124 cases in the United Kingdom. In: IchinoeK, editor. Trophoblastic disease. Tokyo: Igaku-Shoin,1986:129]141.

w x4 Sugawa T, Umesaki N, Hamada K, Tamura T. A studyof cell mediated immune response in trophoblastic dis-ease: with special experience to the clinical significanceof serum immunosuppressive effect. In: Ichinoe K, edi-tor. Trophoblastic disease. Tokyo: Igaku-Shoin,1986:41]51.

w x5 Lavin PT. Clinical trials in gynecologic oncology: Bios-tatistic issues. In: Knapp RC, Berkowitz RS, editors.Practical gynecologic oncology. 2nd ed. New York: Mc-Graw Hill Inc., 1993:150]175.

w x6 Livingston P. Foundations for therapy of cancer withvaccines. In: Bystryn JC, Ferrone S, Livingston P, edi-


tors. Specific immunotherapy of cancer with vaccines.The New York Academy of Sciences, 1993;690:1]5.

w x7 Iaffaioli RV, Frasci G, Rinaldi R et al. Value of theCD25q, CD16q cell determination in defining the prog-nosis of operable breast cancer patients. Eur GynecolOncol 1993;XIV:412]418.

w x8 Soper JT, Hammond CB. Non-metastatic gestationaltrophoblastic disease. In: Hammond CB, editor. ObstetGynecol Clin of North America, 1988;15:505]519.

w x9 Lemeshaw S, Hosmer DW, Klan J, Lwanga SK. Ade-quacy of sample size in health studies. Chichester:WHO-J, Wiley, 1990:129.

Documents

Sema Rang