VAN TROMMEL, N. E.*,†; NGO DUC, H.*,†; MASSUGER, L. F.A.G.†; SCHIJF, C. P.T.†; SWEEP, C. G.J.*; THOMAS, C. M.G.*,†; THE DUTCH WORKING PARTY ON TROPHOBLASTIC TUMORS
A variety of pathologic subtypes of trophoblastic neoplasms are included in gestational trophoblastic disease, comprising villous malformations, hydatidiform mole—subdivided in complete and partial hydatidiform mole—and nonvillous malformations, of which choriocarcinoma is the most frequent(1). In persistent trophoblastic disease (PTD), trophoblastic activity remains after evacuation of a hydatidiform mole as shown by subsequent unaltered high or even rising serum human chorionic gonadotropin (hCG) concentrations. The reported frequency of PTD is 20% in complete hydatidiform mole(2) and 0.5–9.9% in partial hydatidiform mole(3–6).
In 2000, the FIGO revised their definition of PTD and defined PTD when the plateau of hCG lasts for four measurements over a period of 3 weeks or longer, or when there is a rise of hCG of 3 weekly consecutive measurements or longer over a period of at least 2 weeks(7).
The revised FIGO 2000 criteria are practical and intended for use in a majority of hospitals in the world. Applying the FIGO 2000 criteria is potentially biased by the assessment of the “plateau or rise” criteria that is left to the discretion of each physician(8,9). Inadequate patient compliance or suboptimal clinical resources such as incomplete hCG monitoring may lead to delay in recognition of, and therapy for, PTD(10–12).
Although several authors described normal serum hCG regression corridors after uneventful regression(13–16), not every physician who is involved in the follow-up of a patient with hydatidiform mole can make use of such a normogram because each individual hCG assay requires the construction of its own specific normogram which is not always available.
Our hypothesis is that there is a substantial difference in the calculated ratio from two serum hCG concentrations obtained after evacuation of a mole in patients with PTD versus the ratio in patients without PTD.
The aim of the present study was to develop a procedure less prone to interindividual observer bias than the “plateau or rise” criteria as part of the FIGO 2000 definition for PTD. Thus, we assessed the ability of serum hCG ratios obtained after evacuation of the mole to distinguish patients who will develop PTD from those patients who will not.
Materials and methods
Between 1977 and 2004, 2872 patients were registered with the Dutch Central Registry for Hydatidiform Moles. After informed consent, patients are registered through their referring clinician. We evaluated all patients' data in our database and included all patients with available serum hCG data obtained after evacuation of the mole and assayed in our institution. In retrospect, we classified the hCG regression according to the FIGO 2000 guideline(7) (ie, mola hydatidosa with serum hCG plateauing for three consecutive weeks or rise for two consecutive weekly measurements) and confined all the patients with proven PTD and according to these criteria to the study group (n= 86, 84 of which were treated with chemotherapy or a hysterectomy and 2 were cured by a second curettage only). Those with spontaneous regression of serum hCG after evacuation of the mole according to FIGO 2000 criteria were assigned to the control group (n= 118). Additionally, to be included in the evaluation, the first hCG measurement should have been performed in a serum specimen taken within the first 2 weeks after evacuation of the mole. Reasons for exclusion were (a) the histologic diagnosis of choriocarcinoma for which chemotherapy was initiated without awaiting a plateau or rise in postmolar serum hCG concentrations, or (b) persistence of low hCG serum levels around 10 ng/mL after spontaneous regression. For each patient in the study group, we scored the week that PTD was diagnosed according to FIGO 2000 criteria. The week in which therapy for PTD was started was scored as well, and data were censored from that time point onwards.
A privately developed (“in-house”) radioimmunoassay (RIA) that measured “total” hCG (ie, intact hCG and free ß-subunit, hCG + hCGß) was used exclusively in the authors' laboratory(17). Thus, this assay has been utilized centrally for all measurements in sera sent to the Dutch Central Registry for Hydatidiform Moles. The RIA was calibrated with the World Health Organization third International Standard for intact hCG (WHO third IS hCG 75/537 obtained from the National Institute for Biological Standards, Potters Bar, UK). The measuring range for the standard line of the assay was 1–80 μg/L (0.027–2.14 nmol/L, equivalent to 9.29–743 International Units per Liter (IU/L) of the WHO third IS hCG 75/537)(15). The hCG + hCGß-RIA cross-reacts 100% on a mol/mol basis with intact hCG and 1000% with hCG ß-subunit. Serum hCG concentrations were considered to be normalized if below 2 μg/L (0.053 nmol/L or 18.6 IU/L of the WHO third IS hCG 75/537). The within- and between-assay coefficients of variation (CVw, CVb) for means of duplicate measurements were at a level of 10 μg/L (equivalent to 0.267 nmol/L or 93 IU/L) 7.5% and 10.3%, respectively, and at a level of 56 μg/L (equivalent to 1.50 nmol/L or 520 IU/L), 7.3% and 12%.
For each patient, we calculated the ratio of the hCG concentrations obtained in weeks 1 through 11 after evacuation. The hCG ratio of week 1 after evacuation was calculated by dividing the first by the second available hCG concentration taken in week 1. In week 2, the hCG ratio was calculated by dividing the first available hCG concentration of weeks 1 or 2 by the hCG concentration of a specimen taken in week 2. From week 3 onwards until week 11, the hCG ratios were calculated by dividing the first available hCG concentration of a specimen taken within the first 2 weeks by the hCG concentration of a specimen taken in that particular week (these hCG ratios were designated hCG ratio week 1, 2–11). In the study group, the hCG ratios were excluded from analysis as from the week that the curative therapy (chemotherapy, hysterectomy or a curative second curettage) was initiated or performed. The accuracy to diagnose PTD as derived from the calculated hCG ratios week 1 through 11 was investigated by receiver operating characteristic (ROC) curve analysis of the hCG ratios per week and resulted in calculations of area under curves (AUCs).
To make a direct comparison between the accuracy of both diagnostic tools, we compared for each week the number of patients in the study group identified by FIGO 2000 criteria as having PTD with those identified by the hCG ratio. We correlated, for the FIGO 2000 and the hCG ratio criteria, the cumulative percent rate of identified patients with PTD as a function of the time after evacuation of the mole. All statistical analyses were performed using the SPSS statistical software package version 12.0.1 (SPSS Inc., Chicago, IL).
Ratios obtained from hCG measurements in the study group and the control group were explored in ROC curve analysis to demonstrate the diagnostic ability of the hCG ratio in predicting PTD.
Table 1 depicts the ROC curve analyses of all available hCG ratios of weeks 1 through 11 for the study and control group. For each week analyzed, the number of patients included in the PTD group ranged from 15 to 71 and in the control group from 33 to 89 patients. The AUC of the hCG ratio was 0.568 in week 1 and increased steadily to a plateau greater than or equal to 0.9 by week 5. The calculated sensitivities at 95% specificity increased from 20% in week 1 to 79% in week 5 and declined to 48–52% in weeks 10–11. The cutoff values represent the hCG ratios below which PTD can be diagnosed with 95% specificity in a given week after evacuation.
Figure 1 shows the ROC analyses from the hCG ratios of weeks 1, 5, and 10. To compare the diagnostic accuracy of the two diagnostic tools, we conducted a cross-sectional analysis in which we scored only the number of patients (ie, the study group) that were identified with either test as having PTD and expressed these as a proportion of the total number of PTD patients with an hCG ratio in that week.
Patients of the study group were censored in the week that PTD was identified by means of an hCG ratio that was below the cutoff value at that week. The study group fulfilled the criteria for the diagnosis of PTD at a median time point of 4.7 weeks when using FIGO 2000 criteria, and this was 3.0 weeks in the case that the hCG ratio was used.
In Table 2, we show at weekly intervals the proportion of patients with a diagnosis of PTD as identified with FIGO 2000 criteria as well as with the hCG ratio. FIGO 2000 criteria identified 0 out of 15 (0%) patients in week 1, while the hCG ratio of week 1 detected three patients (20%) to have PTD. In week 3, the FIGO 2000 criteria identified 11 (16%) and the hCG ratio 35 patients (52%) out of 67 patients with PTD. In week 7, FIGO 2000 criteria identified 70% (30/43) and the hCG ratio 32 patients (74%) out of 43 patients with PTD. Afterward, the diagnostic ability of FIGO 2000 criteria increased steadily up to 89% in week 11, whereas the hCG ratios detected 48% and 63% of PTD patients in weeks 10 and 11, respectively, after evacuation of the mole. The main reason why the sensitivity of the hCG ratio decreased after week 7 was that most of the PTD patients (74%) were then identified as having PTD.
Figure 2 depicts the cumulative number of patients with PTD once identified by FIGO 2000 criteria, and by the hCG ratio of the corresponding week.
Out of the 86 patients with PTD, FIGO 2000 criteria cumulatively identified 0 patients in week 1 and 12 patients in week 3 (14%) to reach 60% by week 5, and more than 90% in weeks 10–11 after evacuation. According to the FIGO 2000 criteria, less than 10% of 86 patients in the study group could not be identified as having PTD in the first 11 weeks after the evacuation date.
Out of the 86 patients with PTD, the hCG ratios of week 1 already identified 3.5% (three cases) of the patients with PTD, 51% in week 3 (44 cases), while the maximum cumulative detection rate of the hCG ratio (83%, 71 cases) was already reached at week 8. About 7% of 86 PTD patients could not be identified by the hCG ratio in the first 11 weeks after evacuation.
The aim of our retrospective study was to explore the value of serum hCG concentration ratios to distinguish patients who will develop PTD from patients who will have spontaneous regression of serum hCG after evacuation of a hydatidiform mole, and to compare the diagnostic potential of these hCG ratios with the revised FIGO 2000 criteria for the diagnosis of postmole hydatidiform mole trophoblast neoplasia(7). To our knowledge, no studies have been performed on the potential of hCG ratios in the diagnosis of PTD. Ratios obtained from hCG measurements in the study group and the control group were explored in ROC curve analysis and various AUCs thereof, were compared for diagnostic accuracy in predicting PTD. The hCG ratios obtained in week 1 through week 11 displayed an increasing diagnostic potential from week 1 onwards until a plateau was reached in week 5 after evacuation. For each week after evacuation, a cutoff value for this ratio was calculated, below which PTD can be diagnosed at 95% specificity. The AUC of the hCG ratios increased from 0.568 in week 1 to 0.935 in week 5 and stabilized at AUCs of greater than or equal to 0.9 afterward. The hCG ratio and FIGO 2000 criteria identified 52% and only 16%, respectively, of patients to have PTD in week 3 after evacuation. In week 7, both the hCG ratio and the FIGO 2000 criteria identified greater than or equal to 70% of patients with PTD.
To compare the potential of the two diagnostic tools in this study, we analyzed for each patient in the study group at which time point after the evacuation the criteria for PTD were met according to the internationally accepted FIGO 2000 criteria and to the hCG ratio. We found that the median time to reach the diagnosis of PTD in the study group was 4.7 weeks when using FIGO 2000 criteria for PTD, whereas the median time to diagnose PTD by using the hCG ratio was 3.0 weeks. Thus, the hCG ratio identified patients earlier than FIGO 2000 criteria. According to the FIGO 2000, the individual assessment of “plateau or rise” to confirm or reject the diagnosis of PTD after evacuation is presently left to the discretion of each physician, which is subject to inter- and intraobserver bias.
To add a quantitative measure to the diagnosis of PTD, several normograms for spontaneous regression of serum hCG have been developed(13–16). The use of these normograms has been advocated since their use may be an aid to prevent overtreatment of patients(13), but bears the disadvantage that each individual hCG assay requires the construction of its own specific normogram. This is because hCG in blood is not present as one clear distinct entity but as a multifactorial complex of heavily glycosylated proteins. As the different commercially available hCG assays use different sets of (monoclonal and/or polyclonal) antibodies to bind and detect hCG molecules, all with their own specific characteristics (specificity and affinity), it theoretically is not possible to calibrate these assays using one common standard or hCG reference preparation, as all assays differently will recognize the peptides. Moreover, the reference preparation obligatory will differ from hCG-like immunoreactivity present in blood.
Another qualitative, though arbitrary, addition to the definition of the FIGO 2000 criteria for PTD was put forward by Kohorn(18) who proposed to define a “rise” of hCG as an increase in terms of 10% or more for three values or more over a period of at least 2 weeks. However, up till now, no universal agreement has been reached on this proposal(19).
It is still unclear why some hydatidiform moles develop into PTD while others show a spontaneous regression. It is not very likely that residual tissue after evacuation is the only cause to develop PTD. Normally, extravillous trophoblasts invade the myometrium up to one third of the entire uterine wall(20). With the commonly used suction curettage for the evacuation of a hydatidiform mole, it is highly unlikely that this molar tissue bearing the extravillous trophoblasts will be completely removed. The study by Lao et al.(21) compared the “histology result” (trophoblastic tissue, or no trophoblastic tissue) of a routinely performed second curettage in patients with a hydatidiform mole with the subsequent need for chemotherapy for PTD. The authors reported absence of such a correlation, meaning that this result contradicts the hypothesis that residual tissue after evacuation of a hydatidiform mole is the entirely unique cause of PTD.
The search for genetic factors that can identify hydatidiform moles at risk to develop into PTD is continuing(22). Recently, several cell growth regulating factors like MCL-1, EGFR, c-erbB-2 and c-erbB-3, c-ras, Nm-23, p53, cyclin E, and telomerase(23–27) were found to have different expression patterns in hydatidiform mole, which will develop PTD as compared to hydatidiform mole with spontaneous regression of serum hCG. This may lead to the development of new biochemical tools in the future that will aid the diagnosis of patients who will develop PTD. At present, no such tools are available.
Although remission is safely ensured by the current recommendations for hCG follow-up, the advised monitoring can sometimes be suboptimal due to poor compliance, as in most cases long-term follow-up of serum hCG levels is required to rule out PTD. In the present study, we demonstrate that we can identify PTD already in the first weeks after evacuation in a substantial proportion of patients by using serum hCG concentrations from which we calculated hCG ratios.
In this retrospective study, we found that the ratio of two serum hCG measurements identified more than 50% of patients with PTD at the 95% specificity level in week 3 after evacuation, whereas FIGO 2000 identified not more than 16% of patients with PTD at that week. With an hCG ratio, we were able to identify up to 79% of patients by week 5 after evacuation at the 95% specificity level.
The findings in this study provided evidence for the potential diagnostic ability of the hCG ratio, but direct clinical conclusions cannot be drawn. The hCG ratio may be helpful in detecting PTD earlier than the “plateau or rise” criteria of FIGO 2000. Further prospective study should be explored for the clinical applicability of the hCG ratio in the definition for PTD.
The authors thank Henk Ariaens and Rob van der Steen, RUNMC Department of Chemical Endocrinology, for expert technical assistance and the members of the Dutch Working Party on Trophoblastic Tumours for their continuous support in gathering the clinical data.
1 Bentley RC. Pathology of gestational trophoblastic disease. Clin Obstet Gynecol
2 Berkowitz RS, Goldstein DP. Gestational trophoblastic disease. Cancer
3 Bagshawe KD, Lawler SD, Paradinas FJ, Dent J, Brown P, Boxer GM. Gestational trophoblastic tumours following initial diagnosis of partial hydatidiform mole. Lancet
4 Berkowitz RS, Goldstein DP, Bernstein MR. Natural history of partial molar pregnancy. Obstet Gynecol
5 Goto S, Yamada A, Ishizuka T, Tomoda Y. Development of postmolar trophoblastic disease after partial molar pregnancy. Gynecol Oncol
6 Rice LW, Berkowitz RS, Lage JM, Goldstein DP, Bernstein MR. Persistent gestational trophoblastic tumor after partial hydatidiform mole. Gynecol Oncol
7 FIGO Oncology Committee. FIGO staging for gestational trophoblastic neoplasia 2000. Int J Gynaecol Obstet
8 Kohorn EI, Goldstein DP, Hancock BW et al. Workshop report: combining the staging system of the International Federation of Gynecology and Obstetrics with the scoring system of the World Heath Organization for trophoblastic neoplasia. Report of the working committee of the International Society for the Study of Trophoblastic Disease and the International Gynecologic Cancer Society. Int J Gynecol Cancer
9 Ngan HY. The practicability of FIGO 2000 staging for gestational trophoblastic neoplasia. Int J Gynecol Cancer
10 Massad LS, Abu-Rustum NR, Lee SS, Renta V. Poor compliance with postmolar surveillance and treatment protocols by indigent women. Obstet Gynecol
11 Allen JE, King MR, Farrar DF, Miller DS, Schorge JO. Postmolar surveillance at a trophoblastic disease center that serves indigent women. Am J Obstet Gynecol
12 Kim DS, Moon H, Kim KT, Moon YJ, Hwang YY. Effects of prophylactic chemotherapy for persistent trophoblastic disease in patients with complete hydatidiform mole. Obstet Gynecol
13 Rotmensch J, Rosenshein NB, Block BS. Comparison of human chorionic gonadotropin regression in molar pregnancies and post-molar nonmetastatic gestational trophoblastic neoplasia. Gynecol Oncol
14 Schlaerth JB, Morrow CP, Kletzky OA, Nalick RH, D'Ablaing GA. Prognostic characteristics of serum human chorionic gonadotropin titer regression following molar pregnancy. Obstet Gynecol
15 Shigematsu T, Kamura T, Saito T, Kaku T, Nakano H, Kinugawa N. Identification of persistent trophoblastic diseases based on a human chorionic gonadotropin regression curve by means of a stepwise piecewise linear regression analysis after the evacuation of uneventful moles. Gynecol Oncol
16 Behtash N, Ghaemmaghami F, Honar H et al. Is normal beta-hCG regression curve helpful in the diagnosis of persistent trophoblastic disease? Int J Gynecol Cancer
17 Thomas CM, Segers MF, Houx PC. Comparison of the analytical characteristics and clinical usefulness in tumour monitoring of fifteen hCG (-beta) immunoassay kits. Ann Clin Biochem
18 Kohorn EI. The new FIGO 2000 staging and risk factor scoring system for gestational trophoblastic disease: description and critical assessment. Int J Gynecol Cancer
19 Ng TY, Wong LC. Diagnosis and management of gestational trophoblastic neoplasia. Best Pract Res Clin Obstet Gynaecol
20 Huppertz B, Hemmings D, Renaud SJ, Bulmer JN, Dash P, Chamley LW. Extravillous trophoblast apoptosis—a workshop report. Placenta
21 Lao TT, Lee FH, Yeung SS. Repeat curettage after evacuation of hydatidiform mole. An appraisal. Acta Obstet Gynecol Scand
23 Cheung AN, Zhang DK, Liu Y, Ngan HY, Shen DH, Tsao SW. Telomerase activity in gestational trophoblastic disease. J Clin Pathol
24 Fong PY, Xue WC, Ngan HY et al. Mcl-1 expression in gestational trophoblastic disease correlates with clinical outcome: a differential expression study. Cancer
25 Fukunaga M. Immunohistochemical characterization of p57(KIP2) expression in early hydatidiform moles. Hum Pathol
26 Tuncer ZS, Vegh GL, Fulop V, Genest DR, Mok SC, Berkowitz RS. Expression of epidermal growth factor receptor-related family products in gestational trophoblastic diseases and normal placenta and its relationship with development of postmolar tumor. Gynecol Oncol
27 Yang X, Zhang Z, Jia C, Li J, Yin L, Jiang S. The relationship between expression of c-ras, c-erbB-2, nm23, and p53 gene products and development of trophoblastic tumor and their predictive significance for the malignant transformation of complete hydatidiform mole. Gynecol Oncol
Copyright © 2008 Blackwell Publishing Ltd.