Advertisement

Ultrasound imaging after evacuation as an adjunct to β-hCG monitoring in posthydatidiform molar gestational trophoblastic neoplasia

Published:February 09, 2009DOI:https://doi.org/10.1016/j.ajog.2008.11.032

      Objective

      The purpose of this study was to identify prognostic factors associated with development of gestational trophoblastic neoplasia (GTN) after hydatidiform mole (HM).

      Study Design

      A retrospective analysis of 189 patients with HM was performed. We recorded features such as maternal age, HM history, blood group, gestational age, uterine volume at evacuation, presence of theca lutein cysts, vaginal bleeding, and transvaginal ultrasonography with color Doppler imaging. We considered risk predictors to be the presence of nodules and hypervascularization within the myometrium or endometrium (positive ultrasound imaging). An univariate and multivariate analysis, with the COX nominal logistic model, was performed.

      Results

      Fourteen patients experienced GTN (7.4%). After univariate analysis, uterine size (P = .0139) and positive ultrasound results (P < .0001) were associated significantly with GTN development. At multivariate analysis, only positive ultrasound results maintained significance (likelihood ratio test: χ2 = 0.0000).

      Conclusion

      The risk of GTN is increased in patients with uterine involvement that is assessed by ultrasound imaging. None of the other prognostic factors that were evaluated was predictive of GTN development.

      Key words

      Gestational trophoblastic neoplasia (GTN) represents a rare complication of pregnancy. It can develop after a full-term delivery, a spontaneous miscarriage, or a termination of pregnancy; however, the risk of occurrence is 2000-fold greater after hydatidiform mole.
      • Berkowitz R.S.
      • Goldstein D.P.
      Presentation and management of molar pregnancy.
      • Seckl M.J.
      • Fisher R.A.
      • Salerno G.
      • et al.
      Choriocarcinoma and partial hydatidiform moles.
      • Ngan S.
      • Seckl M.J.
      Gestational trophoblastic neoplasia management: an update.
      It was estimated that after molar evacuation 6-36% of patients progress to GTN.
      • Morrow C.P.
      Postmolar trophoblastic disease: diagnosis, management, and prognosis.
      • Kim S.J.
      • Na Y.J.
      • Jung S.G.
      • Kim C.J.
      • Bae S.N.
      • Lee C.
      Management of high-risk hydatidiform mole and persistent gestational trophoblastic neoplasia: the Korean experience.
      • Goldstein D.P.
      • Berkowitz R.S.
      • Bernstein M.R.
      Management of molar pregnancy.
      • Lurain J.R.
      • Brewer J.I.
      • Torok E.E.
      • Halpern B.
      Natural history of hydatidiform mole after primary evacuation.
      • Niemann I.
      • Petersen L.K.
      • Hansen E.S.
      • Sunde L.
      Predictors of low risk of persistent trophoblastic disease in molar pregnancies.
      For Editors' Commentary, see Table of Contents
      GTN is defined on the basis of a clinically invasive disease or in presence of persistently elevated serum human chorionic gonadotrophin (β-hCG) concentration after diagnosis of molar pregnancy. Current treatment of GTN is effective; however, occasionally, progression and death can occur in patients with metastatic disease.
      • Berkowitz R.S.
      • Goldstein D.P.
      Chorionic tumors.
      The importance of prognostic factor identification in hydatidiform mole has been reported in several previous studies.
      • Tow W.S.
      The influence of the primary treatment of hydatidiform mole on its subsequent course.
      • Tsukamoto N.
      • Iwasaka T.
      • Kashimura Y.
      • Uchino H.
      • Kashimura M.
      • Matsuyama T.
      Gestational trophoblastic disease in women aged 50 or more.
      • Xia Z.F.
      • Song H.Z.
      • Tang M.Y.
      Risk of malignancy and prognosis using a provisional scoring system in hydatidiform mole.
      • Ayhan A.
      • Tuncer Z.S.
      • Halilzade H.
      • Küçükali T.
      Predictors of persistent disease in women with complete hydatidiform mole.
      • Stone M.
      • Bagshawe K.D.
      An analysis of the influences of maternal age, gestational age, contraceptive method, and the mode of primary treatment of patients with hydatidiform moles on the incidence of subsequent chemotherapy.
      • Feltmate C.M.
      • Growdon W.B.
      • Wolfberg A.J.
      • et al.
      Clinical characteristics of persistent gestational trophoblastic neoplasia after partial hydatidiform molar pregnancy.
      • Berkowitz R.S.
      • Bernstein M.R.
      • Laborde O.
      • Goldstein D.P.
      Subsequent pregnancy experience in patients with gestational trophoblastic disease: New England Trophoblastic Disease Center, 1965-1992.
      • Seckl M.J.
      • Dhillon T.
      • Dancey G.
      • et al.
      Increased gestational age at evacuation of a complete hydatidiform mole: does it correlate with increased risk of requiring chemotherapy?.
      • Parazzini F.
      • Mangili G.
      • Belloni C.
      • La Vecchia C.
      • Liati P.
      • Marabini R.
      The problem of identification of prognostic factors for persistent trophoblastic disease.
      • Parazzini F.
      • La Vecchia C.
      • Franceschi S.
      • et al.
      ABO blood-groups and the risk of gestational trophoblastic disease.
      • Curry S.L.
      • Hammond C.B.
      • Tyrey L.
      • Creasman W.T.
      • Parker R.T.
      Hydatidiform mole: diagnosis, management, and long-term follow-up of 347 patients.
      • Berkowitz R.S.
      • Goldstein D.P.
      • DuBeshter B.
      • Bernstein M.R.
      Management of complete molar pregnancy.
      In fact, it appears useful for subgroup patients in different classes according to the risk of the development of GTN, thus permitting an individualization of the follow-up and eventual treatment.
      However, features of the high-risk population are still controversial, and literature reports discrepant results on the effectiveness of different prognostic factors. At present, β-hCG monitoring remains the standard method of follow-up in patients with hydatidiform mole.
      • Berkowitz R.
      • Ozturk M.
      • Goldstein D.
      • Bernstein M.
      • Hill L.
      • Wands J.R.
      Human chorionic gonadotropin and free subunits' serum levels in patients with partial and complete hydatidiform moles.
      In this study, we evaluated clinical prognostic factors in a series of 189 patients with hydatidiform mole, and we present an effective ultrasound-based method of selecting those high-risk cases that need closer follow-up evaluation.

      Materials and Methods

      This is a study of 189 patients who were affected by hydatidiform mole and followed in our outpatient clinic for gestational trophoblastic disease in San Raffaele Hospital between 1992 and 2004. All cases that were referred to our center for persistence of β-hCG levels and chemotherapy were excluded.
      Surgical evacuation was performed in 186 patients; 2 patients with ectopic molar pregnancy were submitted to laparoscopic salpingectomy, and 1 patient underwent hysterectomy. All patients had a confirmed histologic diagnosis.
      Follow-up was started 21 days after surgery. At that time, the patient's history, pelvic examination, β-hCG measurement, and transvaginal ultrasound findings were recorded. Subsequent controls were set weekly (with β-hCG assay) until β-hCG levels decreased below 5 mUI/mL, then 2 further specimens were obtained at weekly intervals. Afterwards, the patient was tested monthly for 6-12 months. Transvaginal ultrasound scanning was used to investigate presence of theca lutein cysts, retained trophoblastic tissue, and myometrial nodules 3 weeks after evacuation. The uterus was scanned in several longitudinal and transverse planes, and a careful search for areas with abnormal echoes was made. The morphologic condition of the uterus was considered to be abnormal (positive ultrasound) when hyperechoic lesions were identified within the myometrium or when power Doppler imaging showed increased signal that suggested an increased vascularization within the endometrium or the myometrium. An color-Doppler ultrasound machine (Eidos; Esaote, Genova, Italy) was used with an endocavitary probe of 6.5 MHz frequency. The pulsed repetition frequency ranged between 1000 and 5000 MHz, power Doppler imaging gains were set at average level, and a wall filter of 100 MHz was used. Areas with either increased echogenicity or increased vascularization were identified, measured, and enumerated; their localization was classified according to the side and anterior or posterior wall of the uterus.
      Cystic ovarian masses > 30 mm were considered compatible with theca lutein cysts. Ultrasonography was repeated at fixed intervals until complete resolution in patients with uterine involvement. In the other cases, ultrasound scanning was repeated at the end of follow-up or when there was an alteration in the normal decline of β-hCG (increment, plateau, or slow decline).
      After 2000, GTN was defined according to FIGO 2000
      • Benedet J.L.
      • Bender H.
      • Jones 3rd, H.
      • Ngan H.Y.
      • Pecorelli S.
      FIGO staging, classification, and clinical practice guidelines in the management of gynecologic cancers: FIGO Committee on Gynecologic Oncology.
      in the presence of 1 of the following criteria: serum β-hCG at > 20000 U/L > 3 weeks after evacuation, plateau of β-hCG values (values of persistently elevated β-hCG in days 1, 7, 14, and ≥ 21) sequential rise of β-hCG for ≥ 2 weeks or β-hCG that was persistent 4-6 months after evacuation. Before 2000, diagnosis was based according to criteria similar to those of Stone and Bagshawe (in Parazzini et al).
      • Parazzini F.
      • Mangili G.
      • Belloni C.
      • La Vecchia C.
      • Liati P.
      • Marabini R.
      The problem of identification of prognostic factors for persistent trophoblastic disease.
      We considered the possible risk factors of persistent disease: previous molar pregnancy, maternal age > 40 years, gestational age at evacuation, blood group A or AB, large-for-date uterus, theca lutein cysts, ultrasonographic myometrial involvement, and persistence of vaginal bleeding after evacuation.
      Data were analyzed by the use of the statistical software (JMP, version 5.1; SAS Institute Inc, Cary, NC). A univariate analysis was performed, and we considered a correlation that was expressed by a Pearson index of < 0.05 (P < .05) to be significant. We performed a multivariate analysis using the COX nominal logistic model and the likelihood ratio test.

      Results

      We included 189 patients of whom 14 women (7.4%) experienced GTN and required chemotherapy.
      Maternal age at diagnosis was analyzed, and we grouped the patients in 4 clusters. Thus, the proportion of patients with GTN, if the maternal age was less than 24, 25-34, 35-39, or > 40 years old, was 14.3%, 50,0%, 14.3%, and 21.4%, respectively. In those patients who did not require treatment, the proportion was 6.9%, 64.5%, 21.4%, and 7.4%, respectively. The difference of maternal age between patients with GTN and patients with spontaneous resolution was not statistically different (P = .19), even if we consider a risk factor to be only maternal age > 40 years. In this case, the Pearson index was lower (P = .07) and showed a trend towards significance.
      The mean value of gestational age at evacuation is 10 weeks (range, 6-18 weeks of gestation). We considered a potential risk factor to be a gestational age > 12 weeks. Univariate analysis did not show a difference between the 2 groups (P = .59).
      A previous molar pregnancy was reported in only 1 patient of those who needed treatment (7.1%) and in 6 of all the other patients (3.43%). The difference is not statistically different (P = .42). Also, A or AB blood groups are not related with a different risk of the development of GTN (P = .61).
      A greater uterine size for gestational age was reported in 46 of patients (26.3%) with spontaneous resolution and in 8 patients (57.1%) who needed chemotherapy treatment. The incidence is significantly different with a probability value of .013.
      Theca lutein cysts and persistence of vaginal bleeding did not present a statistically significant difference in the 2 groups of patients (P = .078 and .0615).
      The presence of endometrial or myometrial involvement at the first ultrasound examination was observed in 53.5% and 3.5% of patients with GTN and spontaneous resolution, respectively. The difference is statistically significant with a Pearson index < 0.0001. The accuracy of sonographic examination for detection of molar pregnancy complicated by GTN presents a high specificity (96.5%) but a low sensitivity (53.9%). FIGURE 1, FIGURE 2 show an hyperechoic myometrial nodule at transvaginal sonography and at color Doppler examination.
      Figure thumbnail gr1
      FIGURE 1Myometrial nodule at ultrasound
      Garavaglia. Ultrasound imaging after evacuation as an adjunct to β-hCG monitoring in posthydatidiform molar GTN. Am J Obstet Gynecol 2009.
      Figure thumbnail gr2
      FIGURE 2Myometrial nodule at Doppler ultrasound
      Garavaglia. Ultrasound imaging after evacuation as an adjunct to β-hCG monitoring in posthydatidiform molar GTN. Am J Obstet Gynecol 2009.
      Univariate statistic results are shown in Figure 3 and summarized in Table 1.
      Figure thumbnail gr3
      FIGURE 3Prognostic factors in GTN development
      Garavaglia. Ultrasound imaging after evacuation as an adjunct to β-hCG monitoring in posthydatidiform molar GTN. Am J Obstet Gynecol 2009.
      TABLE 1Univariate analysis of clinical prognostic factors of hydatidiform mole
      Prognostic factorSpontaneous resolution (%)GTN (%)P value
      Age > 40 y7.421.4.0702
      Increased uterine size26.357.1.0139
      Thecalutein cysts12.028.6.0782
      Persistent vaginal bleeding16.035.7.0615
      ABO blood group53.546.1.615
      History of molar pregnancy3.47.4.4789
      Gestational age > 12 wk15.88.3.59
      Positive transvaginal ultrasonography3.553.9< .0001
      GTN, gestational trophoblastic neoplasia.
      Garavaglia. Ultrasound imaging after evacuation as an adjunct to β-hCG monitoring in posthydatidiform molar GTN. Am J Obstet Gynecol 2009.
      To identify significant prognostic factors, we performed a multivariate analysis using the nominal logistic model; the results are reported in Table 2. The only prognostic factor that was confirmed to be significant is the presence of uterine positive ultrasound abnormalities with a likelihood ratio test X2 < 0.0001.
      TABLE 2Multivariate analysis of prognostic factors of hydatidiform mole (likelihood ratio test)
      Prognostic factorLikelihood ratio
      χ2 test.
      P value
      χ2 test.
      Age > 40 y0.18475559.6673
      Increased uterine size3.15323927.0758
      Thecalutein cysts0.01828178.8924
      Persistent vaginal bleeding0.32614973.5679
      Positive transvaginal ultrasonography17.5656868< .0001
      ABO blood group0.93835467.3327
      History of molar pregnancy0.00008822.9925
      Garavaglia. Ultrasound imaging after evacuation as an adjunct to β-hCG monitoring in posthydatidiform molar GTN. Am J Obstet Gynecol 2009.
      a χ2 test.

      Comment

      The importance of identifying the potential risk factors for the development of GTN is due to the necessity of the submission of only some patients to a close clinical and laboratory follow-up.
      Tow
      • Tow W.S.
      The influence of the primary treatment of hydatidiform mole on its subsequent course.
      and Tsukamoto et al
      • Tsukamoto N.
      • Iwasaka T.
      • Kashimura Y.
      • Uchino H.
      • Kashimura M.
      • Matsuyama T.
      Gestational trophoblastic disease in women aged 50 or more.
      reported an incidence of GTN of 37% and 56%, respectively, in patients with hydatidiform mole who were > 40 or 50 years. The influence of maternal age at diagnosis was reported in a more recent investigation.
      • Berkowitz R.S.
      • Goldstein D.P.
      Chorionic tumors.
      • Xia Z.F.
      • Song H.Z.
      • Tang M.Y.
      Risk of malignancy and prognosis using a provisional scoring system in hydatidiform mole.
      • Ayhan A.
      • Tuncer Z.S.
      • Halilzade H.
      • Küçükali T.
      Predictors of persistent disease in women with complete hydatidiform mole.
      • Stone M.
      • Bagshawe K.D.
      An analysis of the influences of maternal age, gestational age, contraceptive method, and the mode of primary treatment of patients with hydatidiform moles on the incidence of subsequent chemotherapy.
      An interesting study of Feltmate et al
      • Feltmate C.M.
      • Growdon W.B.
      • Wolfberg A.J.
      • et al.
      Clinical characteristics of persistent gestational trophoblastic neoplasia after partial hydatidiform molar pregnancy.
      analyzed the presence of predictive factors for persistent disease in 2 groups of patients: patients who were registered between 1973 and 1989 and patients who were registered between 1990 and 2003. In the first group, 2 prognostic factors were identified: maternal age and history of hydatidiform mole. In the recent group, there was no significant difference between patients with spontaneous resolution and patients who needed chemotherapy. In our study, maternal age was not confirmed as a prognostic factor, probably because of the limited number of patients and because the Pearson index approximated statistical significance (P = .07).
      A previous hydatidiform mole was reported to be associated with an increased risk of the development of GTN.
      • Berkowitz R.S.
      • Goldstein D.P.
      Chorionic tumors.
      • Ayhan A.
      • Tuncer Z.S.
      • Halilzade H.
      • Küçükali T.
      Predictors of persistent disease in women with complete hydatidiform mole.
      • Berkowitz R.S.
      • Bernstein M.R.
      • Laborde O.
      • Goldstein D.P.
      Subsequent pregnancy experience in patients with gestational trophoblastic disease: New England Trophoblastic Disease Center, 1965-1992.
      In our series, no influence of this factor was observed.
      Unlike other neoplastic diseases in which late treatment of the lesion might worsen the outcome, in our series this event did not occur. In agreement with our findings, a previous study of Seckl et al
      • Seckl M.J.
      • Dhillon T.
      • Dancey G.
      • et al.
      Increased gestational age at evacuation of a complete hydatidiform mole: does it correlate with increased risk of requiring chemotherapy?.
      demonstrated that there is not evidence that delayed evacuation of molar miscarriage influences the prognosis. Our study confirms this result; patients who underwent evacuation after the 12 weeks of gestation did not have an increased risk of requiring chemotherapy.
      Several investigators observed that persistence of vaginal bleeding and some associations of ABO blood groups were more common in patients who experienced persistent trophoblastic disease. Our study found no consistent influence in malignancy onset.
      • Parazzini F.
      • Mangili G.
      • Belloni C.
      • La Vecchia C.
      • Liati P.
      • Marabini R.
      The problem of identification of prognostic factors for persistent trophoblastic disease.
      • Parazzini F.
      • La Vecchia C.
      • Franceschi S.
      • et al.
      ABO blood-groups and the risk of gestational trophoblastic disease.
      Uterine size greater than gestational age and theca lutein cysts were reported to be risk factors in several studies
      • Morrow C.P.
      Postmolar trophoblastic disease: diagnosis, management, and prognosis.
      • Curry S.L.
      • Hammond C.B.
      • Tyrey L.
      • Creasman W.T.
      • Parker R.T.
      Hydatidiform mole: diagnosis, management, and long-term follow-up of 347 patients.
      • Berkowitz R.S.
      • Goldstein D.P.
      • DuBeshter B.
      • Bernstein M.R.
      Management of complete molar pregnancy.
      as being indirect signs of marked trophoblastic proliferation. In our patients, only large-for-date uterus maintains statistical significance in univariate analysis (P = .0132); ovarian masses are not associated with the increased incidence of GTN.
      In the last 20 years, mainly because of the ultrasound introduction, the clinical presentation of hydatidiform mole has changed.
      • Mangili G.
      • Garavaglia E.
      • Cavoretto P.
      • Gentile C.
      • Scarfone G.
      • Rabaiotti E.
      Clinical presentation of hydatidiform mole in northern Italy: has it changed in the last 20 years?.
      This can explain the reason that features that previously were identified as predictive of persistence disease are not confirmed in recent series.
      • Feltmate C.M.
      • Growdon W.B.
      • Wolfberg A.J.
      • et al.
      Clinical characteristics of persistent gestational trophoblastic neoplasia after partial hydatidiform molar pregnancy.
      In our study, ultrasound evaluation had a fundamental role in the recognition of patients with a high probability of the development of persistence. In fact, 54% of patients with positive ultrasound results subsequently had GTN. In patients with spontaneous remission, < 4% received positive ultrasound results during follow-up. The difference is highly statistically significant (P < .0001). Interestingly, those patients with positive ultrasound results and without experiencing GTN showed a slow declining of β-hCG, therefore requiring intensive follow-up. Multivariate analysis demonstrates that the prediction of GTN can be based on the ultrasound report. The absence of uterine disease could not be considered to be a protecting factor because of the low sensibility of the examination.
      Volume of myometrial nodules was variable; in our patients, the diameter range was 15-40 mm. Serial ultrasound examinations were able to identify modifications of nodule size and echogenic pattern.
      We hypothesize that GTN onset after hydatidiform mole is associated with myometrial involvement and eventual subsequent evolution in invasive mole, which can be considered the first step before progression and distant metastases. Myometrial nodules can also appear during follow-up, and we think that this event can also provide a negative prognostic impact. However, in this study, only ultrasound evaluation that was executed after 21 days from evacuation was considered.
      The early recognition of GTN clinical predictors allows better treatment of patients by stratifying them in 2 risk groups: high-risk patients who must be monitored closely with β-hCG measurements and ultrasound examinations and low-risk patients who can be reassured at the time of counseling and who can be observed for a shorter period of time. In fact, hydatidiform mole is a disease with an important emotional impact; the majority of patients are young and asymptomatic and desire pregnancy. During the whole length of β-hCG surveillance, patients must avoid pregnancies. Therefore, compliance during the postmolar follow-up period is stressful for these patients, because of the need of postponing pregnancy.
      Other studies described the need of a shorter follow-up period after molar diagnosis.
      • Sebire N.J.
      • Foskett M.
      • Short D.
      • et al.
      Shortened duration of human chorionic gonadotrophin surveillance after complete or partial hydatidiform mole: evidence for revised protocol of a UK regional trophoblastic disease unit.
      Effective prognostic factors are needed to subgroup patients with different risk of GTN. Sebire et al
      • Sebire N.J.
      • Foskett M.
      • Short D.
      • et al.
      Shortened duration of human chorionic gonadotrophin surveillance after complete or partial hydatidiform mole: evidence for revised protocol of a UK regional trophoblastic disease unit.
      classified patients by the time that was needed for β-hCG normalization. We are proposing a new adjunctive method of selection of patients that is based on myometrial features at diagnosis together with β-hCG regression curve. These results should be followed by further work to establish a new method of monitoring patients after a molar evacuation.

      References

        • Berkowitz R.S.
        • Goldstein D.P.
        Presentation and management of molar pregnancy.
        in: Hancock B.W. Newlands E.S. Berkowitz R.S. Gestational trophoblastic disease. Chapman and Hall, London1997: 127-146
        • Seckl M.J.
        • Fisher R.A.
        • Salerno G.
        • et al.
        Choriocarcinoma and partial hydatidiform moles.
        Lancet. 2000; 356: 36-39
        • Ngan S.
        • Seckl M.J.
        Gestational trophoblastic neoplasia management: an update.
        Curr Opin Oncol. 2007; 19: 486-491
        • Morrow C.P.
        Postmolar trophoblastic disease: diagnosis, management, and prognosis.
        Clin Obstet Gynecol. 1984; 27: 211-220
        • Kim S.J.
        • Na Y.J.
        • Jung S.G.
        • Kim C.J.
        • Bae S.N.
        • Lee C.
        Management of high-risk hydatidiform mole and persistent gestational trophoblastic neoplasia: the Korean experience.
        J Reprod Med. 2007; 52: 819-830
        • Goldstein D.P.
        • Berkowitz R.S.
        • Bernstein M.R.
        Management of molar pregnancy.
        J Reprod Med. 1981; 26: 208-212
        • Lurain J.R.
        • Brewer J.I.
        • Torok E.E.
        • Halpern B.
        Natural history of hydatidiform mole after primary evacuation.
        Am J Obstet Gynecol. 1983; 145: 591-595
        • Niemann I.
        • Petersen L.K.
        • Hansen E.S.
        • Sunde L.
        Predictors of low risk of persistent trophoblastic disease in molar pregnancies.
        Obstet Gynecol. 2006; 107: 1006-1011
        • Berkowitz R.S.
        • Goldstein D.P.
        Chorionic tumors.
        N Engl J Med. 1996; 335: 1740-1748
        • Tow W.S.
        The influence of the primary treatment of hydatidiform mole on its subsequent course.
        J Obstet Gynaecol Br Commonw. 1966; 73: 544-552
        • Tsukamoto N.
        • Iwasaka T.
        • Kashimura Y.
        • Uchino H.
        • Kashimura M.
        • Matsuyama T.
        Gestational trophoblastic disease in women aged 50 or more.
        Gynecol Oncol. 1985; 20: 53-61
        • Xia Z.F.
        • Song H.Z.
        • Tang M.Y.
        Risk of malignancy and prognosis using a provisional scoring system in hydatidiform mole.
        Chin Med J (Engl). 1980; 93: 605-612
        • Ayhan A.
        • Tuncer Z.S.
        • Halilzade H.
        • Küçükali T.
        Predictors of persistent disease in women with complete hydatidiform mole.
        J Reprod Med. 1996; 41: 591-594
        • Stone M.
        • Bagshawe K.D.
        An analysis of the influences of maternal age, gestational age, contraceptive method, and the mode of primary treatment of patients with hydatidiform moles on the incidence of subsequent chemotherapy.
        BJOG. 1979; 86: 782-792
        • Feltmate C.M.
        • Growdon W.B.
        • Wolfberg A.J.
        • et al.
        Clinical characteristics of persistent gestational trophoblastic neoplasia after partial hydatidiform molar pregnancy.
        J Reprod Med. 2006; 51: 902-906
        • Berkowitz R.S.
        • Bernstein M.R.
        • Laborde O.
        • Goldstein D.P.
        Subsequent pregnancy experience in patients with gestational trophoblastic disease: New England Trophoblastic Disease Center, 1965-1992.
        J Reprod Med. 1994; 39: 228-232
        • Seckl M.J.
        • Dhillon T.
        • Dancey G.
        • et al.
        Increased gestational age at evacuation of a complete hydatidiform mole: does it correlate with increased risk of requiring chemotherapy?.
        J Reprod Med. 2004; 49: 527-530
        • Parazzini F.
        • Mangili G.
        • Belloni C.
        • La Vecchia C.
        • Liati P.
        • Marabini R.
        The problem of identification of prognostic factors for persistent trophoblastic disease.
        Gynecol Oncol. 1988; 30: 57-62
        • Parazzini F.
        • La Vecchia C.
        • Franceschi S.
        • et al.
        ABO blood-groups and the risk of gestational trophoblastic disease.
        Tumori. 1985; 71: 123-126
        • Curry S.L.
        • Hammond C.B.
        • Tyrey L.
        • Creasman W.T.
        • Parker R.T.
        Hydatidiform mole: diagnosis, management, and long-term follow-up of 347 patients.
        Obstet Gynecol. 1975; 45: 1-8
        • Berkowitz R.S.
        • Goldstein D.P.
        • DuBeshter B.
        • Bernstein M.R.
        Management of complete molar pregnancy.
        J Reprod Med. 1987; 32: 634-639
        • Berkowitz R.
        • Ozturk M.
        • Goldstein D.
        • Bernstein M.
        • Hill L.
        • Wands J.R.
        Human chorionic gonadotropin and free subunits' serum levels in patients with partial and complete hydatidiform moles.
        Obstet Gynecol. 1989; 74: 212-216
        • Benedet J.L.
        • Bender H.
        • Jones 3rd, H.
        • Ngan H.Y.
        • Pecorelli S.
        FIGO staging, classification, and clinical practice guidelines in the management of gynecologic cancers: FIGO Committee on Gynecologic Oncology.
        Int J Gynecol Obstet. 2000; 70: 209-262
        • Mangili G.
        • Garavaglia E.
        • Cavoretto P.
        • Gentile C.
        • Scarfone G.
        • Rabaiotti E.
        Clinical presentation of hydatidiform mole in northern Italy: has it changed in the last 20 years?.
        Am J Obstet Gynecol. 2008; 198: 302.e1-302.e4
        • Sebire N.J.
        • Foskett M.
        • Short D.
        • et al.
        Shortened duration of human chorionic gonadotrophin surveillance after complete or partial hydatidiform mole: evidence for revised protocol of a UK regional trophoblastic disease unit.
        BJOG. 2007; 114: 760-762