Fertility outcomes following surgery and multiagent chemotherapy in malignant ovarian germ cell tumor survivors: a survey study.

OBJECTIVE: To assess fertility outcomes in long-term survivors of malignant ovarian germ cell tumors treated with fertility-sparing surgery with or without additional chemotherapy. METHODS: Women diagnosed and treated for malignant ovarian germ cell tumors at Charing Cross Hospital or Mount Vernon Cancer Centre between 1977 and 2015 were included. Questionnaires assessing fertility issues were sent to patients treated with fertility-sparing surgery. Fertility outcomes were evaluated according to the treatment received. The effect of the mean total dose of cyclophosphamide and cisplatin was assessed. RESULTS: A total of 146 patients were sent the questionnaire; 77 (56.5%) patients were included in the analysis. A total of 49 (64%) patients received platinum-based chemotherapy after surgery, 39 (79.6%) of these with cisplatin, vincristine, methotrexate, bleomycin, actinomycin D, cyclophosphamide, and etoposide, while 10 (20.4%) with bleomycin, etoposide, and cisplatin. After any treatment, 39/46 patients (85%) became pregnant: the conception rate was not different between those receiving surgery only and those receiving also chemotherapy (85.7% vs 84.4%, p=1.0). Live birth rate was 80.4% (37/46), with no statistically significant difference between the treatment groups (p=0.42). Median age of women achieving conception was 29 years (IQR 26-33). The probability of live birth at 5 years was 48% and 40% for patients in the surgery only and chemotherapy group, respectively (p=0.55). Infertility and miscarriage rates did not differ significantly between the two treatment groups (p=0.30 and p=0.32). The mean doses of cisplatin and cyclophosphamide received by patients failing and achieving conception were not different (p=0.10, p=0.47). CONCLUSIONS: Our results suggest that fertility may not be hampered in patients with malignant ovarian germ cell tumor treated with fertility-sparing surgery or receiving additional chemotherapy.

Following chemotherapy for gestational trophoblastic neoplasia, do residual lung lesions increase the risk of relapse?

BACKGROUND: There remains uncertainty about the prognostic significance of residual lung lesion on imaging after completion of treatment of low- or high-risk gestational trophoblastic neoplasia (GTN). Here, we determine if such residual lung lesions are associated with an increased risk of relapse. METHODS: We retrospectively screened our electronic database to identify patients with low- or high-risk GTN and lung metastases between 2004 and 18. Recurrences among patients with or without residual lung lesions on imaging were compared. Chi square analysis and Kaplan-Meier survival curves were constructed. As the numbers of cases were low, we combined this data with our previously published and non-overlapping patient cohort (1995-2004). RESULTS: Of 1304 GTN patients treated at our centre between 2004 and 18, 99 had lung metastases without other distant sites. There were 40 patients (40.4%) with residual lung lesions. Whilst an increased rate of relapse was observed among patients with residual lung lesions (4/40; 10.0%) compared to without such lesions (3/59; 5.1%), this difference was not statistically significant (p = .35). By combining the data with our previous cohort, there was an increase in relapse rate of patients with residual lung lesions (5/63; 7.9%) compared to those without such lesions (4/112; 3.6%). However, this difference was also not statistically significant (p = .21). CONCLUSION: Residual lung lesions on imaging after completion of GTN treatment are common. However, this finding did not statistically increase relapse rate. Due to low number of recurrent events, a multi-centre, larger dataset would be needed to provide more definitive evidence.

Intensified therapies improve survival and identification of novel prognostic factors for placental-site and epithelioid trophoblastic tumours.

BACKGROUND: Placental-site trophoblastic (PSTT) and epithelioid trophoblastic tumours (ETT) are the rarest malignant forms of gestational trophoblastic disease (GTD). Our prior work demonstrated that an interval of ≥48 months from the antecedent pregnancy was associated with 100% death rate, independent of the stage. Here, we assess whether modified treatments for these patients have increased survival and identify new prognostic factors. METHODS: The United Kingdom GTD database was screened to identify all PSTT/ETT cases diagnosed between 1973 and 2014. Data and survival outcomes from our prior patient cohort (1976-2006) were compared to our new modern cohort (2007-2014), when intensified treatments were introduced. RESULTS: Of 54,743 GTD patients, 125 (0.23%) were diagnosed with PSTT and/or ETT. Probability of survival at 5 and 10 years following treatment was 80% (95% CI 72.8-87.6%) and 75% (95% CI 66.3-84.3%), respectively. Univariate analysis identified five prognostic factors for reduced overall survival (age, FIGO stage, time since antecedent pregnancy, hCG level, mitotic index) of which stage IV disease (HR 6.18, 95% CI 1.61-23.81, p = 0.008) and interval ≥48 months since antecedent pregnancy (HR 14.57, 95% CI 4.17-50.96, p < 0.001) were most significant on multivariable analysis. No significant differences in prognostic factors were seen between the old and new patient cohort. However, the new cohort received significantly more cisplatin-based and high-dose chemotherapy, and patients with an interval ≥48 months demonstrated an improved median overall survival (8.3 years, 95% CI 1.53-15.1, versus 2.6 years, 95% CI 0.73-4.44, p = 0.·005). CONCLUSION: PSTT/ETT with advanced FIGO stage or an interval ≥48 months from their last known pregnancy have poorer outcomes. Platinum-based and high-dose chemotherapy may help to improve survival in poor-prognosis patients.

Emergency Etoposide-Cisplatin (Em-EP) for patients with germ cell tumours (GCT) and trophoblastic neoplasia (TN).

BACKGROUND: Etoposide (E) at 100 mg/m2 combined with Cisplatin (P) at 20 mg/m2 represents an induction 2-day regimen embedded in our clinical practice for patients with advanced GCT or TN at high risk of early death. We evaluated 24/7 Em-EP administration to a combined GCT-TN cohort at our Emergency Cancer Treatment Centre (ECTC) to determine its efficacy within the acute setting. METHODS: Patients who received Em-EP during a five-year interval were identified from electronic databases at Imperial College Healthcare NHS Trust. Data collected included demographics, treatment details and clinical outcome. RESULTS: Em-EP was administered in the emergency setting to 104 patients, predominantly young adults (median age 35, range 17-71). Half the cases were GCT (n = 52): 22 male (6 seminomas, 13 non-seminomas); 30 female (2 dysgerminomas, 28 non-dysgerminomas). The other 50% were treated for TN (n = 52): 45 gestational (GTN) and 7 non-gestational. Most patients received Em-EP for a new cancer diagnosis (n = 100, 96%), within 24 h (n = 93, 89%) and out-of-hours (n = 74, 70%). Indications for Em-EP included symptomatic disease (n = 66, 63%), high-burden disease, (n = 51, 49%) and organ failure requiring Intensive Care Unit support (n = 9, 9%). Neutropenic sepsis was observed in 5%. Four-week overall survival after Em-EP administration was 98%. CONCLUSIONS: Despite the potentially fatal complications encountered in the acute setting, early mortality with Em-EP is low at our ECTC. Specialist units that treat unwell patients with advanced GCT or TN should consider making Em-EP available 24/7 for emergency administration. Its efficacy within a prospective cohort and in other platinum-sensitive malignancies requires evaluation.

When to stop human chorionic gonadotrophin (hCG) surveillance after treatment with chemotherapy for gestational trophoblastic neoplasia (GTN): A national analysis on over 4,000 patients.

OBJECTIVE: To determine the optimal duration of human chorionic gonadotrophin (hCG) surveillance following treatment for low and high risk gestational trophoblastic neoplasia (GTN) and establish whether the current surveillance protocol that recommends life-long hCG monitoring requires revision. METHODS: A population-based cohort study was undertaken using a national registry, comprising patients from both tertiary trophoblastic disease treatment units in the UK (London and Sheffield). All patients who received chemotherapy for low or high risk GTN in the UK between 1958 and 2014 in London and 1973 and 2015 in Sheffield (n = 4201) were included in the study. Patients with placental site trophoblastic tumours and epithelioid trophoblastic tumours were excluded due to their distinct clinical behavior, treatment and follow-up requirements. The risk of recurrence with time following completion of chemotherapy for low or high risk GTN was measured. RESULTS: The overall risk of relapse in this national cohort of 4201 patients was 4.7% (198/4201) with a median time to recurrence of 117.5 days (range 9 days to 6.54 years). The greatest risk of recurrence occurred in the first year after completing treatment for either low or high risk GTN measuring 72.7% (n = 112) or 86.4% (n = 38), respectively. The subsequent recurrence risk reduced over time with none observed beyond 7 years. CONCLUSIONS: The absence of any recurrences beyond seven years following completion of chemotherapy for GTN indicates that the UK policy of life-long hCG surveillance is unnecessary. Our revised conservative protocol recommends stopping after 10 years.

What is the optimal duration of human chorionic gonadotrophin surveillance following evacuation of a molar pregnancy? A retrospective analysis on over 20,000 consecutive patients.

OBJECTIVE: To quantify the risk of developing post-molar gestational trophoblastic neoplasia (pGTN) beyond the first normal human chorionic gonadotrophin (hCG) in women who have had a complete (CHM) or partial molar pregnancy (PHM) and to re-evaluate the current UK Hydatidiform mole hCG surveillance guidelines. METHODS: The Charing Cross Hospital Trophoblast Disease Centre database was screened to identify all registered cases of hydatidiform mole (HM) between 1980 and 2009. RESULTS: We identified 20,144 cases of HM, comprising 8400 CHM, 9586 PHM, and 2158 cases of unclassified hydatidiform mole (UHM). Twenty-nine cases (20 CHM, 3 PHM and 6 UHM) developed pGTN after the first normal hCG. For CHM the risk of pGTN at the point of hCG normalisation was 1 in 406, and fell rapidly in the first six months of monitoring. For PHM the risk of pGTN at the point of hCG normalisation was 1 in 3195. Women with CHM where hCG normalisation occurred beyond 56days after uterine evacuation of molar tissue were found to have a 3.8-fold higher risk of pGTN. CONCLUSIONS: Our results show that pGTN can occur after hCG normalisation following PHM but the risk is extremely low. Women with CHM have a comparatively higher risk of pGTN after hCG normalisation. Those with CHM where hCG normalises within 56days have a lower risk of pGTN. We have revised the current UK hCG surveillance protocol for PHM to a single additional confirmatory normal urine hCG measurement one month after first normalisation. The protocol for CHM remains unchanged.

Brain metastases in gestational trophoblast neoplasia: an update on incidence, management and outcome.

OBJECTIVE: To update the demographic data, treatment details and outcomes for GTN patients with brain metastases managed with the modern treatment protocols at the UK centre for gestational trophoblast neoplasia at Charing Cross Hospital in London. METHODS: The hospital database and pharmacy records were reviewed to identify GTN patients treated with brain metastases. Data was assembled on the specific GTN diagnosis, staging, prognostic scores, chemotherapy regimens, additional interventions and outcomes. RESULTS: During the 22 year study period, 27 GTN patients with brain metastases were treated. One case clearly resulted from a prior molar pregnancy, 3 were of uncertain aetiology and 23 cases had no prior molar pregnancy. The standard chemotherapy regimens were EMA-CO or EMA-EP given with an enhanced CNS methotrexate dose combined with intrathecal methotrexate. Five patients required emergency neurosurgery and routine radiotherapy was not employed. Twenty three (85%) patients are long term survivors and four patients died. Of the patients who died, all four had chemotherapy refractive disease and two had extended intervals of 18 and 30 years from their antecedent pregnancy. CONCLUSION: The incidence of brain metastases in postmolar pregnancy GTN is extremely low. Patients with non-molar choriocarcinoma have an approximate 20% risk of CNS disease and should have routine CNS imaging. Treatment with CNS doses of EMA-CO or EMA-EP appears curative for most patients.

Atypical placental site nodule (APSN) and association with malignant gestational trophoblastic disease; a clinicopathologic study of 21 cases.

The WHO Classification of Gestational Trophoblastic Tumors classifies placental site nodule (PSN) as a benign tumor-like trophoblastic neoplasm. Cases of PSN with atypical features were described [atypical placental site nodule (APSN)] and we started registering APSN in our unit in 2005. The aim of this study is to present our initial experience with these lesions. The Trophoblastic Disease Unit database was searched to identify all patients who were either referred with, or on review were diagnosed with, APSN from September 2005 to May 2013. Case notes and the pathology findings for these patients were retrieved and reviewed. A total of 21 cases of APSN were included, 3 of which were associated with gestational trophoblastic neoplasm on follow-up or review. Malignant gestational trophoblastic disease was associated with 3/21 (14%) cases of APSN, either concurrently or developing/manifesting within 16 mo of APSN diagnosis. None of these patients had raised serum hCG levels either at presentation or follow-up. Presence of APSN should indicate a thorough clinical and radiologic investigation and follow-up if diagnosed on curettage specimens. With increased recognition of this entity and corresponding larger series with longer follow-up, more accurate patient counseling will be possible.

Histopathological and immunohistochemical features of early hydatidiform mole in relation to subsequent development of persistent gestational trophoblastic disease.

OBJECTIVE: To examine histomorphological and immunohistochemical findings in hydatidiform moles to determine whether any features can reliably predict clinical behavior. STUDY DESIGN: Blinded semiquantitative review of histological and immunohistochemical findings in cases of partial hydatidiform mole (PHM) (N = 50) and complete hydatidiform mole (CHM) which either spontaneously resolved (N = 50) or required chemotherapy (N = 50). Immunostains assessed included MLH1, MSH2, nm23, TERT, p53, EGFR, and CerbB2 based on previous data. RESULTS: There were marked morphological differences in various criteria between CHMs and PHMs, including the proportion of villi with abnormal trophoblast hyperplasia (29% vs. 6%, respectively). However, there were no significant differences in any morphological parameters between CHMs that spontaneously resolved and those that subsequently required chemotherapy. Similarly, there were no clinically useful differences regarding any immunostaining scores between CHM groups. CONCLUSION: Neither morphological nor immunohistochemical features can reliably predict subsequent requirement of chemotherapy in CHMs.

Effect of early pregnancy following chemotherapy on disease relapse and fetal outcome in women treated for gestational trophoblastic neoplasia.

OBJECTIVE: To examine the effects of early pregnancy (< 12 months following chemotherapy) on a recent cohort of women treated with modern therapies for gestational trophoblastic neoplasia (GTN). STUDY DESIGN: The Charing Cross GTN database was screened between 1998-2012 to identify 1,204 patients treated with either single-agent (61.9%) or multiagent (38.1%) chemotherapy. RESULTS: A total of 23% of single-agent and 15.4% of the multiagent treatment groups conceived within 12 months of chemotherapy, resulting in 255 early pregnancies, with 73.3% resulting in live births. There was no significant increased risk of miscarriage, ectopic pregnancy, second molar pregnancy or stillbirth as compared to the general U.K. population. Intriguingly, the incidence of relapse was only 1.7% in the early pregnancy group as compared to 5.2% in the 963 patients who did not conceive early. CONCLUSION: Women who become pregnant within 12 months postchemotherapy for GTN can be reassured of a likely favorable outcome, although the safest option is still to delay pregnancy for a year.

Management and survival of patients with FIGO high-risk gestational trophoblastic neoplasia: the U.K. experience, 1995-2010.

OBJECTIVE: To present survival rates of high-risk gestational trophoblastic neoplasia (GTN) (FIGO score > 7) patients treated between 1995 and 2010 in the U.K. Death due to GTN is largely confined to patients with high-risk disease. In the U.K. a national system ensures that all patients are treated at only 2 specialist centers: Charing Cross Hospital (CXH) in London and Weston Park Hospital (WPH) in Sheffield. STUDY DESIGN: A total of 196 high-risk patients were identified using the CXH and WPH GTN databases, based on the risk score at the time of presentation. RESULTS: In all, 140 CXH and 56 WPH high-risk patients were treated with EMA/CO (etoposide, methotrexate, actinomycin D alternating with cyclophosphamide and vincristine) and MEA (methotrexate, etoposide, actinomycin D), respectively. The FIGO score at presentation ranged from 6-23. Eight patients (7from WPH and 1 from CXH) who were treated prior to 2002 as high-risk based on their pre-2002 scoring scored a 6 using FIGO 2002. Two (1%) patients died within 4 weeks of starting treatment (early death), 12 (6%) relapsed, and 9 patients subsequently died due to drug resistance. The overall survival was 94%, with a median follow-up of 4.69 years. CONCLUSION: In the context of a national trophoblastic disease service, patients with high-risk GTN have an excellent prognosis with EMA/CO or MEA.

Survival of women with gestational trophoblastic neoplasia and liver metastases: is it improving?

OBJECTIVE: To determine whether survival outcomes of women with liver metastases from gestational trophoblastic neoplasia (GTN) have improved from the previous finding of 27% at 5 years. STUDY DESIGN: The Charing Cross GTN database was searched for patients with liver metastases treated between 1975 and 2007. Prognostic variables were recorded and analyzed for effect on survival. RESULTS: Thirty-eight (1.8%) of 2,100 GTN patients had liver metastases. One patient with placental site trophoblastic tumor was excluded. In the remaining 37 cases the overall survival was 48% at 5 years. Seven patients with very advanced disease died <4 weeks after admission, and 12 late deaths occurred, 5 due to non-GTN causes (1 stroke and 4 second cancers). After exclusion of the early deaths and censoring for the non-GTN related deaths, the cause-specific survival was 68%. No prognostic variable was significant on univariate analysis. However, patients presenting >2.8 years and <2.8 years from the antecedent pregnancy had a 32% and 75% (p = 0.08) chance of long-term survival, respectively. CONCLUSION: The prognosis of patients with liver metastases from GTN has improved. Outcome may be best in those patients presenting within 2.8 years of the causative pregnancy and without very large volumes of disease.

Uterine artery pulsatility index and serum BMP-9 predict resistance to methotrexate therapy in gestational trophoblastic neoplasia: A cohort study.

BACKGROUND: Methotrexate is the most common first-line chemotherapy for low-risk gestational trophoblastic neoplasia (GTN). Uterine artery pulsatility index (UAPI) is an ultrasound marker for tumor vascularity that has been associated with an increased risk of methotrexate resistance. The combination of circulating angiogenic factor levels with UAPI data may improve the capacity of this model to predict chemoresistance. METHODS: This was a single-center cohort study of women newly diagnosed between January 2008 and June 2012 with low-risk GTN during postmolar surveillance and treated with single-agent methotrexate at Charing Cross Hospital, a UK national center for treatment of gestational trophoblastic disease. Two hundred seventeen women underwent an ultrasound for UAPI measurement prior to initiation of chemotherapy. To examine serologic markers of methotrexate resistance among this cohort, we performed a case-control study using archived serum from 76 patients who could be matched based on prognostic risk score. Serum samples were examined by immunoassay to measure 8 different angiogenic factors (VEGF-A, FGF-basic, PLGF-1, PDGF-BB, EGF, ANGPT2, BMP-9, and ENG). Receiver-operator characteristic area under the curve (AUC) values were calculated for the ability of each analyte to correctly classify patients as methotrexate sensitive (MTX-S) or resistant (MTX-R). RESULTS: Total human chorionic gonadotropin levels were similar between the MTX-S and MTX-R groups. UAPI values were significantly higher in MTX-S (median 1.30 [interquartile range {IQR} = 0.80-1.90]) compared to MTX-R patients (median 0.875 [IQR = 0.60-1.30]; P < 0.0001) with AUC 0.68 (95% confidence interval 0.61-0.76; P < 0.0001). In univariate analysis, only BMP-9 concentrations were significantly different between groups, lower among MTX-S (median of 225 ng/L, IQR = 170-287) compared to MTX-R patients (median 280 ng/L [IQR = 200-339]; P= 0.03). Combining UAPI with BMP-9 concentration improved prediction for chemoresistance with AUC 0.77 (95% confidence interval 0.66-0.88; P < 0.0001). CONCLUSION: Circulating levels of BMP-9 are elevated in newly diagnosed women with low-risk GTN destined to fail primary methotrexate therapy. A combined test using serum BMP-9 plus UAPI might improve prediction of MTX-R in low-risk GTN.

The demographics of molar pregnancies in England and Wales from 2000-2009.

OBJECTIVE: To analyze the overall incidence of molar pregnancies and that of complete and partial molar pregnancies across the reproductive age range for England and Wales for the period 2000-2009. STUDY DESIGN: The cases of all patients with molar pregnancies registered with the UK Trophoblast disease service from England and Wales were identified. The overall number of molar pregnancies registered from 1998-2007 was compared to the number of maternities (live births and still births) and total viable conceptions for each year. For the series 2000-2009 the number of complete and partial molar pregnancies were compared to the number of maternities and terminations occurring for women across the age range < 14 to 50+ years, allowing an accurate estimate of the risk of molar pregnancy for women conceiving at any age. RESULTS: The results indicate that for the period 1998-2007 the overall incidence of molar pregnancies was 1 case per 591 viable conceptions. The incidence increased from 1:611 in 1997 to 1:528 in 2008. The age-specific data for the period 2000-2009 confirms a risk level of < 0.2%for women aged 18 {N dash} 39 years, with a modest excess risk for young teenagers but a much more significant increase for women > 40, where the risk is 1% at 45 and 17% at > or = 50. CONCLUSION: This study provides detailed data regarding the risk of partial and complete molar pregnancies with increasing maternal age. It confirms that the risk of partial molar pregnancy varies relatively little with age, with complete molar pregnancies contributing the main component of the overall increase with age.

Can we replace adjuvant chemotherapy with surveillance for stage IA-C immature ovarian teratomas of any grade? an international multicenter analysis.

BACKGROUND: The role of surveillance after surgery for stage IA-C grade 2 (G2) or grade 3 (G3) immature teratomas (ITs) is controversial with many guidelines advocating adjuvant chemotherapy. Here, we investigate the safety of surveillance in stage IA-C G1-3 ITs. METHODS: Clinicopathological data were analysed on postpubertal patients with stage I pure ITs in Multicenter Italian Trials in Ovarian Cancer centres and at Charing Cross Hospital, UK, between January 1985 and January 2018. RESULTS: Of 108 stage I patients, 66 (61.1%), 3 (2.8%) and 39 (36.1%) were International Federation of Gynecology and Obstetrics IA, IB, IC, respectively, with 31 (28.7%), 41 (38%) and 36 (33.3%) having grade 1 (G1), 2 and 3 disease, respectively. After surgery, 27 patients (25%) had adjuvant chemotherapy and 81 (75%) surveillance. There was no significant increase in the risk of malignant (G2-3 IT) relapse (9/81 vs 2/27; p = 0.72) or in disease-free survival (DFS) or overall survival in the surveillance vs chemotherapy groups. The median time to relapse was 17.8 months (range: 3-47) with no significant difference between surveillance or chemotherapy groups. The median follow-up was 64.3 months (Interquartile range (IQR) 22.2-101.7). Chemotherapy induced cures in all except for one patient who did not follow the surveillance protocol due to pregnancy and died of disease. Univariate and multivariate analyses revealed that only tumour grade (hazard ratio [HR] = 3.11; p = 0.02) and complete surgical staging (HR = 0.2; p = 0.01) were independent prognostic factors for decreased DFS. CONCLUSION: The present study suggests that in the adult setting careful surveillance appears to be an acceptable alternative to adjuvant chemotherapy for stage IA-C ITs of any grade, properly staged and with negative postoperative tumour markers.

Effects of single-agent and combination chemotherapy for gestational trophoblastic tumors on risks of second malignancy and early menopause.

PURPOSE: To assess the risks of second malignancy and early menopause in a large cohort of patients with gestational trophoblastic tumor who were treated with chemotherapy. PATIENTS AND METHODS: A survey of patients treated at Charing Cross Hospital between 1958 and 2000 was performed in 2006 to assemble incidence data for subsequent malignancies and the age at menopause. Treatment records were reviewed for the regimens and durations, and the incidence of subsequent malignancies was compared with that in the national age-matched population. RESULTS: Data were obtained for 1,903 patients, with a mean follow-up of 16.9 years. Eighty-six patients developed a subsequent malignancy compared with an expected number of 79 (standardized incidence ratio [SIR], 1.1; 95% CI, 0.9 to 1.3). The overall risk was low for patients treated with single-agent methotrexate and folinic acid (MTX-FA; SIR, 0.7; 95% CI, 0.5 to 1.1) and also for patients treated with etoposide, methotrexate, and dactinomycin followed by cyclophosphamide and vincristine on alternating weeks (EMA-CO) with an SIR of 0.9 (95% CI, 0.4 to 2.2), but there were significantly increased risks of oral cancer, melanoma, meningioma, and leukemia. The cumulative risk of early menopause was low after MTX-FA but was substantial after EMA-CO, reaching 13% by age 40 years and 36% by age 45 years. CONCLUSION: Subsequent cancer risks for patients cured of gestational trophoblastic tumors with modern chemotherapy appear similar to those of the normal population with no overall increased risk of malignancy after MTX-FA or EMA-CO. However, there was evidence of an increased risk of leukemia after EMA-CO and some evidence of other site-specific increased risks based on small patient numbers. All major treatments except MTX-FA increased the risk of early menopause.

Clinical utility of selective molecular genotyping for diagnosis of partial hydatidiform mole; a retrospective study from a regional trophoblastic disease unit.

AIMS: Hydatidiform moles (HMs) are genetically abnormal conceptions, associated with increased risk of gestational trophoblastic neoplasia. Diagnosis is usually based on histopathological criteria but in a minority definitive histological diagnosis is not possible; in such cases molecular genotyping may be diagnostic. This study describes the clinical usefulness of such an approach. METHODS: Cases in which central histology review demonstrated abnormal villous morphological features insufficient for definite diagnosis of partial HM (PHM) ('favour PHM' or 'PHM not excluded') underwent molecular genotyping of villous and maternal tissue, using short tandem repeats, to determine ploidy and parental origin of the placental tissue. RESULTS: Of 251 cases with non-diagnostic morphological villous abnormalities, molecular investigation was not possible in 14 (6%; limited material or technical issues). Overall, 124 (49%) were triploid including 71/86 (85%) of those morphologically favouring PHM, and 53/165 (32%) of those favouring non-molar miscarriage. Of 85 cases of triploidy in whom sufficient material was available, 84 had an additional paternal contribution. Single cases of digynic triploidy, tetraploid PHM and two mosaic conceptions were also identified. Twenty-three non-molar diploid cases (21%) exhibited trisomy. CONCLUSIONS: Molecular genotyping allows definitive diagnosis of PHM for cases in which specialist histopathology review remains equivocal. While this approach provides definite diagnosis it is considerably more expensive than a pragmatic management approach of human chorionic gonadotrophin surveillance in all such cases.