The influence of lung metastases on the clinical course of gestational trophoblastic neoplasia: a historical cohort study

Transcription

1 DOI: / Gynaecological oncology The influence of lung metastases on the clinical course of gestational trophoblastic neoplasia: a historical cohort study M Vree, a N van Trommel, b G Kenter, b F Sweep, c M ten Kate-Booij, d L Massuger, e C Lok b a Department of Gynaecologic Oncology, The Netherlands Cancer Institute, Amsterdam, The Netherlands b Department of Gynaecologic Oncology, Centre of Gynaecologic Oncology Amsterdam, Amsterdam, The Netherlands c Department of Laboratory Medicine, Radboud University Medical Centre, Nijmegen, The Netherlands d Department of Gynaecologic Oncology, Erasmus Medical Centre, Rotterdam, The Netherlands e Department of Obstetrics and Gynaecology, Radboud University Medical Centre, Nijmegen, The Netherlands Correspondence: C Lok, Department of Gynaecologic Oncology, Centre for Gynaecologic Oncology Amsterdam, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands. c.lok@nki.nl Accepted 23 July Published Online 12 October Objective To evaluate whether gestational trophoblastic neoplasia (GTN) patients with lung metastases have more adverse outcomes such as resistance to chemotherapy, recurrence or death of disease compared with patients without lung metastases. Design Historical observational cohort study. Setting The Netherlands. Population We identified 434 GTN patients (72 patients with lung metastases, 362 patients without metastases) between 1990 and 2012 registered in the Dutch national databases. Methods Baseline characteristics, recurrence rates, Methotrexate (MTX) remission rates and deaths from disease were compared between patients with lung metastases (group I) and without lung metastases (group II) using the Fisher exact test or Mann Whitney U-test where applicable. Main outcome measures Methotrexate resistance, recurrences and survival. Results Methotrexate resistance did not differ between group I and group II (62.9 versus 72.7% P = 0.19). However, the observed recurrence rate was significantly increased in patients with lung metastases compared with patients without metastases (16.7 versus 2.2% P < ), also after correction for antecedent pregnancy and interval (from the end of the antecedent pregnancy until the start of treatment). Disease-specific survival was 91.7% in the group with lung metastases and 100% in the patients without metastases (P < ). Conclusions Although lung metastases are considered to be associated with a low risk of adverse outcomes, their presence appears to increase the risk for recurrence and death of disease. Further research is needed to evaluate whether the presence of lung metastases is an independent risk factor that needs adjustment in the FIGO scoring system and clinical classification system. Keywords Gestational trophoblastic disease, gestational trophoblastic neoplasia, lung metastases, methotrexate, recurrence. Tweetable abstract In gestational trophoblastic neoplasia (GTN) recurrence is more often observed in the case of lung metastases. Please cite this paper as: Vree M, van Trommel N, Kenter G, Sweep F, ten Kate-Booij M, Massuger L, Lok C. The influence of lung metastases on the clinical course of gestational trophoblastic neoplasia: a historical cohort study. BJOG 2016;123: Introduction Gestational trophoblastic disease (GTD) is a heterogeneous group of diseases arising from an abnormal proliferation of trophoblastic tissue and includes hydatidiform mole, choriocarcinoma, placental site trophoblastic tumour (PSTT) and epithelioid trophoblastic tumour (ETT). 1 Both PSTT and ETT are very rare tumours with <1% of all forms being a PSTT or an ETT. 2,3 When GTD persists or when evidence of metastatic disease is present, or in the case of recurrent disease the term gestational trophoblastic neoplasia (GTN) is often used. 1 When GTN is diagnosed, accurate identification of metastases is necessary to identify patients with a high risk of developing resistance to singleagent chemotherapy. High-risk patients will be treated with multi-agent chemotherapy directly and can often be cured with this treatment, even in the presence of widespread metastases. 4,5 In contrast, low-risk patients are treated with 1839

2 Vree et al. single-agent chemotherapy. Thus, careful risk assessment before the start of chemotherapy is essential. In the Netherlands, a clinical classification system is used (Table S1). Another widely used prognostic scoring system is the FIGO 2000 scoring system. 6 In the last decades, the prognosis of GTN has improved as a result of developments in diagnostics and treatment. GTN has changed from a disease with a fatal prognosis to a disease with overall cure rates up to almost 100% in lowrisk disease. 7,8 However, in the case of high-risk disease the survival rate is less favourable, with a cure rate of 86 94% Known adverse prognostic factors in GTN are: liver metastases, brain metastases, long interval from preceding pregnancy to therapy and an antecedent term delivery. 11,12 Both in the Dutch clinical classification system and in the FIGO scoring system, lung metastases are not considered to be an adverse prognostic factor. However, the exact influence of lung metastases on the course of the disease is unknown. Although previous studies by Darby et al. (2009) and Garner et al. (2004) reported no significant influence in clinical outcome in patients with micro-lung metastases detected on computed tomography (CT) scan in otherwise non-metastatic GTN, a greater need to switch to secondline treatment was found in these studies. 13,14 However, this has not been investigated in macro-lung metastases. As patients with lung metastases have a greater tumour bulk, higher hcg concentrations, more courses of chemotherapy and possibly a more protracted course of disease can be expected. The aim of the present study was to evaluate whether GTN patients with lung metastases have a more complicated course of disease compared with patients without metastases. Methods In this historical cohort study, patient data were collected from two registries: the Dutch Working Party on Trophoblastic Disease and the Dutch Central Registry for Hydatidiform Moles (DCRHM). The working party holds quarterly meetings to discuss new patients with GTN, provides treatment advice to medical specialists and collects data for research purposes. The DCRHM is a voluntary national registry for patients with GTD and provides a national hcg assay service to gynaecologists. For this study, no medical ethical approval was required because patients previously consented to registration and data was completely anonymised. In Figure 1 the patient selection process is shown; all GTN patients registered in the DCRHM and the Dutch Working party between 1 January 1990 and 1 January 2013 were selected. Duplicates in the two databases were eliminated by checking date of birth, evacuation date and surnames. Patients with PSTT, ETT or unknown metastatic status were excluded from this study. Patients were divided into two groups: patients with only lung metastases (group I) and patients without any metastases (group II). Patients with lung metastases were included in group I if one or more lung metastases were present on X-ray and/or computed tomography (CT)-thorax upon first GTN staging, without other metastases at any time during the course of disease. For inclusion in group I, no discrimination was made between patients with micro- or macro-lung Figure 1. Flowchart illustrating the patient selection process. 1840

3 Influence of lung metastases on clinical course of GTN metastases. Patients that presented with recurrent disease with lung metastases were excluded to make sure only patients with lung metastases at initial presentation were included. Exclusion of patients with lung metastases at recurrence was necessary to prevent a selection bias for patients with recurrent disease in group I. The exclusion prevents patients with progressive disease due to failure of therapy, from being included in group I; this is a group of patients that is also not scored by FIGO. Patient files, discharge papers and official letters, meeting reports and digital registration documents were evaluated for all GTN patients with only lung metastases. In group II, all patients without metastases during the complete course of the disease were included. For the definition of GTN the Dutch criteria were used: a plateau in serum hcg concentrations for 3 weeks or an increase for two consecutive weeks and at least one hcg serum measurement above the 95th percentile of the hcg regression curve After diagnosing GTN, a chest X-ray is performed. When further staging is necessary, the National Guideline recommends CT abdomen, pelvis and chest and, in the case of lung metastases, an MRI of the brain. In the case of low-risk disease, singleagent chemotherapy with methotrexate (MTX) is the treatment of choice, whereas high-risk disease patients should receive multi-agent chemotherapy with etoposide, MTX and actinomycin D alternating with cyclophosphamide and vincristine (EMA/CO). The pre-evacuation serum hcg concentrations, regression of hcg on treatment, the need for single- and multi-agent chemotherapy, and the number of courses necessary to achieve normalisation of hcg concentrations were recorded. A large number of hcg concentrations measured by referring hospitals was available, but because of the use of different assay types, these values could not be used for comparison of hcg concentrations between groups. For this reason, only standardised hcg measurements from the central laboratory were reported in this study to enable comparison. Furthermore, recurrence (defined as an increase of hcg after normalisation, after any treatment measured in at least one sample) and death of disease were registered. Statistical analysis was performed using IBM SPSS software (version 20). Baseline characteristics were compared between group I and group II using the Fisher exact test where applicable. Differences in pre-evacuation hcg concentrations, number of weeks to achieve remission, and number of required chemotherapy regimens in patients cured with MTX were calculated and compared between the patients in group I and group II, using the Mann Whitney U-test. Choice of first treatment and effectiveness of treatment, recurrence of disease and disease-specific survival were compared between group I and group II using Fisher s exact test. Multivariate analysis was hampered by the fact there is a very low incidence of death and recurrence in these groups; therefore a propensity-matched analysis was performed as described by Austin. 18 In this analysis, patients in group I were matched one-to-one with patients in group II based on antecedent pregnancy and interval from evacuation until treatment. Statistical tests were repeated on these matched patients to correct for confounding by these factors. Results A total of 434 patients were included in this study, 72 patients with only lung metastases (group I) and 362 patients with no metastases (group II). Three patients had lung metastases at recurrence but not at initial risk classification and therefore these patients were excluded from group I. Patient baseline characteristics of group I and group II are shown in Table 1. No significant differences were found in age. Other characteristics such as antecedent term pregnancy and a histological diagnosis of choriocarcinoma were more frequently seen in the patients with lung metastases. In contrast, antecedent molar pregnancy and partial mole were less frequently observed in group I than group II. In 21 patients only micro-metastases were detected. In this subgroup, outcome was not significantly different from patients with macro-metastases, which could be due to the small size of the subgroup. In the patients with micro-metastases, five (23.8%) had a recurrence, three (14.3%) died of disease, and nine of 15 patients that started on MTX went into remission. Therefore, both groups were pooled in group I. The pre-evacuation serum hcg concentrations of patients in group I were significantly higher than those of patients in group II with a median ng/ml [interquartile range (IQR) 25 75%: ng/ml] versus median ng/ml [IQR: ng/ml, P = 0.02). The hcg concentrations before the start of chemotherapy did not differ significantly between both groups. Group I had a median hcg concentration of 2000 ng/ml (IQR: ng/ml) and group II a median hcg concentration of 870 ng/ml (IQR: ng/ml, P = 0.07). The time from evacuation to remission in patients only treated with chemotherapy, between patients in group I (median = 18.9 weeks, interval weeks) and group II (median = 16.1 weeks, interval ) did not differ significantly (P = 0.09). In patients only treated with MTX as first line therapy (patients needing multichemotherapy or Actinomycin D were not included in this comparison) there was no significant difference in the number of courses needed to achieve remission (median: 1841

4 Vree et al. Table 1. Baseline characteristics of included GTN patients in sorted by patient group Lung metastases (group I) No metastases (group II) n = 72 % n = 362 % Age (years) N.S. Antecedent pregnancy Mole Term Miscarriage N.S. Unknown N.S. Other N.S. Histopathological diagnosis Complete mole N.S. Partial mole Mole* N.S. Choriocarcinoma < Unknown N.S. Pre-evacuation hcg n = 16** n = 52** Median, ng/ml *** Prechemotherapy hcg Median, ng/ml *** n = 24** n = 32** *** MTX n = 51** n = 281** six courses in group I versus five courses in group II; P = 0.09). Treatment with MTX was initiated in 76.4% of the patients in group I and 90.1% in group II (P < ; Table 2). No significant difference in MTX chemotherapy remission was found between these two groups (62.9 versus 72.7%; P = 0.19; Table 2). The response to MTX is shown in Figure 2. Of the patients in group I, 13 patients (18.3%) started with multi-chemotherapy compared with 10 patients P *** Median MTX courses 6 2*** 5 4*** 0.09 Clinical classification Low-risk High-risk *Mole, not further specified. **Calculated on a subset of patients based upon availability. ***Interquartile ranges: 75 25% are provided. Table 2. Treatment regimen Lung metastases (group I) No metastases (group II) n = 72 % n = 362 % Primairy treatment MTX < EMA/CO < Hoog-Brabant* N.S. Hysterectomy N.S. Other N.S. None N.S. Unknown N.S. Outcome registered Remission 34/54** / 72.7 N.S. after MTX 300** Recurrence < Disease-specific death < *Hoog-Brabant: EMA/CP multi-agent chemotherapy regimen containing etoposide, methotrexate, Actinomycin D, cyclophosphamide and cisplatin. 25 **After start of treatment, 27 patients were lost to follow up (group I: 1 patient, group II: 26 patients) (2.8%) in group II (P < 0.001) because they were considered to be high risk. In the remaining patients the first treatment was a hysterectomy followed in 40.0% of the cases by chemotherapy (n = 6/15) and in nine patients in group II, remission was achieved with a second curettage. In Table 2 an overview is shown of the treatment regimens. After MTX treatment, eight patients in group I and 14 in group II were treated with Actinomycin D as second line treatment and eight (100%) and 10 (71.4%) patients, respectively, were cured. A significantly higher recurrence rate was observed in the patients in group I (n = 12) than group II (n = 8) (16.7 versus 2.2%; P < ). After first line mono- or multi-agent chemotherapy treatment, 63.6% of the patients in group I achieved remission compared with 72.5% in group II (P = 0.18). All except one of the 20 patients with recurrent disease were initially classified as low-risk patients. Two of the patients with a recurrence, died of disease. Survival was significantly higher in group II than in group I: 362 patients versus 66 patients, respectively (100 versus 91.7%; P < ). Clinical characteristics of deceased patients are shown in Table S2. In all included patients at least a chest X-ray was available. In 79 patients, both chest X-ray and CT-thorax were performed and documented. In 20 (27.7%) patients, P 1842

5 Influence of lung metastases on clinical course of GTN Figure 2. Remission rates after MTX treatment. lung metastases were diagnosed on CT-scan but were undetectable with chest X-ray. In the propensity-matched analysis, the 72 patients of group I were matched to 72 patients of group II based on antecedent pregnancy and interval from evacuation to treatment. In this analysis, again no significant difference in remission to MTX was found between patients and controls (62.7 versus 61.5%; P = 0.53) or in the number of administered MTX courses (median six versus five; P = 0.06). In contrast, recurrence of disease occurred more often in patients in group I (15.3%) than group II (4.2%) (n = 11 versus n = 3, respectively; P = 0.02) and death of disease remained significantly elevated in this propensity analysis: six patients (8.3%) with lung metastases died as a consequence of disease or complications of the treatment, compared with none in group II (P = 0.01). Discussion Main findings The aim of our study was to investigate the effect of lung metastases on the course of GTN. We hypothesised that the presence of lung metastases could have a negative influence on the course of the disease, unlike what is presumed in current risk classifications. In this study, adverse outcomes such as recurrences and mortality were more often observed in patients with lung metastases than in patients without metastases. Strengths and limitations For the detection of lung metastases, a chest X-ray is required, although CT-chest is known to be more sensitive in detecting micro-lung metastases. In earlier studies the presence of micro-lung metastases has shown no effect on clinical outcome. 13,14,19,22 However, a higher rate of chemotherapy resistance was reported. 13,20 In this study, not all patients underwent the same imaging procedures, as some patients underwent no CT-chest, so we cannot exclude that group II may contain patients with micro- lung metastases, left undetectable on regular chest X-ray. Because death of disease and recurrence were rare in group II, these small metastases are unlikely to influence prognosis. In group I, patients with only micro-lung metastases could have been included if no chest X-ray was available and the CT-scan reported the presence of lung metastases. Micro-metastases were defined as lungmetastases not detected by chest X-ray but present on CT-chest. Patients with lung metastases had a recurrence rate of 16.7%, which is higher than the previously reported 8.3%. 21 Future studies in larger patient groups will need to confirm this difference because no previous studies have reported a separate recurrence rate for patients with lung metastases with either low- or high-risk disease. This study has several limitations. Due to its retrospective nature, some data were missing. The registration in DCRHM is voluntary and it is known from a former study 1843

6 Vree et al. that not all GTD patients are registered in DCRHM. 2 In an attempt to identify all GTN cases in the Netherlands, data from two extensive national databases were used, and the combination of both minimises the risk of missing data. Because of the use of the clinical risk classification in the Netherlands, exact FIGO scores could not be calculated in all patients. However, all patients with FIGO scores >7 were also designated as high-risk in the clinical classification. Therefore, the use of the clinical system does not change the main results of our study. Unfortunately, correction for the effect of size and number of lung metastases could not be performed due to missing data and small subgroups. In this study, low- and high-risk patients were both included because the aim of the study was to assess the effect of lung metastases on the course of disease and not to validate the risk classification system. Not all patients were classified and subsequently treated according to the Dutch classification system; nine patients were mistakenly started on MTX after a term pregnancy. It is to be expected that misclassification occurred equally in both groups and so no bias was introduced. As CT-scans of the abdomen are not standard procedure, it cannot be excluded that a few patients included in group II have intra-abdominal metastases. However, the excellent survival and response to therapy in this group suggest otherwise. Although the registration of trophoblastic disease is nationwide and death of disease is likely to be reported, under-reporting cannot be excluded completely; however, this could have occurred in both groups. Interpretation Term pregnancy is described as a risk factor for a complicated course of disease and is one of the factors in the FIGO scoring system used to determine whether treatment with multi-chemotherapy is indicated. 4,12 Patients with lung metastases more often had an antecedent term pregnancy. This is not an explanation for the unfavourable outcome in group I, as the same results were found after correcting for antecedent pregnancy in the propensity-based analysis. The association of lung metastases, choriocarcinoma and antecedent term pregnancy suggests that a choriocarcinoma has a higher tendency to disseminate to the lungs compared with molar pregnancy. Hypothetically, another factor that could influence the risk for a more unfavourable outcome is a longer interval from the end of the pregnancy to the start of therapy. Lung metastases may have more time to become visible or symptomatic, suggesting that patients with lung metastases have a more unfavourable course of disease caused by a longer interval until treatment. However, after correction for the interval evacuation to treatment, the recurrence and disease-specific death remained significant. Another risk factor in the FIGO classification is the hcg serum concentration; higher serum hcg concentrations were measured in patients with lung metastases than in group II patients. This may reflect tumour burden, as hcg is a measurement of the bulk of the disease. Although in the Netherlands a central laboratory is available for standardised hcg measurements, many community hospitals measure hcg concentrations first in their own laboratories using different types of assays. These measurements cannot be compared and cannot be used for research purposes. This underlines the importance of central hcg measurements for patients with GTN. Only a few previous studies have investigated the effect of the presence of lung metastases on the tumour response to chemotherapy. Growdon et al. (2009) reported a higher number of administered chemotherapy courses in patients with lung metastases. 22 Nevin et al. (2000) and Chapman-Davis et al. (2012) observed a higher percentage of single-agent chemotherapy resistance in patients with lung metastases than in controls. 23,24 In the current study the MTX resistance and number of administered courses did not differ significantly. Conclusion Results from this study suggest that the presence of lung metastases is a risk factor for an unfavourable course of disease. This can only partially be explained by the higher incidence of other risk factors such as term antecedent pregnancy and higher serum hcg concentrations. Further research is needed to confirm these findings in larger patient cohorts from specialised centres. Independent of these future developments, patients with lung metastases need to be very closely monitored during and after therapy. However, it is too soon to propose an adjustment of the FIGO scoring system in which lung metastases score more than zero points. Disclosure of interests Full disclosure of interests available to view online as supporting information. Contribution to authorship Mireille Vree: data collection, data analysis, writing of the article. Nienke van Trommel: involved in development of the project, support with writing the article. Gemma Kenter: critical review of the article. Fred Sweep: collection and analysis of the serum hcg values, critical review of article. Marianne ten Kate-Booij: involved in data collection in past, critical review of the article. Leon Massuger: support of project and active involvement of registration of patients. Christianne Lok: initiator of the study, involved in development of project, support with data analysis and writing the article. 1844

7 Influence of lung metastases on clinical course of GTN Details of ethics approval Medical ethics approval of this study was not necessary because only routinely collected data were used and participants were not imposed a specific deed. Funding The study was funded by The Netherlands Cancer Institute, Amsterdam, The Netherlands, and the Radboud University Medical Centre, Nijmegen, the Netherlands. Acknowledgements No further acknowledgements. Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Dutch clinical classification system. Table S2. Characteristics of deceased patients. & References 1 Lurain JR. Gestational trophoblastic disease I: epidemiology, pathology, clinical presentation and diagnosis of gestational trophoblastic disease, and management of hydatidiform mole. Am J Obstet Gynecol 2010;203: Lybol C, Thomas CMG, Bulten J, van Dijck JAAM, Sweep FCGJ, Massuger LFAG. Increase in the incidence of gestational trophoblastic disease in The Netherlands. Gynecol Oncol 2011;121: Schmid P, Nagai Y, Agarwal R, Hancock B, Savage PM, Sebire NJ, et al. Prognostic markers and long-term outcome of placental-site trophoblastic tumours: a retrospective observational study. Lancet 2009;374: Bagshawe KD. Risk and prognostic factors in trophoblastic neoplasia. Cancer 1976;38: Altieri A, Franceschi S, Ferlay J, Smith J, La Vecchia C. Epidemiology and aetiology of gestational trophoblastic diseases. Lancet Oncol 2003;4: Hancock BW. Staging and classification of gestational trophoblastic disease. Best Pract Res Clin Obstet Gynaecol 2003;17: McNeish IA, Strickland S, Holden L, Rustin GJS, Foskett M, Seckl MJ, et al. Low-risk persistent gestational trophoblastic disease: outcome after initial treatment with low-dose methotrexate and folinic acid from 1992 to J Clin Oncol 2002;20: Lybol CD, Centen W, Thomas CMG, ten Kate-Booij MJ, Verheijen RHM, Sweep CGJ, et al. Fatal cases of gestational trophoblastic neoplasia over four decades in the Netherlands: a retrospective cohort study. BJOG 2012;119: Escobar PF, Lurain JR, Singh DK, Bozorgi K, Fishman DA. Treatment of high-risk gestational trophoblastic neoplasia with etoposide, methotrexate, actinomycin D, cyclophosphamide, and vincristine chemotherapy. Gynecol Oncol 2003;91: van der Houwen C, Rietbroek RC, Lok CA, Ten Kate-Booij MJ, Lammes FB, Ansink AC. Feasibility of central co-ordinated EMA/CO for gestational trophoblastic disease in the Netherlands. BJOG 2004;111: Bower M, Newlands ES, Holden L, Short D, Brock C, Rustin GJ, et al. EMA/CO for high-risk gestational trophoblastic tumours: results from a cohort of 272 patients. J Clin Oncol 1997;15: Miller JM Jr, Surwit EA, Hammond CB. Choriocarcinoma following term pregnancy. Obstet Gynecol 1979;53: Darby S, Jolley I, Pennington S, Hancock BW. Does chest CT matter in the staging of GTN? Gynecol Oncol 2009;112: Garner El, Garrett A, Goldstein DP, Berkowitz RS. Significance of chest computed tomography findings in the evaluation and treatment of persistent gestational trophoblastic neoplasia. J Reprod Med 2004;49: Yedema KA, Verheijen RH, Kenemans P, Schijf CP, Borm GF, Segers MF, et al. Identification of patients with persistent trophoblastic disease by means of a normal human chorionic gonadotropin regression curve. Am J Obstet Gynecol 1993;168: WOG. Richtlijn:Molazwangerschap. wangerschap, 2010 July 13 (1.3) Available from: URL: Kohorn EI. The new FIGO 2000 staging and risk factor scoring system for gestational trophoblastic disease: Description and critical-aassessment. Int J Gynecol Cancer 2001;11: Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivar Behav Res 2011;46: Ngan HYS, Chan FL, Au VWK, Cheng DKL, Ng TY, Wong LC. Clinical outcome of micrometastasis in the lung in stage IA persistent gestational trophoblastic disease. Gynecol Oncol 1998;70: Mutch DG, Soper JT, Baker ME, Bandy LC, Cox EB, Clarke-Pearson DL, et al. Role of computed axial tomography of the chest in staging patients with nonmetastatic gestational trophoblastic disease. Obstet Gynecol 1986;68: Lybol C, Sweep FCGJ, Harvey R, Mitchell H, Short D, Thomas CMG, et al. Relapse rates after two versus three consolidation courses of methotrexate in the treatment of low-risk gestational trophoblastic neoplasia. Gynecol Oncol 2012;125: Growdon WB, Wolfberg AJ, Goldstein DP, Feltmate CM, Chinchilla ME, Lieberman ES, et al. Evaluating methotrexate treatment in patients with low-risk postmolar gestational trophoblastic neoplasia. Gynecol Oncol 2009;112: Nevin J, Silcocks P, Hancock B, Coleman R, Nakielny R, Lorigan P. Guidelines for the stratification of patients recruited to trials of therapy for low-risk gestational trophoblastic tumor. Gynecol Oncol 2000;78: Chapman-Davis E, Hoekstra AV, Rademaker AW, Schink JC, Lurain JR. Treatment of nonmetastatic and metastatic low-risk gestational trophoblastic neoplasia: factors associated with resistance to singleagent methotrexate chemotherapy. Gynecol Oncol 2012;125: Lybol C, Thomas CM, Blanken EA, Sweep FC, Verheijen RH, Westermann AM, et al. Comparing cisplatin-based combination chemotherapy with EMA/CO chemotherapy for the treatment of high risk gestational trophoblastic neoplasia. Eur J Cancer 2013;49: