Prognosis of patients with relapsed and chemoresistant gestational trophoblastic neoplasia transferred to the Peking Union Medical College Hospital
Dr Y Xiang, Department of Obstetrics and Gynecology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Beijing 100730, China. Email firstname.lastname@example.org
Objective To analyse and compare the treatment outcome of the patients with gestational trophoblastic neoplasia (GTN) who were transferred to the Peking Union Medical College Hospital (PUMCH) because of chemoresistant or relapsed GTN.
Design A retrospective study of the clinical features and treatment outcome of GTN after treatment failure elsewhere at the PUMCH.
Setting The Department of Obstetrics and Gynecology, PUMCH, Peking Union Medical College and Chinese Academy of Medical Sciences, Beijing, China.
Population A total of 81 women with chemoresistant or relapsed GTN transferred and treated between January 2005 and December 2007.
Methods A review of medical records of consecutively collected, referral cases of GTN was performed. The patients were classified according to the reasons for referral as having chemoresistant GTN if they had an incomplete response to previous treatment, and as having relapsed GTN if they had elevated serum β-human chorionic gonadotropin (β-hCG) levels, in the absence of a pregnancy, after the completion of initial treatment.
Main outcome measures Serologic complete remission (SCR) and relapse rates.
Results The patients with chemoresistant and relapsed GTN achieved 52.6 and 76.7% of the long-term SCR rate, respectively; P = 0.035. Relapse rate and median time to relapse were similar between the chemoresistant GTN cohort and the relapsed GTN cohort after initial SCR (24 versus 35.7%, P = 0.417; 4 versus 3 months, P = 0.969).
Conclusions The patients with chemoresistant GTN had a worse outcome compared with patients with relapsed GTN. The importance of obtaining a normal β-hCG level with treatment for GTN should be emphasised, and novel therapies should be pursued for these patients.
As a result of the introduction of effective chemotherapy, the ability to monitor treatment progress with human serum β-human chorionic gonadotrophin (β-hCG) levels and the individualised chemotherapy based on prognostic factors, gestational trophoblastic neoplasia (GTN) has become a highly curable tumour. At present, 90% or more of patients with GTN are cured with chemotherapy because of the inherent chemosensitivity of this disease: a minority of patients have an incomplete response to chemotherapy or relapse after the completion of initial treatment.1–4
As the largest GTN centre in China, the Peking Union Medical College Hospital (PUMCH) often admits patients with GTN after treatment elsewhere, particularly those who have become chemoresistant or have relapsed after failure of treatment elsewhere.5–7 The treatment outcome of the chemoresistant or relapsed patients after treatment failure would have an adverse impact on the ultimate cure rate of GTN. However, to date, there is limited information about comparison of patients with chemoresistant and relapsed GTN after failure of treatment elsewhere. Additionally, a growing body of evidence in other curable cancer literature suggests that patients with refractory or relapsed disease have a different prognosis.8,9 In this study, we analysed and evaluated the treatment outcomes of patients with relapsed and chemoresistant GTN transferred to the PUMCH after treatment elsewhere.
This study was retrospective, conducted between January 2005 and December 2007 at the PUMCH. The cut-off date of follow-up was 28 February 2009.
On admission to the PUMCH, each patient underwent a complete history-taking and physical examination, a complete blood count, renal and liver function tests, serum β-hCG level, pelvic and liver ultrasonography, chest X-ray and/or computed tomography (CT). If the chest X-ray and/or chest CT showed the presence of pulmonary metastases, brain/abdomino-pelvic magnetic resonance imaging (MRI) or CT were performed. Patients were staged according to the internationally accepted criteria for GTN, and were then classified into low-risk or high-risk categories.10 Additionally, patients were classified according to the reasons for referral: as chemoresistant GTN if they had a plateau (<50% fall in the β-hCG levels) or an increased β-hCG level and/or development of new metastases for at least two consecutive cycles of combined chemotherapy, and relapsed GTN if they had two elevated serum β-hCG levels in the absence of a pregnancy after achieving a minimum of 4 weeks of normal β-hCG values with treatment. The patients with placental site trophoblastic tumours (PSTTs) were excluded from this analysis because PSTT has a rather variable chemosensitivity.
After admission to the PUMCH, all patients were primarily treated with combined chemotherapy. Several combined chemotherapy regimens were used, based on the previous chemotherapy protocols that patients had received in other hospitals. Commonly used chemotherapy regimens included: vincristine, floxuridine (FUDR), and dactinomycin (FAV); vincristine, FUDR, dactinomycin, and etoposide (FAEV); etoposide, methotrexate, and dactinomycin, alternating with cyclophosphamide and vincristine (EMA-CO); or etoposide, methotrexate, and dactinomycin, alternating with etoposide and cisplatin (EMA-EP). Serum β-hCG levels were measured before each cycle of chemotherapy. Chemotherapy was changed if serum β-hCG levels reached a plateau or began to increase for two consecutive cycles of combined chemotherapy.
In general, the majority of patients were primarily treated with the FAEV regimen. EMA-CO or EMA-EP was used if the response to FUDR-based combination chemotherapy was unsatisfactory. If patients remained refractory, almost all salvage regimens were platinum-based. When central nervous system (CNS) metastases were present, the administration of intrathecal methotrexate (MTX) was often used concomitantly with systemic chemotherapy. Surgery was used in selected patients with limited resistant or persistent disease. Before deciding to perform surgical resection, CT or MRI of the brain, chest, abdomen, and ultrasonography of the pelvis were performed to document the presence of uterine disease or disease elsewhere. If there was no evidence of distant metastasis, hysterectomy or hysterotomy for excision of the lesion in the uterus was performed, based on a desire to maintain fertility. If the disease was isolated in distant metastasis sites, and there was no evidence of an active tumour at other sites, surgical extirpation of metastatic disease was performed. On the other hand, the preoperative β-hCG level should be as low as possible.
Serum β-hCG levels were measured weekly to monitor the response during the treatment. A comprehensive evaluation had to be performed once normal serum β-hCG levels (of <2 iu/l) were attained. Serologic complete remission (SCR) was defined as normal serum β-hCG levels in at least four consecutive weekly determinations. A partial remission (PR) was defined as a greater than 50% decrease in serum β-hCG levels, as compared with the pre-treatment value. No response (NR) was defined as a 50% or less decrease in serum β-hCG levels, compared with the pre-treatment value, or an increase in the serum β-hCG level that continued for at least two consecutive cycles of chemotherapy. After the completion of treatment, the serum β-hCG level was measured weekly for 4 weeks, then biweekly up to 3 months, followed by monthly measurements for the rest of the first year, then every 3 months in the second year, and finally every 6 months to 1 year for the rest of their lives.
Categorical variables were evaluated by Pearson’s chi-square test or Fisher’s exact test, as appropriate for category size. Continuous variables were evaluated by a Student’s t-test or a Wilcoxon–Mann–Whitney U test for normally or non-normally distributed variables, respectively. All analyses were performed using spss 12.0 (SPSS, Chicago, IL, USA), and all reported P-values were two-sided.
The characteristics of the chemoresistant or relapsed patients
From January 2005 to December 2007, 81 patients had treatment failure before referral to the PUMCH. Of the 81 patients, 38 were defined as having chemoresistant GTN and 43 were defined as having relapsed GTN. The previous treatments and patient’s characteristics in the two groups are summarised in Table 1. Patient characteristics were similar. All patients had received chemotherapy, and the number of cycles was equivalent. In addition, 11 patients (28.9%) in the chemoresistant group and 11 (25.6%) in the relapsed group had also undergone surgical management (P = 0.734). However, the number of previous chemotherapy regimens was significantly higher in the chemoresistant GTN group compared with the relapsed GTN group (3 versus 2; P = 0.006).
Table 1. Summary of the chemoresistant and relapsed patient characteristics (n = 81)
|Median age (range)||31 (15–47)||28 (19–50)||0.545|
|Hydatidiform mole||15 (39.5%)||21 (49%)||0.397|
|Term pregnancy||8 (21%)||10 (23%)||0.812|
|Abortion/unknown||15 (39.5%)||12 (28%)||0.270|
|Median cycle (range)||9 (2–25)||8 (4–26)||0.965|
|Median regimens (range)||3 (1–7)||2 (1–6)||0.006|
|Previous surgery||11 (28.9%)||11 (25.6%)||0.734|
|Sites of metastases|
|Lung||29 (76%)||30 (70%)||0.508|
|Liver||1 (3%)||2 (5%)||0.631|
|Brain||4 (11%)||3 (7%)||0.570|
|Vagina||2 (5%)||1 (2%)||0.598|
|Others||4 (1 spleen, 1 adrenal, 1 bone, 1 parametrium)||4 (2 kidney, 2 skin)||–|
|No metastasis||8 (21%)||12 (28%)||0.475|
|No. of disease sites|
|0||4 (11%)||6 (14%)||0.743|
|1||16 (42%)||24 (56%)||0.218|
|2||15 (39%)||10 (23%)||0.115|
|3||2 (5%)||3 (7%)||1.000|
|Median serum β-hCG, U/l (range)||1887 (6–247 041)||217 (3–1 239 310)||0.296|
|Median FIGO prognostic score (range)||10 (2–18)||9 (4–21)||0.098|
|High risk (≥7)||35 (92%)||37 (86%)||0.490|
|Low risk (<7)||3 (8%)||6 (14%)||0.490|
Initial response of the chemoresistant or relapsed patients at the PUMCH
The treatment and the corresponding results of these 81 patients at the PUMCH are presented in Table 2. Treatment protocols were formulated based on a number of factors, the most important of which were the previous treatment plan and stage of disease. The majority of patients received FUDR-based combination chemotherapy, in which the most common regimen was FAEV, used in approximately 97% of the patients in the two groups. EMA-CO was preferred in patients who failed with the FAEV regimen, although other regimens were also used.
Table 2. Management and outcome of the 81 patients at Peking Union Medical College Hospital (PUMCH)
|Type of regimens|
| FAV||3 (7.9%)||3 (7%)||1.000|
| FAEV||34 (89.5%)||39 (90.7%)||1.000|
| Others||3 (1 AE, 1 MBE, 1 PVB)||1 (AE)||–|
|Median regimens||2 (1–3)||1 (1–2)||0.000|
|No. of patients needing to change regimens||24||11||0.001|
| Because of resistance||24 (63.2%)||10 (23.3%)||0.000|
| Because of toxicity||0||1||–|
|Surgery||20 (52.6%)||18 (41.9%)||0.332|
|Hysterotomy + thoracotomy||1||0|| |
|Resection of vaginal metastasis||1||0|| |
|SCR||25 (65.8%)||42 (97.7%)||0.000|
|NR||12 (31.6%)||1 (2.3%)||0.000|
Surgical management combined with chemotherapy was performed in 20 (52.6%) patients with chemoresistant GTN, and in 18 (42%) patients with relapsed GTN. Among patients who underwent surgery, 75% (15/20) in the chemoresistant group and 100% (18/18) in the relapsed group achieved SCR.
Significant differences were noticed in the number of regimens and number of patients who needed to change regimens because of resistance between the chemoresistant and relapsed groups (2 versus 1, P = 0.000; 63.2 versus 23.3%, P = 0.000). At the completion of treatment, higher numbers of SCR were achieved in the patients with relapsed GTN (97.7%, 42/43) compared with those of chemoresistant GTN (65.8%, 25/38), P = 0.000.
Outcome and follow-up of the chemoresistant or relapsed patients at the PUMCH
At a median follow-up of 22 months (6–48 months), 21 patients in total (31.3%) from the two groups had relapsed disease: six patients from the chemoresistant group and 15 patients from the relapsed group. However, there is no difference in the relapse rate in these two groups (24 versus 35.7%; P = 0.417), and there is also no difference in the median time to relapse (4 months versus 3 months; P = 0.969). The majority of relapses (about 80%) occurred within 6 months of completion of treatment, and only one patient relapsed after more than 1 year of follow-up (Table 3).
Table 3. Time to recurrence after stopping initial treatment at the Peking Union Medical College Hospital (PUMCH)
|0–3||2 (33.3)||33.3||9 (60)||60|
|3–6||3 (50)||83.3||3 (20)||80|
|6–12||1 (16.7)||100||2 (13.3)||93.3|
|12–24|| || ||1 (6.7)||100|
At the last follow-up, of the 21 relapsed patients after the completion of initial treatment at the PUMCH, 13 patients progressed and died of GTN, including five patients from the chemoresistant group and eight patients from the relapsed group, even though the serum β-hCG level was temporarily normalised in three of these patients; seven patients again achieved SCR by salvage chemotherapy with or without surgery; one patient developed a relapse for a third time, and was receiving chemotherapy. One patient with PR was lost to follow-up and presumed dead. Thirteen patients with NR died of disease. Ultimately, 53 (65.4%) patients were in SCR, including 20 (52.6%) of the 38 patients in the chemoresistant group and 33 (76.7%) of the 43 patients in the relapsed group (P = 0.035).
Owing to the rarity and inherent chemosensitivity of GTN, there are a small number of patients who relapse after treatment or develop chemoresistant disease after chemotherapy, and publications examining the management and outcome of these patients often do not distinguish between chemoresistant and relapse disease.11,12 In this study, we compared the response with treatment, relapse rate, and interval to relapse, between the chemoresistant and relapsed patients. The main results are that the overall remission rate of the relapsed patients (76.7%) is significantly higher than that of the chemoresistant patients (52.6%), and there is no difference in relapse rate and median time to relapse after an initial remission achieved between the two groups.
Gestational trophoblastic neoplasia (GTN) was highly sensitive to chemotherapy, but needed to be treated by physicians experienced in the management of this disease: the overall cure rate was over 98% in specialised cancer centres.1 As the largest GTN centre in China, the PUMCH often admits patients with GTN after treatment failure elsewhere. The main causes of failure were inappropriate chemotherapy regimens, inadequate timing and dosing of chemotherapy, inadequacy and unavailability of the β-hCG test, and a lack of experience in managing such disease in some hospitals.13,14 Despite subsequent treatment at the PUMCH for patients who had already received treatment elsewhere, the number of SCRs was significantly lower than in the patients treated primarily at the PUMCH, and more patients relapsed after the treatment. Therefore, treatment of GTN in specialised medical centres should be emphasised.
Analysis of these patient characteristics in this study showed that the majority had been treated by currently available chemotherapy protocols, with or without surgery, before being referred to our hospital. Based on the data from these patients, we could not perform different treatment protocols between the chemoresistant and relapsed patients. Therefore, after being referred to our hospital, the patients only received an aggressive combination chemotherapy including FAEV, EMA/CO, EMA/EP, ICE, and TP/TE, with between two and four additional courses of consolidation chemotherapy. Although individual cases have been reported where high-dose chemotherapy with autologous bone marrow or stem cell support has induced SCR in patients with refractory GTN, the potential role of high-dose chemotherapy is still unproven.15,16 Therefore, in the present study, no patients were treated with high-dose chemotherapy. On the other hand, the effectiveness of salvage surgery in the selected patients with chemoresistant and relapsed GTN has been demonstrated by several authors.2,17,18 In our study, most of the patients also underwent additional surgical resection of chemoresistant and relapsed disease. However, the long-term remission rate in these patients was only 65.4%, significantly lower than the overall cure rate of GTN at our hospital.6 Therefore, there is a need to develop other effective regimens, and further therapy should be explored and evaluated in this group in which treatment has failed elsewhere.
As a specialist centre for GTN, a group from the Charing Cross Hospital, UK, have reported on the outcome of patients with relapsed and chemorefractory GTN who had received initial treatment at the Charing Cross Hospital. The overall 5-year survival for patients with relapsed and chemorefractory GTN were 93 and 43%, respectively.19 As we have shown in this study, chemoresistant and relapsed patients have differing outcomes, and the outcome of patients with relapsed GTN is good. Thus, the importance of obtaining a normal β-hCG level with treatment for GTN was underlined. Additionally, although patient characteristics and treatment protocols in the two groups were similar, because the median number of previous and present chemotherapy regimens in the relapsed group was lower, the higher remission rate in this group is more likely to result from less chemotherapy regimens. Moreover, analysis of patients who had chemotherapy regimens changed because of resistance in the relapsed and chemoresistant groups (23.3 versus 63.2%, respectively) also showed that the patients in the relapsed group were more sensitive to chemotherapy. This information emphasised the importance of classifying GTN with previous failed treatments as either chemoresistant or relapsed GTN according to serial measurement of the serum β-hCG level after previous treatment.
It has been reported that relapse from remission occurs in 2–3% of patients with non-metastatic GTN, in 4–5% of metastatic patients with favourable prognostic factors, and in 13–20% of metastatic patients with unfavourable prognostic factors.7,20,21 The median interval from remission to relapse was 6–6.5 months.22 In the present study, although remarkable differences were found in the remission rate between the chemoresistant and relapsed groups, there was no difference in relapse rate and median time to relapse after the initial remission achieved at the PUMCH between the two groups, with a relapse rate of 24 and 35.7%, and median time to relapse of 4 and 3 months, in the chemoresistant group and the relapsed group, respectively. The relapse rate and median time to relapse in these two groups were higher and shorter than patients with poor-prognosis GTN reported in previous reviews.7,14,20 Moreover, at median follow-up of 22 months, there were 13 patients who had again progressed to NR after initial CR. This result suggests that the patients with treatment failure elsewhere may have occult chemoresistant metastases, which relapse more than once within a short period of time after the completion of treatment, even after a normal β-hCG level was achieved with salvage treatment. Therefore, in the present study we emphasise the need for close surveillance of these chemoresistant or relapsed patients, to assure sustained remission.
In conclusion, the present study confirmed the importance of the management of GTN in specialist centres. As there were different outcomes for the patients with the chemoresistant GTN, as opposed to patients with relapsed GTN, it is important to distinguish between the chemoresistant and relapsed GTN groups. Closely monitoring patients after SCR is recommended, especially in the first year. Meanwhile, there is a need to develop other effective therapeutic modalities to improve the outcome of these patients. Because of the rarity of GTN, we support the combined study of chemoresistant patients and relapsed patients in GTN treatment trials, but believe that the results on chemoresistant and relapsed GTN should be reported independently.
Disclosure of interests
The authors declare no conflict of interest.
Contribution to authorship
All authors contributed to writing the paper, revising it and approved the final version to be published. In addition, FF and YX contributed to the conception and design of the project, and FF carried out the data acquisition and analysis.
Details of ethics approval
Ethics approval was not applied for, as this was a retrospective analysis of clinical work.
The study was financially supported by the Key Projects in National Science & Technology Pillar Program in the 11th 5-year plan period (grant No. 2008BAI57B05).
We thank all of the patients. Without their continuing data collection, this study would not have been possible.