Background
Gestational trophoblastic diseases include premalignant (partial and complete hydatidiform moles) and malignant entities referred to as gestational trophoblastic neoplasia. Use of the International Federation of Gynecology and Obstetrics prognostic score is encouraged in cases of gestational trophoblastic neoplasia to predict the potential for the development of resistance to single-agent chemotherapy. An International Federation of Gynecology and Obstetrics score of ≥7 defines a high-risk patient and requires combination chemotherapy. Appropriate and rapid diagnosis, treatment by specialized centers, and reduction of early deaths at the time of chemotherapy initiation have led to significant improvements in survival for patients with high-risk gestational trophoblastic neoplasia. There is a crucial need for the early identification of high-risk patients with gestational trophoblastic neoplasia who have an increased death risk to organize their treatment in highly specialized centers.
Objectives
The purpose of this study was to describe cases of gestational trophoblastic neoplasia that have resulted in death, particularly in a subgroup with an International Federation of Gynecology and Obstetrics prognostic score of ≥13, for whom low-dose etoposide and cisplatin induction chemotherapy recently was shown to reduce early death rate.
Study Design
We identified 974 patients from the French Center for Trophoblastic Diseases who had a diagnosis of gestational trophoblastic neoplasia from November 1999 to March 2014. Among 140 patients who were at high risk of resistance to single-agent chemotherapy (International Federation of Gynecology and Obstetrics score, ≥7), 29 patients (21%) had a score of ≥13. Mortality rate was estimated with the use of the Kaplan-Meier method.
Results
The 5-year overall mortality rate, after the exclusion of placental site trophoblastic tumors and epithelioid trophoblastic tumors, was 2% for patients with gestational trophoblastic neoplasia (95% confidence interval, 1.25–3.13%). High-risk patients had a 5-year mortality rate of 12% (95% confidence interval, 7.49–18.9%). Patients with an International Federation of Gynecology and Obstetrics score of ≥13 had a higher 5-year mortality rate (38.4%; 95% confidence interval, 23.4–58.6%) and accounted for 52% of the deaths in the entire cohort. Early deaths, defined as those that occur within 4 weeks after treatment initiation, occurred in 8 patients of the entire cohort. Six of them had an International Federation of Gynecology and Obstetrics score of ≥13 at presentation, of whom 5 patients had brain and/or liver metastases.
Conclusion
Gestational trophoblastic diseases with an International Federation of Gynecology and Obstetrics score of ≥13 have an increased risk of early death. We suggest that an International Federation of Gynecology and Obstetrics score of ≥13 becomes a consensual criterion for prediction of patients with gestational trophoblastic neoplasia with increased risk of death, particularly early death. These patients justify treatment in highly specialized gestational trophoblastic disease centers and may benefit from the use of induction low-dose etoposide and cisplatin.
Gestational trophoblastic diseases (GTD) include premalignant (partial and complete hydatidiform moles) and malignant entities that are referred to as gestational trophoblastic neoplasia (GTN), which encompass invasive mole, gestational choriocarcinoma, placental site trophoblastic tumor (PSTT), and epithelioid trophoblastic tumor (ETT), the latter 2 being very rare entities. The French Center for Trophoblastic Diseases has been receiving voluntary declarations of GTD cases since 1999; it provides pathologic reviewing, supports clinicians for the choice of treatment according to the International Federation of Gynecology and Obstetrics (FIGO) score, and organizes prospective follow-up evaluations. To date, >5000 patients with GTD have been registered in the French Center for Trophoblastic Diseases, of whom >1000 patients had GTN. Use of the FIGO prognostic scoring system is encouraged strongly in cases of GTN to predict the potential for the development of resistance to single-agent chemotherapy. This prognostic score does not apply to PSTT and ETT because single-agent chemotherapy should not be the first therapeutic choice for these rare tumors. A score of ≤6 defines a low-risk of resistance, and patients should be placed under methotrexate or dactinomycin. A score of ≥7 defines a high risk and requires combination chemotherapy. The regimen that consists of alternating EMA (etoposide, methotrexate, and dactinomycin) every week with CO (cyclophosphamide and vincristine) is the most widely used regimen worldwide, even in the presence of other possible regimens, such as bleomycin, etoposide, and cisplatin (BEP) or dactinomycin, cisplatin, and etoposide.
Among the prognostic criteria that influence survival, the largest tumor mass diameter, the nature of the antecedent pregnancy, and the sites of metastases take part in the FIGO score. A high FIGO score is associated with poor survival, where death is not only linked to chemoresistance but also to early and severe complications, such as hemorrhagic metastases, infection, multisystem organ failure, or tumor lysis syndrome. Appropriate and rapid diagnosis, treatment by specialized centers, and reduction of early deaths because of chemotherapy initiation have led to significant improvements in survival for patients with high-risk GTN. A recent study showed a 10-fold reduction in early deaths with the use of induction low-dose etoposide and cisplatin (EP) in 33 patients with high-risk GTN after exclusion of nongestational diseases by genetic analysis. In this study, induction chemotherapy was given to patients with a large disease burden, particularly within the thorax, and to patients with brain and/or liver metastases. These patients had a higher number of metastases, higher levels of serum human chorionic gonadotropin (hCG), and/or higher total FIGO scores. Therefore, a low-dose EP-induction protocol was implemented in 2012 in the French Center for Trophoblastic Diseases.
Here we present a retrospective analysis of the GTN cohort from the French Center for Trophoblastic Diseases (from 1999-2014), focusing on death rates and the influence of FIGO scores of ≥13 as a threshold for the identification of patients who are at risk of early death.
Material and methods
This study was carried out at the French Center for Trophoblastic Diseases in Lyon. Informed consent for the prospective registration at the French Center for Trophoblastic Diseases was obtained from each patient, and ethical approval was guaranteed from the local ethical authority. Every registered patient’s diagnosis was confirmed by our referent pathologists. We identified 974 patients who had been diagnosed with a GTN between November 1999 and March 2014 from our electronic database. Patient characteristics are listed in Tables 1 and 2 .
International Federation of Gynecology and Obstetrics | Total number of patients (n = 974), n (%) | International Federation of Gynecology and Obstetrics score, n (%) | Placental site trophoblastic tumor-epithelioid trophoblastic tumor (n = 33), n (%) | Spontaneous regression (n = 7), n (%) | |
---|---|---|---|---|---|
≤6 (n = 794) | 7 (n = 140) | ||||
stage | |||||
I | 743 | 663 (83.6) | 56 (40) | 19 (70.4) | 5 |
II | 39 | 31 (3.9) | 6 (4.3) | 2 (7.4) | |
III | 140 | 97 (12.2) | 39 (27.9) | 4 (14.9) | |
IV | 43 | 2 (0.25) | 39 (27.9) | 2 (7.4) | |
Unknown | 9 | 1 | 0 | 6 | 2 |
score | |||||
0-4 | 685 (73) | 682 (86) | 0 | NA | 3 (60) |
5-6 | 113 (12) | 112 (14) | 0 | NA | 1 (20) |
7-12 | 112 (12) | 0 | 111 (79) | NA | 1 (20) |
≥13 | 29 (3) | 0 | 29 (21) | NA | 0 |
Unknown or not applicable | 35 | 0 | 0 | NA | 2 |
Clinical characteristic | Total patients (n = 941) | International Federation of Gynecology and Obstetrics score | ||
---|---|---|---|---|
≤6 (n = 794 ) | ≥7 (n = 140) | ≥13 (n = 29 ) | ||
Median age, y (range) | 33 (12–58) | 32 (12–57) | 37 (18–58) | 32 (18–50) |
Median human chorionic gonadotropin before treatment, IU/L | 10,290 | 6,723 | 205,450 | 203,500 |
Range a | 4–4,022,380 | 4–1,916,972 | 107–4,022,380 | 107–3,228,579 |
Missing data, n | 23 | 10 | 5 | 0 |
Antecedent pregnancy, n | ||||
Mole | 787 | 731 | 53 | 1 |
Abortion | 77 | 39 | 37 | 11 |
Birth | 72 | 19 | 50 | 17 |
Unknown | 5 | 5 | 0 | 0 |
International Federation of Gynecology and Obstetrics, n | ||||
Interval from antecedent pregnancy, y | ||||
0 | 761 | 681 | 76 | 6 |
1 | 101 | 88 | 13 | 6 |
2 | 41 | 21 | 18 | 6 |
4 | 38 | 4 | 33 | 11 |
Number of metastases score | ||||
0 | 750 | 682 | 61 | 0 |
1 | 137 | 101 | 36 | 3 |
2 | 18 | 10 | 8 | 2 |
4 | 36 | 1 | 35 | 24 |
Site of metastases score | ||||
0 | 903 | 792 | 104 | 8 |
1 | 5 | 2 | 3 | 0 |
2 | 6 | 0 | 6 | 1 |
4 | 27 | 0 | 27 | 20 |
Liver metastasis | ||||
n/N (%) | 15/902 (1.7) | 0/765 | 15/133 (11.3) | 10/29 (34.5) |
Not evaluated, n | 36 a | 29 | 7 | 0 |
Brain metastasis | ||||
n/N (%) | 17/833 (2.0) | 0/698 | 17/132 (12.9) | 14/29 (48) |
Not evaluated, n | 104 a | 96 | 8 | 0 |
Median follow-up period, y (range) | 5 (0–15.1) | 5.2 (0.6–15.1) | 4.5 (0–15.1) | 2.85 (0–12.8) |
Total deaths, n | 18 | 2 | 16 | 11 |
Early deaths, n | 7 | 0 | 7 | 6 |
a Cases of human chorionic gonadotropin spontaneous normalization were excluded.
At the time of survival calculation, 4 patients were still under treatment at 15, 17, 20, and 25 months after treatment initiation. The minimum follow-up duration after data extraction in March 2014 was 1 year. The median follow-up period of the entire cohort was 5 years (range, 0-15 years); subgroup-specific follow-up duration is reported in Table 2 . For 5 patients who had 2 distinct cases of GTN, we analyzed only the latest 1. A FIGO score of ≥13 was chosen as a cut-point because of its reproducibility to identify patients with increased risk of early death because of a large disease burden. This criterion was present in 45% of patients who were treated by induction low-dose EP by Alifrangis et al. Among patients with a FIGO score of ≥13, 16 patients who were identified by the French Reference Center for Trophoblastic Diseases were treated in regional expert centers; 6 patients were treated in specific anticancer centers, and 7 patients were treated in general hospitals, of whom 3 were treated outside metropolitan France. First-line treatment of high-risk patients with a FIGO score of ≥13 consisted in EMA-CO (n = 14), induction low-dose EP (n = 6), BEP (n = 4) and dactinomycin, cisplatin, and etoposide (n = 4); 1 patient died before treatment initiation. The total serum hCG concentration was measured by automated immunochemiluminescence analyzers from the laboratory of each patient center weekly until 1 month after the end of treatment, then monthly during 12 or 18 months for single-agent or combination chemotherapy, respectively. Clinical and hCG follow-up evaluation was then eased. For PSTT and ETT, hCG follow-up evaluation was extended. After serum hCG normalization, consolidation consisted of 2 cycles of chemotherapy. FIGO scoring and staging was enabled by pelvic Doppler ultrasound scanning, abdomen and chest computed tomography (followed by chest X-ray in case of lung metastasis), and brain and pelvic magnetic resonance imaging. The latest investigation was considered as reference for tumor size assessment. Any suspicion of liver metastasis was further explored by abdomen magnetic resonance imaging. FIGO score and stage were extracted from our database with PARADOX 9 software (Corel, Ottawa, Canada), except for patients who were registered before 2002, whose cases subsequently were rescored (n = 49). Early death was defined as that which occurred within 4 weeks after treatment initiation. Survival was calculated from the first day of treatment to 1 year after data extraction or to the date of death. Survival was estimated according to the FIGO score with the use of the Kaplan-Meier method, and a Wilcoxon test was used to compare survival curves among groups. Statistical analysis was performed with SAS software (version 9.2; SAS Institute, Cary, NC).