Discussion
The current study found that liver or brain metastases were exceptionally rare, which is consistent with previous reports6, 9, 10.However, early death and a dramatically high drug-resistance rate with initial treatment remains a critical problem in ultra-high-risk GTN patients with liver and brain metastases.
We found that liver or brain metastases accounted for 1.3% and 2.7% in all GTN patients, respectively, which is supported by previous reports from China9, 10, France12, and the United Kingdom (UK) 7, 11. At Peking Union Medical College Hospital, GTN patients with liver metastases reportedly accounted for 1.9%9 and brain metastases accounted for 3.4% of patients 10, respectively. Similarly, the incidence of liver metastases was only 1.8% (38/2100) of all GTN patients in the Charing Cross GTN database from 1975 to 200716 and the incidence of brain metastases in GTN patients was only 1.7% (21/1251) in a 17-year retrospective study in France12. The difference between studies in Asia and Europe could be owing to differences in prevalence, discrepancies between hospital-based and population-based data, or disparities in the availability of central pathology review. The reason for the rare prevalence of liver or brain metastases might be related to sensitivity to chemotherapy in patients with GTN.
However, liver metastases (HR: 34.05; 95% CI: 1.65–703.7;P =0.02) and brain metastases (HR: 49.19; 95% CI: 5.6–432.1;P =0.01) were found to be independently significant risk factors for death in GTN patients in our study. Similar findings have been reported, showing that the presence of liver or brain metastases is a strong indicator of a poor outcome in GTN 5, 7, 9, such as cerebral hemorrhage 12, neurological sequelae12, less than 40% 5-year survival rate7 and early death 2. The survival rate of GTN patients with brain metastases is only 35%–70%10, 17, and 27%–48% 7, 16 for liver metastases. Additionally, we found a significantly lower CR rate (0% vs. 28.6%, P <0.01) and higher drug-resistance rate (90.0% vs 42.9%, P <0.01) after initial treatment between the subgroups with and without liver or brain metastases in ultra-high-risk GTN patients. Hence, this confirms that liver or brain metastases is a crucial risk factor for ultra-high-risk GTN patients, which can lead to poorer outcome than ultra-high-risk GTN patients without liver or brain metastases.
In this study, one ultra-high-risk patients with concomitant liver and brain metastases had early death only four days after treatment initiation (before adequate chemotherapy could be given). The patient died from cerebral hernia and multiple organ failure. Similarly, other studies have reported that early death remains a critical problem in GTN patients with liver and brain metastases. One study found that early death was significantly associated with ultra-high-risk GTN, occurring in 13.8% of these patients 2. Many deaths happened soon after admission for hemorrhage or metabolic results of overwhelming disease4. When deaths within four weeks were excluded, survival in patients with brain metastases (86%) was equivalent to that for other patients 18. Similarly, in 37 patients with liver metastases treated between 1977 and 2005, OS increased to 48% at five years. However, when early deaths were excluded, survival was 68% in a UK study 16. For GTN patients with liver or brain metastases and massive disease, starting with standard first-line multidrug chemotherapy may cause sudden tumor collapse with severe bleeding, metabolic acidosis, myelosuppression, septicemia, and multiple organ failure, any or all of which can result in early death6. To minimize early deaths in patients with very advanced disease, starting chemotherapy with low-dose induction EP chemotherapy (etoposide 100 mg/m2 and cisplatin 20 mg/m2 on days 1 and 2) before commencing EMA-CO may help to reduce tumor edema4 and can remarkably reduce the early death rate from 7.8% to 0.7%19. Additionally, EP chemotherapy can enable a more gradual reduction in tumor bulk during the initial weeks of treatment to minimize the risk of early death 19. Several series have also suggested that the use of low-dose induction etoposide and cisplatin may benefit GTN patients with a FIGO score ≥13 2, 6, 12, 20, 21, especially if their increased risk score is due to a large tumor burden or metastases to the brain, liver, or extensive metastases6, 14 owing to hemorrhagic sequelae at the tumor sites22.
Ten patients among 11 ultra-high-risk GTN patients with liver or brain metastases were treated with EMA-CO as first-line chemotherapy in our study. The results were consistent with previous studies, showing that the most commonly used multiple-agent chemotherapy for high-risk GTN worldwide remains the EMA-CO protocol 6, 11, 14, 19, 23-25, which is considered to have the best effectiveness-to-toxicity ratio13. The cumulative 5-year survival rate of patients given EMA-CO is between 75% and 90% 4, 19. The OS rate was 86.2% in all GTN patients after EMA-CO and 85.4% in the high-risk group in a UK study 19. However, Bolze et al. reported that the 5-year death rate approached 38.4% in patients with FIGO score ≥13 treated with EMA-CO, with or without low-dose EP20.Furthermore, a Cochrane review found that 20% of patients do not achieve a complete response with EMA-CO therapy6, and most can ultimately be salvaged with TP/TE or EP-EMA 6, 12, 21, 26.
In this study, the drug-resistance rate reached 90.9% with initial treatment in ultra-high-risk GTN patients with liver or brain metastases. A retrospective series in India including 82 high-risk GTN patients suggested that the resistance or relapse rate was only 31 (37.8%) after EMA-CO therapy 27. This discrepancy might be owing to differences in study participants because the study in India only reported the drug-resistance rate with initial treatment in high-risk GTN patients rather than ultra-high-risk patients. The extremely high drug-resistance rate with initial treatment in GTN patients with liver or brain metastases in our study suggests that the combination of chemotherapy with adjuvant immunological therapy or surgery is needed, in comparison with chemotherapy alone, as initial treatment 28. Immunological reactions might be involved in the development of GTN from complete hydatidiform mole (CHM)29. Recent work suggests that checkpoint immunotherapies represent an important new approach for the management of drug-resistant GTN 2, 3, 5, 8. Additionally, poly (ADP-ribose) polymerase inhibitors (PARP) and anti-angiogenesis agents are novel treatments for drug-resistant GTN 15.
In the present study, adjuvant surgical procedures were performed in eight (72.7%) ultra-high-risk GTN patients with liver or brain metastases as a component of their therapy, among whom six ultimately achieved long-term survivors. Similarly, several studies have found that adjuvant surgery might be an effective choice for high-risk GTN patients5, 6, 10-12, 15, 21, 30, which is in accordance with our results. Previous studies have confirmed that surgery not only reduces the tumor burden but also removes isolated chemoresistant lesions, such as in the lung or brain 2, 3, 5, 31, 32. Hence, surgery (RR 0.336, 95% CI 0.177–0.641, P =0.001) is a protective factor in the prognosis of ultra-high-risk GTN patients5. Generally, nearly 50% of GTN patients with high-risk disease require some surgical intervention to achieve a cure33. Hysterectomy can be considered with uncontrolled uterine bleeding and laparotomy might be needed to stop bleeding in organs such as the liver, gastrointestinal tract, kidneys, and spleen6, 21, 31. To decrease the early death rate, neurosurgery is needed if there is bleeding into the brain or increased intracranial pressure 6, 11, 12, 21. However, with the availability of uterine artery embolization, hysterectomy can often be avoided 31. Additionally, selective angiographic embolization is used to control hemorrhage from multiple liver metastases30.
In our study, no patients had whole-brain radiation therapy, but three patients had stereotactic radiotherapy for residual brain metastases at the end of treatment. The strategy for treating GTN patients with brain metastases at our center is intravenous multidrug chemotherapy and intrathecal methotrexate. Stereotactic radiotherapy was used for 37.5% (3/8) GTN patients with brain metastases for residual brain metastases at the end of treatment. However, concomitant whole-brain radiotherapy has rarely been used at our center owing to intellectual impairment in patients over the long term. Our results are consistent with those of previous studies showing that radiotherapy plays a limited role in the treatment of brain 6, 12, 21 and liver metastases7 in GTN patients. An observational study performed at Peking Union Medical College Hospital found that only 0.9% (2/109) of GTN patients with brain metastases received brain irradiation during 1990–2013 because this can induce long-term intellectual impairment in patients who are cured 10. However, some centers may administer whole-brain radiotherapy (3000 cGy in 200 cGy daily fractions) concurrent with chemotherapy or use stereotactic or gamma knife radiation to treat existing or residual brain metastases after chemotherapy 6. Neubauer et al. recommended an approach using whole-brain irradiation combined with systemic multi-agent chemotherapy to treat patients with brain metastases; however, their reported OS was only 50%34. In the UK, only one of 46 GTN patients with liver metastases received liver irradiation between 1958 and 19947.
Although this study was limited owing to its retrospective design and relatively small sample size, we reported our experience in the management of ultra-high-risk GTN patients with liver or brain metastases. Early death and the remarkably high drug-resistance rate to initial chemotherapy remain two critical problems in GTN patients with liver or brain metastases. We speculate that the combination of chemotherapy with adjuvant immunological therapy or surgery, compared with chemotherapy alone, as initial treatment may improve patient prognosis. RCTs are needed to confirm our hypothesis. Furthermore, starting chemotherapy with low-dose induction EP chemotherapy before commencing EMA-CO might be helpful in these patients. Further work is essential to explore optimal treatment strategies in this patient population. Specialized multidisciplinary teams and tertiary specialist centers are critical to the management of GTN.