Gestational trophoblastic disease with liver metastases: the Charing Cross experience

Transcription

1 British Journal of Obstetrics and Gynaecology January 1997, Vol. 104, pp Gestational trophoblastic disease with liver metastases: the Charing Cross experience *$Robin A. F. Crawford Senior Registrar (Gynaecological Oncology), TE. Newlands Director/Professor (Medical Oncology), **G. J. S. Rustin Consultant (Medical Oncology), tl. Holden Data Manager, $Roger A'Hern Statistician, tk. D. Bagshawe Professor (Medical Oncology) *St Bartholomew S Hospital, London; #The Royal Marsden Hospital, London; TTrophoblastic Tumour Screening and Treatment Centre, Department of Medical Oncology. Charing Cross Hospital, London; **Mount Vernon Hospital, Northwood, Middlesex Objective To define management options for women presenting with gestational trophoblastic disease (GTD) which had already metastasised to the liver. Design Retrospective analysis of case records between 1958 and Setting A national referral centre for trophoblastic disease. Results The database containing 1676 treated patients was reviewed and 46 patients with hepatic metastases were identified (.7%). The median age was years (range 19-5 years). The antecedent pregnancy to the GTD was normal in 65% (0/46), and the time interval between the antecedent pregnancy and presentation was longer than one year in 50% (/44). Lung metastases were present in 4 patients (9%) and brain deposits in 15 patients (%). Forty-five patients (98%) were high risk by WHO criteria. The five-year overall survival was 7%. The five-year survival of the subgroup of patients having both hepatic and cerebral metastases was 10%. There was no significant survival difference between the different chemotherapy regimens used in the study period (pre CHAMOCA: methotrexate, actinomycin D, cyclophosphamide, doxorubicin, melphalan, hydroxyurea and vincristine; 1979 onwards EMA/CO-EP: etoposide, methotrexate, adriamycin-d/ cyclophosphamide, vincristineetoposide and cis-platinum). Multivariate analysis revealed that a prognostic score > 1 was significant (Hazard ratio 5.4,95% CI ; P = 0.04). Conclusions The outcome for women presenting with hepatic metastases from GTD is poor with an even worse prognosis if cerebral metastases are also present. Alternative therapeutic measures, such as high dose therapy or new drugs, should be explored in these women. INTRODUCTION The modern management of gestational trophoblastic disease (GTD) has resulted in a cure rate in excess of 90%1-. This success has been the result of effective use of the tumour marker human chorionic gonadotrophin (hcg), the inherent sensitivity of trophoblastic tumours to chemotherapy, the referral of patients to centralised specialist units for surveillance and treatment, and the recognition of high risk factors which identify patients who should be treated with aggressive chemotherapy regimens, salvage surgery and occasionally irradiation4. In the low risk group, cure is expected and the quest is for effective chemotherapy with a minimum of long term side effects. Between % and 0% of patients presenting with metastatic GTD have liver inv~lvement~,~ which carries an ominous prognosis with less than 40% s urviving 5 year^^,^.*. ~ Correspondence: Mr R. Crawford, Department of Gynaecology, Second Floor, KGV Building, St Bartholomew's Hospital, West Smithfield, London ECl A 7BE, UK. The FIGO staging system uses an anatomical distribution. Stage I disease is confined to the uterus, Stage I1 is limited to the genital structures, Stage I11 extends to the lungs and in Stage IV, metastases are present at other sites. Therefore liver involvement signifies Stage IV disease using the FIGO criteria9 and generally leads to high risk classification using the modified criteria of the World Health Organization (WH0)'O. The WHO modified scoring system as used at the Charing Cross Hospital allocates points (0, 1, or 6) to features such as age, antecedent pregnancy, time interval between presentation and that pregnancy, the presenting level of hcg, blood group, the number and site of metastases, the size of the largest mass and the use of prior chemotherapy. High risk patients score more > 9 points, while low risk patients score 5 5 points. The optimal treatment for patients with hepatic metastases has not been identified. Multiple agent chemotherapy with or without hepatic irradiation has been proposed. We have reviewed the database at the Trophoblastic Tumour Screening and Treatment Centre at Charing 0 RCOG 1997 British Journal of Obstetrics and Gynaecologv 105

2 106 R. A. F. CRAWFORD ET AL. Table 1. Clinical details of patients presenting with hepatic metastases in GTD. Age (years) (n = 46)? 0 7 (59) < 0 19 (41) Ethnic group (n = 46) Caucasian 41 (89) Other 5 (11) Blood group (n = 4) 0 14 () A 16 (8) B 7 (17) AB 5 (1) Parity (n = 46) Nulliparous (7) 1- (7) () Antecedent pregnancy (n = 46) Term 0 (65) Nonmolar abortion 7 (15) Hydatidiform mole 9 (0) Interval (months) between previous pregnancy and GTD (n = 44) <4 10 () (14) (14) > 1 (50) Status: no evidence n (%) ofdisease P* t 5 0.0: Univariate analysis. thydatidiform mole carries a better prognosis than other groups. :The shorter the time interval between prior pregnancy and GTD the better the prognosis. Cross Hospital in London and analysed the subgroup of patients presenting with liver disease in an attempt to define management options for these women. METHODS A retrospective review of the cases presenting to the Charing Cross Trophoblastic Tumour Screening and Treatment Centre between August 1958 and October 1994 was performed. Using the computerised database, cases with liver involvement were identified. The case notes of these patients were retrieved and the following details obtained: age, ethnic group, blood group, parity, antecedent pregnancy, interval between antecedent pregnancy and presentation with metastatic GTD, the presence of liver and lung metastases, the prognostic score on presentation using WHO criteria, their known disease status and the treatment they had received. Women with tumours of probable nongestational origin were excluded from this series by using molecular genetic techniques, *. Survival curves were calculated using the Kaplan-Meier product limit method and compared using the logrank test. Multivariate analysis was performed using Cox s regression analysis. All patients with GTD in the south of England are registered with Charing Cross and are monitored by serum and urinary PhCG according to our protocol. Since 197 when a nationwide registration scheme was introd~ced ~ referrals have been from the whole of the United Kingdom as well as from overseas. Patients are assessed according to a standard protocol. Following a full physical examination, blood was analysed for full blood count, ABO grouping and biochemistry including liver function tests. PhCG was assayed in both serum and cerebrospinal fluid. Routine radiology was confined to a chest X-ray. Ultrasound examination of the abdomen and pelvis was used to examine the uterus including Doppler blood flow studies of the myometrium, uterine vessels and the liver. Computerised tomography of the brain was performed if the patient had central nervous system symptoms or the cerebrospinal fluid/serum hcg ratio was greater than 1/60. Initially isotope scans were used to image the liver and the brain but this has been replaced by ultrasound and computerised tomography. All these patients were treated with multi-agent chemotherapy with curative intent. Surgery and radiation therapy were used as clinically indicated although liver irradiation was not performed routinely. RESULTS There were 1676 patients reviewed on the database and 46 were identified as having liver metastases. This represented.7% of the patients treated. The median age was years (range 19-5 years). The age, ethnic group, blood group and parity of the patients are shown in Table 1. The type of antecedent pregnancy and the time interval between that event and the onset of GTD are also shown in Table 1 and these features were significant on univariate analysis. All patients reviewed had their metastases defined by scanning, biopsy or at postmortem. Of the 46 with documented liver metastases, 4 (9%) had definite lung metastases and % (15/46) had brain metastases. One woman died of a cerebral haemorrhage, although no central nervous system metastases were documented on presentation. All patients with brain metastases also had lung metastases. All were FIG0 Stage IV by virtue of the liver metastases. Using WHO criteria, 45 patients were high risk (risk score 9) and one case low risk. The range of prognostic scores was 5- with a median score of 0 (Table ). No definite follow up details were available in three cases. Thirty-one women (7%) died (1/4 with follow up and status recorded). The overall five-year 0 RCOG 1997 Br J Obstet Gynnecol 104,

3 ~ ~ GESTATIONAL TROPHOBLASTIC DISEASE WITH LIVER METASTASES 107 Survival rate of women with hepatic metastases (after patients who died within one month of presentation have been excluded: see text) Women with hepatic but no cerebral metastases (n= 1) rl 0 -- metastases (n = 15) -- A I Overall survival rate of all women with hepatic metastases 10 A I I 1 Women with hepatic and cerebral 0 I LO Time since primary diagnosis (years) Fig. 1. Overall survival of women presenting with hepatic metastases in GTD. Table. Prognostic scoring (WHO) and metastatic status in patients with hepatic metastases in GTD. n (Yo) Status: no evidence of disease Pulmonary metastases 4 (9) 11 Cerebral metastases 15 () Prognostic score (WHO) <6 1 () (high risk) 5 (11)? 1 (high risk) 40 (87) 8 survival was 7% (Fig. 1). Nine died within one month of presentation: gastrointestinal haemorrhage (n = 4, including one at laparotomy); disseminated choriocarcinoma (n = ); cerebral infarction (n = 1); respiratory failure (n = 1) and cardiac arrest (n = 1). The five-year survival in the subgroup of patients having both hepatic and cerebral metastases was 10% compared with 4% in the group having hepatic metastases but with no documented cerebral involvement (P = 0.09) (Fig. ). Five of the women who died within one month of presentation had no cerebral metastases. All had received chemotherapy, although those dying before one month were deemed to have died from their initial disease. The regimen of combination chemotherapy has been modified with time. Initially the regimen for high risk patients was CHAMOCA (methotrexate, actinomycin-d, cyclophosphamide, doxorubicin, melphalan, hydroxyurea and vincristine), but in 1979, EMNCO (etoposide, methotrexate, actinomycin-d / cyclophosphamide and vincristine) was introduced with EP (etoposide and cis-platinum) if drug resistance developed. The survival of the women treated with the different regimens is compared in Fig.. There is no significant difference in survival between the two regimens. Only one patient received liver irradiation. Fig.. Survival of women with hepatic and cerebral metastases in GTD versus women with hepatic but no cerebral metastases. I # 1$ CHAMOCA regime before EMAICO-EP 10 after 1979 regime 0. I. :. :. :. :. :. :. :. :. I Time since primary diagnosis (years) Fig.. Survival of women before and after 1979 when the standard high risk chemotherapy regime changed from CHAMOCA to EMA/CO-EP (see text for abbreviations). Mutivariate analysis showed that a prognostic score of 1 was independently significant for cause specific survival (Hazard ratio 5-4,95% CI ; P = 0.04). The presence of brain metastases was not independently significant (Hazard ratio *6,95% CI ; P = 0.06). DISCUSSION This retrospective analysis describes the experience of one specialist unit with respect to the outcome and management of a high risk group of patients with gestational trophoblastic disease. The incidence of.7% of patients presenting with liver metastases to the Charing Cross Unit is at the lower end of the range of published data on hepatic involvement in GTD. As the centre treats all the cases of GTD in its referral area, this is not a result of a referral bias. As the overall survival for all GTD in this unit is similar 0 RCOG 1997 Br J Obstet Gynaecol 104,

4 108 R. A. F. CRAWFORD ET AL. to other ~entresl-~, clinically significant hepatic lesions have been detected appropriately. The presence of liver metastases is a strong indicator of a poor outcome in GTD. In this study two other features relating to poor outcome are noteworthy: the long time interval between the previous pregnancy and the presenting GTD (Table 1) and normal pregnancy before GTD14. Therefore the poor outcome in this study may be related to late presentation in addition to the development of chemoresistant disease. This hypothesis is supported by the widespread nature of the disease, with liver and cerebral metastases, and that 0% of the women (9/46) died before treatment was able to take effect. The poor outlook for women with both hepatic and central nervous system metastases has been previously reported4 although in this study the five-year survival is very much lower (10%). In this study using multivariate analysis the WHO score was the only significant prognostic indicator (hazard ratio of 5.4 for the group with a prognostic score of greater than 1, P = 0-04). The presence of cerebral metastases was not an independent significant predictor of poor outcome using multivariate analysis, although this is likely to have been affected by the heavy weighting cerebral metastases give to the prognostic score and the small numbers in this group. This suggests that the high risk WHO category could be further divided, and this may help in improving management by tailoring appropriate aggressive therapy for the higher risk patient using new agents, such as the taxanes and topoisomerase- 1 inhibitors, and new regimens such as high dose protocols. In other centres it appears to be standard practice to irradiate the liver to 000 cgy over 10 fraction^'^>*^. This is not a tumour-sterilising dose and is intended only to reduce the risk of hepatic haemorrhage during initial treatment. None of the women at the Charing Cross Hospital died from hepatic haemorrhage and it does not, therefore, seem appropriate to give radiation therapy. Lurain supports this view. In the same way, surgery has no role for the initial treatment of hepatic metastases. In our study surgical removal of persistent hepatic abnormalities was not performed. Despite the aggressive use of chemotherapy survival in this study is poor. Multiple agents used in the initial treatment regimen are important. Bakri et al. * reported a small retrospective review showing PEA (cis-platinum, etoposide and actinomycin D) was superior to MAC (methotrexate, actinomycin D and chlorambucil) without hepatic irradiation. In our series, patients up to 1979 were treated with CHAMOCA. As it became apparent that the platinum based compounds had activity against chemo-resistant disease, cis-platinum was used in second-line treatment. Therefore, the patients whose disease was resistant to initial therapy with EMA-CO would be treated with a second regimen including cis-platinumum (EMA-EP)18. The EMA-CO/EP regimen did not appear superior to CHAMOCA in survival (Fig. ), although the myelotoxicity was less. However, the number of patients in each group was small and there was some crossover in treatment schedules between the groups. This study describes the experience of a specialist unit in the management of a group of patients with a very poor prognosis. All staging systems recognise the involvement of the liver with metastatic GTD as a poor prognostic indicator. Further work needs to be done to see whether a more refined scoring system to identify groups of patients at high-risk will help with management. At present we cannot recommend a regimen which is clearly superior for treating these high risk patients. The EMA-CO regimen with platinum and without hepatic irradiation will continue to be used as standard treatment while new chemotherapeutic agents and protocols are evaluated. References Newlands ES, Bagshawe KD, Begent RHJ, Rustin GJS, Holden L, Dent J. Developments in chemotherapy for medium- and high-risk gestational trophoblastic tumours ( ). Br J Obstet Gynaecol 1986; 9: Bagshawe KD, Dent J, Newlands ES, Begent RHJ, Rustin GJS. The role of low-dose methotrexate and folinic acid in gestational trophoblastic tumours. Br JObstet Gynaecoll989; 96: Lurain JR. High-risk metastatic gestational trophoblastic tumorscurrent management. JReprodMed 1994; 9: 17-. Lurain JR, Brewer JI, Torok EE, Halpem B. Gestational trophoblastic disease: treatment results at the Brewer Trophoblastic Disease Center. Obstet Gynecoll98; 60: Wong LC, Choo YC, Ma HK. Hepatic metastases in gestational trophoblastic disease. Obstet Gynecoll986; 67: ll. Soper JT. Identification and management of high-risk gestational trophoblastic disease. Semin Oncoll995; : Bagshawe KD. Risk and prognostic factors in trophoblastic disease. Cancer 1976; 8: Surwitt EA, Hammond CB. Treatment of metastatic trophoblastic disease with poor prognosis. Obstet Gynecol1980; 55: Pettersson F, editor. Annual report on the results of treatment in gynecological cancer. International Journal of Gynecology and Obstetrics fstockholm) 1991 : 6: S World Health Organization Scientific Group on Gestational Trophoblastic Disease. World Health Organ Tech Rep Ser 198; (69). 11 Fisher RA, Newlands ES, Jefieys AJ et al. Gestational and non-gestational trophoblastic tumours distinguished by DNA analysis. Cancer 199; 69: Fisher RA, Soteriou BA, Meredith L, Paradinas FJ, Newlands ES. Previous hydatidiform mole identified as the causative pregnancy of choriocarcinoma following birth of normal twins. Int J Gynecol Cancer 1995; 5: Bagshawe KD, Dent J, Webb J. Hydatidiform mole in England and Wales Lancet 1986; : Tidy JA, Rustin GJS, Newlands ES et al. Presentation and management of choriocarcinoma after nonmolar pregnancy. Br J Obstet Gynaecoll995; 10: RCOG 1997 r J Obstet Gynaecol 104,

5 15 Bamard DE, Woodward KT, Yaney SG, Weed JC, Hammond CB. Hepatic metastases of choriocarcinoma: a report of 15 patients. Gynecol Oncoll986; 5: Hammond CB, Soper JT. Poor prognosis metastatic gestational tro phoblastic neoplasia. Clin Obstet Gynecol 1984; 7: Bakri YN, Subhi J, her M et al. Liver metastases of gestational trophoblastic tumor. Gynecol Oncol199; 48: Newlands ES, Bagshawe KD, Begent RHJ, Rustin GJS, Holden L. GESTATIONAL TROPHOBLASTIC DISEASE WITH LIVER METASTASES 109 Results with the E WCO (etoposide, methotrexate, actinomycin D, cyclophosphamide, vincristine regimen in high risk gestational trophoblastic tumours, Br J Obstet Gynaecol 1991; 98: Received ll December 1995 Accepted August RCOG 1997 Br J Obstet Gynaecol 104,