Treatment of Uveal Melanoma Metastatic to the Liver

Transcription

1 1665 Treatment of Uveal Melanoma Metastatic to the Liver A Review of the M. D. Anderson Cancer Center Experience and Prognostic Factors Agop Y. Bedikian, M.D., Sewa S. Legha, M.D., Giora Mavligit, M.D., Cesar H. Carrasco, M.D., Sunil Khorana, B.S., M.S., Carl Plager, M.D., Nicholas Papadopoulos, M.D., and Robert S. Benjamin, M.D. Bockground. Liver metastasis develops in approximately two-thirds of patients with recurrent uveal melanoma. Despite therapy, the median survival of those with liver metastasis is 5 to 7 months. The recognition of a grave prognosis associated with liver metastasis has led to evaluation of new modalities of therapy, including the use of regional therapies such as intrahepatic arterial chemotherapy and either embolization or chemoembolization of hepatic metastases. In this study, the results of an institutional experience over the past decades are reviewed and prognostic factors that affect survival from the time the liver metastasis is diagnosed are assessed. Methods. In this study of 01 patients with uveal melanoma involving the liver who were treated at M. D. Anderson Cancer Center between 1968 and 1991, the authors retrospectively reviewed the cases and compared the results of systemic therapies, hepatic intra-arterial chemotherapies, and chemoembolization of liver metastases. Cox s multivariate analysis and stepwise logistic regression were then computed to determine significant prognostic variables. Results. The systemic therapies produced a response rate of less than 1%. Chemoembolization was the most effective treatment, inducing responses in 36% of patients. Survival curves were calculated using the life-table method of Kaplan and Meier. Patient- and tumor-related characteristics were examined and their relation to on survival from the time of diagnosis of liver metastasis was determined. Levels of serum alkaline phosphatase, From The University of Texas M. D. Anderson Cancer Center Houston, Texas. The authors thank the Department of Patient Studies for access to patient population database. Address for reprints: Agop Y. Bedikian, M.D., University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Box 77, Houston, TX Received January 6, 1995; revisions received April, 1995, and July 6, 1995; accepted July 6, total bilirubin, and lactic dehydrogenase plus response to treatment showed a strong relation to survival. In contrast, univariate analysis showed that patient age and gender, metastasis free interval, presence of extrahepatic metastasis, and type of therapy for liver metastasis did not influence survival. Multivariate stepwise regression analysis identified serum alkaline phosphatase and metastasis free interval as the main independent prognostic factors for survival after liver metastasis diagnosis. Conclusions. Of the three modalities of therapy used for choroidal melanoma metastatic to the liver, only chemoembolization using cisplatin-based regimens produced a meaningful response rate. Information from this analysis can be used to predict the outcome of patients with uveal melanoma metastatic to the liver. Patients with metastatic ocular melanoma confined to the liver should be treated with chemoembolization and should not be included in chemotherapy trials designed for cutaneous melanoma. Cancer 1995; 76: Key words: eye, neoplasm, metastasis, melanoma. Uveal melanoma is the most common primary ocular malignancy in white people, accounting for 70% of all eye malignancies. It arises from the melanocytes in the uveal stroma. In the United States, the annual age-adjusted incidence of ocular melanoma is approximately 13% that of cutaneous melanomas. Although the incidence of cutaneous melanoma has been increasing markedly, uveal melanoma has not shown a similar increase. The optimum therapy for primary and metastatic uveal melanoma remains controversial. For many years, enucleation of the affected eye was considered appropriate, More recently, however, the efficacy of this therapy in preventing systemic metastasis has been questioned. Conceivably, manipulating the affected eye

2 1666 CANCER November, 995, Volume 76, No. 9 during enucleation may contribute to the metastatic pro~ess.~ During the past decades, investigators have explored new modalities of therapy as alternatives to enucleation of the eye. These include local surgical resection, laser photocoagulation, and, in selected patients with a small primary, radiotherapy. Randomized studies currently in progress may elucidate the relative merit of these modalities for treating primary melanoma. Therapy for metastatic uveal melanoma has been the subject of recent investigations, especially after physicians realized that the chemotherapy used for cutaneous melanoma was less effective against metastatic uveal melan~ma.~-~ Despite therapy, the median survival of patients who first develop liver metastasis is 5-7 months, whereas that of patients who first develop metastasis at other sites is about 18 month^.^-^ Liver metastasis develops in approximately two-thirds of the patients with recurrent uveal melanoma. This realization has led to the evaluation of new therapeutic modalities for patients with liver metastases, including such regional therapies as intrahepatic arterial chemotherapy, embolization of the hepatic metastasis and, recently, chemoembolization.s~9 Although responses to the regional therapies have been higher than those achieved with systemic chemotherapy, the effect on patient survival remains uncertain. In this study, we have reviewed the results of our experience with these treatments over the past decades and assessed the prognostic factors affecting survival from the time liver metastasis is diagnosed. Materials and Methods We reviewed the records of The University of Texas M. D. Anderson Cancer Center registry for the years Patients with a histologically confirmed diagnosis of uveal melanoma whose first extraocular metastasis was hepatic were selected for the study. Their case reports were used to determine the prognostic factors that significantly influenced survival after magnosis of a first metastatic tumor in the liver. Patients who received intravenous (IV) chemotherapy but at no time received intra-arterial (IA) chemotherapy or chemoembolization of the liver metastases made up the IV-therapy group. Patients who had undergone embolization of their liver metastases without hepatic IA chemotherapy made up the embolization-therapy group. Those patients who received IA hepatic chemotherapy but not chemoembolization of the liver metastases made up the HIAtherapy group. The patients who received HIA chemotherapy plus embolization of the liver metastases made up the chemoembolization-therapy group. The patients in the HIA-, chemoembolization-, and embolization- therapy groups may have also received oral or IV chemotherapy for their metastatic disease. The hospital records of these patients were carefully reviewed for clinical data and course of diseases. The survival curves were corrected to eliminate the effect of deaths from causes other than malignant melanoma. Statistical Methods We calculated survival curves using the life-table method of Kaplan and Meier.'' A step-wise logistic regression equation relating length of survival to tumorand patient-related characteristics was computed. A stepwise regression procedure was used in which Cox's life-table regression model was fitted to the survival data. The patient- and tumor-related characteristics of 01 patients were included in the regression equation to determine factors influencing survival after the diagnosis of liver metastasis. The differences in survival between the IV-therapy, HIA-therapy, and chemoembolization-therapy groups were evaluated using Gehan's modification of the Wilcoxon test." A patient was considered to have developed liver metastasis on confirmation by liver biopsy, or when hepatic angiogram, technetium liver scan, liver ultrasound, or computerized tomographic scan findings were diagnostic of liver metastasis, or, in the absence of these confirmatory tests, there was significant hepatomegaly (liver edge extended greater than 3 cm below the right costal margin or xiphoid process) in the absence of congestive heart failure and hepatitis. A patient was considered to have achieved a partial remission if there was a decrease in the sum of the products of two perpendicular greatest dimensions of metastatic nodules in serial scans of more than 50% of the pretreatment value without an increase in the size of any mass or the appearance of new lesions; or a 30% decrease in the sum of measurements of the extent of liver size below the costal margin, at the midclavicular lines, and at the xiphoid process, measured in quiet expiration. Complete remission required the disappearance of all signs and symptoms of disease for at least weeks. A patient was considered to have an adequate trial of chemotherapy if he received at least 1 chemotherapy cycle and survived over weeks thereafter. Results Description of Patients The 01 patients with uveal melanoma who had hepatic metastases at the time of diagnosis of the first extraocular metastases were chosen for the review. Their

3 Uveal Melanoma Metastatic to the Liver/Bedikian et al Table 1. Type of Therapy No. of patients Table. Systemic Chemotherapy Trials First-line Second-line or Therapy therapy subsequent Total IV chemotherapy Chemoembolization 0 6 IA chemotherapy 1 38 Embolization 1 1 Symptom control IV: intravenous; IA: intra-arterial median age was 59 years (range, 5-81 years). There were more men than women (male-to-female ratio, 1.: 1). The primary tumor was located in the right eye more often than the left (119 vs. 8). Information about the primary tumor was available for 75 patients. Twelve patients had epithelioid tumors, 7 had spindle-b tumors, and the tumors of the remaining 36 were of mixed histologic types. Two patients had hepatic metastases at the time the primary tumors were diagnosed. Treatment for the primary tumor was as follows: for 155 patients, enucleation of the eye; and for, exenteration. Four patients treated with enucleation of the eye had received preoperative radiotherapy. Thirty-six patients received radiotherapy only, and 6 received no therapy for the primary tumor. In 6 patients treated with radiation, disease relapsed locally, and the eye was surgically removed subsequently. The metastatic workup at the time of diagnosis of extraocular metastases showed the following: 119 patients had tumor metastases in the liver only, and 8 patients had tumor metastases in the liver plus other sites, including lung in 0 patients, bone in 9, skin/subcutaneous tissues in, abdominal lymph nodes in 3, breast in 3, pancreas in, stomach in, spleen in, brain in, and parotid gland in. Treatment The various therapies received by the patients are summarized in Table 1. Thirty-three patients received treatment for control of tumor-related symptoms only. The therapeutic effects of systemic chemotherapy are summarized in Table. Ninety-nine of 13 chemotherapy regimens were administered to patients with no prior chemotherapy. Only one had a partial response. Fortyfour patients who had received prior arterial intrahepatic chemotherapy or chemoembolization of the liver received second-line IV chemotherapy regimens with no responses. Two patients had embolization of the liver only and both had tumor progression. Table 3 lists the results of the hepatic intra-arterial treatment regimens. In these treatments, cisplatin and occasionally Therauv DTIC VELB + DDP k IFN? tamoxifen DTIC ACT-D k C. Parvum DTlC ACT-D + VCR i hydrea DTIC BCNU + VCR f bleomycin DTIC BCG or C. PARVUM DTlC BCNU + BCG DTIC MeCCNU + BCG? T cells DTIC BCNU DDP + tamoxifen DTIC + BCNU + Hydrea i BCG + C. Total PR MR SV PD Parvum 5 3 DTIC + CTX or AMSA or vindesine or procarbazine DTIC + IL- + LAK cells 1 - _ 1 Other multidrug regimens 10-8 Interferon k IL Other single agents* All treatments PR: partial response; MR: minor response; SD: stable disease; I-?: progressive disease; DTIC: dacarbazine; VLLB: vinblastine; DDP: cisplatin; IF : interferon; ACT-D: dactinomycin; C. Parvum: Corynebacterium Parurn; VCR: vincristine; Hydrea: hydroxyurea; BCNU: carmustine; BCG: Bacille Calmette-Guerin; MeCCNU: methyl CCNU; CTX cyclophospharnide; AMSA amsacrine; 1L-: interleukin ; LAK: lymphokine-activated killer. * Including DDP (5 patients) AMSA (), DTlC (3). nitrosoureas (3), Vindesine (). other uhase 1-11 agents (18). dacarbazine and vinblastine were administered intraarterially to the liver. Bleomycin, amsacrine, and vindesine were infused intra-arterially to the liver only when used as single agents. Twenty-four of the 38 patients Table 3. Intra-arterial Therapy Trials Theravv Total PR MR SV PV DDP DDP + DTIC DDP + DTIC + VELB f ACT-D DDP + DTIC + ACT-D t hydrea DTIC i BCNU or AMSA DTIC + MeCCNU + C. Parvum DTIC + ACT-D + C. Parviim Other multidrug regimen Interleukin- Interferon Other single agents All treatments _ - _ _ 10 PR: partial response; MR: minor response; SD: stable disease; PD: progressive disease; DDP: cisplatin; DTIC: dacarbazine; VELB: vinblastine; ACT-D: dactinomycin; Hydrea: hydroxyurea; BCNU: carmustine; AMSA: amsacrine; MeC- CNU: methyl CCNU; C. Parvum: Corynebacterium Parvum; IFN: interferon; IL-: interleukin.

4 1668 CANCER November I, 1995, Volume 76, No. 9 Table. First-Line Chemoembolization TheraDv Total CR PR MR SD PD DDP DDP + VELB DDP + DTIC + VCR DDP + DTIC DDP + ACT-D - - All treatments PR: partial response; M R minor response; SD: stable disease; PD: progressive disease; DDP: cisplatin; VELB: vinblastine; DTIC: dacarbazine; VCR vincristine: ACT-D: dactinomvdn. received intra-arterial intrahepatic therapy as the first line treatment. Two showed partial responses, both during first-line therapy. Forty-four patients had chemoembolization of the liver metastases as the first treatment, and 0 additional patients had chemoembolization after failing on other treatment programs. Cisplatin was the anticancer agent most often administered intraarterially into the liver and was combined with polyvinyl sponge (Ivalon). About one-third of the patients, in addition, received dacarbazine or vinblastine intravenously or, very rarely, intra-arterially into the liver. The median number of chemoembolization procedures was three (range, 1 to 5). The response rates were 16/ (36%) for first-line and 5/0 (5%) for subsequent treatments. All responses except one were partial (Tables and 5). When the survival of patients who received chemoembolization is compared with that of 77 patients who received only IV chemotherapy, the survival difference is not statistically significant (median, 6 vs. 5 months; P = 0.). However, the chemoembolization responders survived significantly longer than the chemoembolization nonresponders (median, 1.5 vs. 5 months; P = 0.003) or the patients who received systemic chemotherapy (median, 1.5 vs. 5 months; P = 0.003). The median duration of the partial responses was 6 months for first-line therapies and shorter for subsequent therapies. The duration of the complete response was 19 months. Prognostic Factors The median survival from diagnosis of liver metastasis was 7 months (range, 1-59 months). At 1,, and 36 months, the corresponding numbers of living patients were 5 (.5%), 10 (5%), and (%). A single patient lived 59 months after histologically confirmed diagnosis of recurrent melanoma in the liver. This patient had chemoembolization of the liver metastasis with cisplatin 3 times and achieved a complete remission that lasted for 19 months. His recurrent tumor in the liver was monitored without therapy until he became symptomatic, 8 months before his death. Then he was given a course of intrasplenic arterial interleukin-, after which he failed to return for evaluation. He died of progressive disease 8 months later. The influence of tumor- and patient-related factors on survival after diagnosis of liver metastasis was evaluated first by univariate analysis and then using the multivariate stepwise regression method to determine the most important independent factors predicting survival. For the latter test, the characteristics analyzed were ranked in decreasing order of relative importance for predicting survival time after diagnosis of liver metastasis. The patients age at diagnosis of liver metastasis, sex, interval from diagnosis of uveal melanoma to liver metastasis (metastasis free interval), presence or absence of extrahepatic metastases, type of therapy for recurrent cancer, response to therapy, serum levels of total bilirubin, lactic dehydrogenase (LDH), and alkaline phosphatase were evaluated as shown in Table 6. Univariate analysis showed that response to therapy and serum levels of alkaline phosphatase, lactate dehydrogenase, and total bilirubin were significant factors influencing survival after the diagnosis of liver metastasis. The patients who had a partial or complete response lived significantly longer than the nonresponders (median, 10.5 vs. 6 months, P = ). Patients who had normal levels of total serum bilirubin, alkaline phosphatase, and LDH at the time of detection of liver metastasis survived longer than those whose levels of these enzymes were elevated above the normal range (median, 6.5 vs. months; P = ; 10 vs. months, P = ; 9 vs. 5 months, P = 0.0; respectively). Patients age, sex, the metastasis free interval, presence of extrahepatic metastases, and type of therapy for liver recurrence were not prognostically significant individually. Table 5. Second-Line and Subsequent Chemoembolization Therapy Total CR PR MR SD DD DDP DDP VELB 3-1 DDP DTIC DDP + ACT-D All treatments PR: partial response; MR: minor response; SD: stable disease; I D: progressive disease; DDP: cisplatin; VELB: vinblastine; DTIC: dacarbazine; ACT-D: dactinomvcin.

5 Uveal Melanoma Metastatic to the Liver/Bedikian et al Table 6. Factors Evaluated for Their Influence on Survival Duration Factors Definition Sienificance Age Sex Metastasis free interval Extrahepatic metastasis Treatment to therapy Serum alkaline phosphatase level Serum lactic dehydrogenase level Serum total bilirubin level : nut simifirant: IA: intra-arterial. 1 = >SO years old = SO years old or less 1 = female = male 1 = months or less = 5-96 months 3 = 96 months or more 0 = absent 1 = present Systemic therapy 1A therapy chemoembolization 0 = absent 1 = present 1 = within normal range = elevated 1 = within normal range = elevated 1 = within normal level = elevated P = P = P = 0.0 P = A multivariate stepwise regression procedure was used, in which Cox's life-table regression model was fitted to the survival data. Only the metastasis free interval and the serum alkaline phosphatase level appeared to have significant independent influence on survival (P values 0.09 and 0.005, respectively); total serum 1 bilirubin level was of borderline significance (P = 0.086). Discussion Little progress has been made in the treating of advanced cutaneous melanoma. Dacarbazine remains the agent most frequently used to treat metastatic disease. Multidrug combination chemotherapy regimens including dacarbazine have resulted in 35% to 5% response rates but produce greater toxicities and have little impact on patient s~rvivai.'~~'~ Uveal melanoma metastatic to the liver treated on protocols designed for cutaneous malignant melanoma has proved to be particularly resistant to systemic chemotherapy, which has produced response rates less than 1% and has had an insignificant effect on overall s~rvival.~-~ Consequently, investigators have sought new approaches for managing this disease. Administering anticancer agents intraarterially direct to the liver, with and without blocking the blood supply of the liver metastases, has been tried several times, reportedly producing better response In view of inherent differences in accuracy of response evaluation in the liver over the past 0 years, in this study we decided to evaluate the effect of these various modalities of therapy on patient survival from the time the first extraocular liver metastasis was diagnosed. Although the medical literature contains a wealth of statistics regarding prognostic factors affecting the long term survival after surgical removal of uveal melanoma with intent to cure, information about factors that influence survival from the time of diagnosing a liver metastasis is very limited. Patients' age and sex have been suggested as independent influences on survival duration after tumor recurrence. Patients younger than 50 years and women are reported to live longer than those older than 50 years and men.6 Our study, employing univariate analysis, indicates that these factors have no prognostic value after the discovery of liver metastasis, confirming the observation of another recent study. Such tumorrelated factors as length of metastasis free interval and presence of extrahepatic disease were not individually significant, but factors related to liver dysfunction, such as the serum levels of alkaline phosphatase, lactate dehydrogenase, and total bilirubin, were inmvidually of prognostic significance. The relative importance of these prognostic factors as evaluated by Cox's multivariate regression analysis indicates that only the serum alkaline phosphatase level and metastasis free interval before diagnosis of liver metastasis were significant, and the total serum bilirubin level was of borderline prognostic significance. Contrary to the finding of a recent study, our study

6 1670 CANCER November 1,1995, Volume 76, No. 9 showed that the metastasis free survival interval after therapy for the primary tumor significantly influences survival duration after liver metastasis, indicating the influence of the inherent biology of the tumor on these patients, Patients who had delayed tumor recurrence after therapy of the primary tumor survived significantly longer. Although the choice of therapeutic modality was not significant in view of the small number of responses, response to therapy was a significant factor influencing survival. Of the three modalities of therapy in treating choroidal melanoma metastatic to the liver, only chemoembolization using cisplatin-based regimens produced a meaningful response rate. The result of the present retrospective study indicates that, for patients with metastatic uveal melanoma involving the liver, response rates after chemoembolization are significantly higher than those after any other therapeutic modality. The result of a retrospective study recently showed that the response rate to a dacarbazine, cisplatin, carmustine, and tamoxifen regimen was significantly lower in patients with melanoma metastatic to the liver of uveal origin than of cutaneous origin (response rate, 1/15 vs. 19/38; P < 0.05).16 Another regimen, including bleomycin, vincristine, lomustine, dacarbazine, and interferon induced three partial responses in 1 treated patients with uveal melanoma metastatic to the liver, suggesting possible activity.17 A retrospective review of South West Oncology Group studies for metastatic melanoma showed responses in 56 patients with uveal melanoma metastatic to the liver treated with dacarbazine and/or cisplatin." A higher response rate was observed in a small number of patients with extrahepatic metastases. Multivariate analysis for prognostic factors showed that the chemotherapy was not a significant prognostic factor for survival.i8 Although prolonged survival of a few patients who did not respond to therapy has been de~cribed,~,'~ achieving complete remission lasting 19 months did contribute to the 59-month survival of one of our patients. The degree of response may positively influence the survival duration of the responders. In this study, however, regression analysis failed to show any prognostic influence related to type of therapy, most probably because of the small number of patients treated initially with chemoembolization. At this time, we recommend that patients with metastatic ocular melanoma confined to the liver be treated with chemoembolization rather than be included in chemotherapy trials designed for those with cutaneous melanoma. Further studies, including chemoemboliza- tion with cisplatin and other agents, should be pursued to improve the outcome of therapy for ocular melanoma metastatic to the liver. References Cutler SJ, Young JL, Connelly RR. Third national cancer survey: incidence data. Bethesda, MD: National Cancer Institute Monograph 1975; 1: 1-5. Shields JA. Management of uveal melanoma, a continuing dilemma. Cancer 1993; 7: McLean IW, Foster WD, Zimmerman LE. Uveal melanoma: location, cell type and enucleation as risk factors in metastasis. Hum Pathol 1981;13:13-3. Kath R, Hayungs J, Bornfeld N, Sauenvein W, Hoffken K, Seeber S. Prognosis and treatment of disseminated uveal melanoma. Cancer 1993; 7:19-3. Bedikian AY, Kantarjian H, Young SE, Bodey GP. Prognosis in metastatic choroidal melanoma. South MedJ 1981; 7:57-7. Rajpal S, Moore R, Karakousis CP. Survival in metastatic ocular melanoma. Cancer 1983; 5:33-6. Nathan FE, Sato T, Hart E, Berd D, Mastrangelo MJ. to combination chemotherapy of liver metastases from choroidal melanoma compared with cutaneous melanoma [abstract]. Proc ASCO 199; 13:39. Patt YZ, Calvo DB, Benjamin RS. Treatment of regionally confined metastatic tumors with arterial infusion of cisplatin. In: Prestayko AW, Crooke ST, Carter SK, editors. Cisplatin: current status and new developments. New York: Academic Press, 1980: Mavligit GM, Charnsangavej C, Carrasco CH, Patt YZ, Benjamin RS, Wallace S. Regression of ocular melanoma metastatic to the liver after hepatic chemoembolization with cisplatin and polyvinyl sponge. JAM 1988; 60:97-6. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. ]Am Stat Assoc 1958;53: Cox DR. Regression models and life tables. Journal of the Royal Statistical Society 197; 3: Gehan EA. A generalized Wilcoxon test for comparing arbitrarily singly censored samples. Biometrika 1965;5:03-. Taylor CW, Hersh EM. Chemotherapy of advanced melanoma. In: Leong SPL, editor. Malignant melanoma: advances in treatment. Austin, Tx: R.G. Landes, Buzaid AC, Murren J. Chemotherapy for advanced malignant melanoma. Int ]C[in Lab Res 199;1:05-9. Legha SS. Current therapy for malignant melanoma. Sernin On- C O 1989; ~ 16~3-. Nathan FE, Sat0 T, Hart E, Berd D, and Mastrangelo MJ. to combination chemotherapy of liver metastases from choroidal melanoma compared with cutaneous melanoma [abstract]. Proc ASCO 199; 13:39. Nathan FE, Sat0 T, Berd D, and Mastrangelo MJ. BOLD + interferon; effective systemic therapy for metastatic uveal melanoma [abstract]. Proc ASCO 1995; 1:10. Katato K, Liu PY, Sondak V, and Flaherty LE. Survival and response to treatment in patients (pts) with metastatic melanoma from intraocular primaries (MMIP) on SWOG studies [abstract]. Proc ASCO 1995; 1:10. Minor DR, Cohen RJ, West W, Oldham R. 5-year survival of an ocular melanoma patient with hepatic metastases treated with interleukin- (IL-), LAK, & interferon [abstract]. Proc ASCO 199;13:00.