1599 Front-Line Chemotherapy with Cisplatin and Etoposide for Patients with Brain Metastases from Breast Carcinoma, Nonsmall Cell Lung Carcinoma, or Malignant Melanoma A Prospective Study Vittorio Franciosi, M.D. Giorgio Cocconi, M.D. Maria Michiara, M.D. Francesco Di Costanzo, Vinicio Fosser, M.D. Maurizio Tonato, M.D. Paolo Carlini, M.D. Corrado Boni, M.D. Sofia Di Sarra, Ph.D. M.D. Medical Oncology Units of Parma, Terni, Vicenza, Perugia, Roma, and Reggio Emilia, Italy. This study was conducted by the Italian Oncology Group for Clinical Research (GOIRC), Parma, Italy (G. Cocconi, Chairman). Supported by Associazione Italiana per la Ricerca sul Cancro (AIRC) and by the Finalized Project of CNR ACRO. The authors thank Andrea Manni for his critical review of the article and Anna Quadretti for her secretarial assistance. Address for reprints: Vittorio Franciosi, M.D., Medical Oncology Division, Azienda Ospedaliera di Parma, Via Gramsci, 14, 43100 Parma, Italy. Received May 4, 1998; revision received October 1, 1998; accepted November 16, 1998. © 1999 American Cancer Society BACKGROUND. The conventional treatment of brain metastases not amenable to surgery is most often radiotherapy. Until now, pharmacologic issues related to the blood brain barrier (BBB) prevented a wide evaluation of chemotherapy. The authors previously reported that the combination of cisplatin (P) and etoposide (E) had strikingly high activity in patients with brain metastases from breast carcinoma. The purpose of this study was to assess, in a larger prospective study, the front-line activity of that combination against brain metastases from breast carcinoma (BC), nonsmall cell lung carcinoma (NSCLC), and malignant melanoma (MM) in patients previously untreated with radiotherapy. METHODS. From December 1986 to July 1993, 116 patients received P 100 mg/m2 on Day 1 and E 100 mg/m2 on Days 1, 3, and 5 or on Days 4, 6, and 8 every 3 weeks. An insignificant change in the E schedule using the same dose on a random basis assured the prospective enrollment and the registration of all cases. Six patients were not eligible and three patients were excluded from the analysis because they were lost to follow-up shortly after the date of registration. One-hundred seven patients were considered for analysis. The distribution according to the primary tumor site was BC in 56 patients (52%), NSCLC in 43 (40%), and MM in 8 (8%). The first evaluation of response was performed after two cycles. In cases of no disease progression, chemotherapy was continued to a maximum of six cycles. RESULTS. Among the 56 patients with BC, 7 achieved complete response (CR) (13%), 14 achieved partial response (PR), 12 had no change (NC), 15 had progressive disease (PD), and 8 had insufficient treatment or response was not assessed. The CR plus rate was 38%. Among the 43 patients with NSCLC, 3 achieved CR (7%), 10 achieved PR, 15 had SD, 7 had PD, and 8 had insufficient treatment or response was not assessed. The CR plus PR rate was 30%. None of the eight patients with MM achieved an objective response. The median survival was 31 weeks for patients with BC (range, 0–287), 32 for patients with NSCLC (0–3921), and 17 for patients with MM (2–48). CONCLUSIONS. The combination of P and E is effective for patients with brain metastases from BC and NSCLC. In this study, the response rate was of the same order as that reported for disseminated disease without central nervous system involvement. The survival figures compare favorably with some others reported in the literature for patients given radiotherapy. A randomized study is warranted to compare this chemotherapy followed by radiotherapy with radiotherapy alone for patients with brain metastases from BC or NSCLC not amenable to surgery or radiosurgery. Cancer 1999;85:1599 – 605. © 1999 American Cancer Society. KEYWORDS: brain metastases, breast carcinoma, nonsmall cell lung carcinoma, melanoma, chemotherapy. 1600 CANCER April 1, 1999 / Volume 85 / Number 7 B rain metastases (BM) account for 5–15% in the clinical series and 14% in the autopsy series of intracranial tumors. About 20 – 40% of patients with solid tumors present with BM, and two-thirds of them become symptomatic during their lifetimes.1 BM are a major clinical problem for patients with breast carcinoma (BC), occurring in 10 –20% of these patients,2 and they are associated with a poor prognosis.3 Nonsmall cell lung carcinoma (NSCLC) is complicated by BM in 16 – 65% of the patients.4 Malignant melanoma (MM) metastasizes to the brain in 12–20% of the cases in the clinical series and in 36 –54% in the autopsy series, with a very short average survival.5 Patients with inoperable BM from solid tumors have been traditionally treated with radiation combined with corticosteroids. A median survival of 3–6 months and a 1-year survival rate of 15% were reported.6 For a long time, the idea existed that systemic chemotherapy could not be effective against central nervous system tumors because the blood brain barrier (BBB) would preclude the non-lipid-soluble antineoplastic drugs to reach the central nervous system. However, it is currently accepted that the integrity of the BBB is impaired or absent in large areas of the tumor vasculature.7 In the early 1980s, a few reports suggested that chemotherapy was effective in the treatment of BM from solid tumors.8 –12 A Phase II study conducted by our group demonstrated strikingly good activity of the cisplatin (P) and etoposide (E) combination in BM from BC (CR, 23%; CR and PR, 55%).13 Indeed, the PE combination, which was initially designed to treat small cell and nonsmall cell lung carcinoma, was demonstrated by our group to be very active even against systemic nonbrain metastases from BC.14 P is active against MM, and its combination with E was demonstrated to be synergistic in experimental studies.15 For these reasons, we decided to conduct a large, multicentric, prospective study in which this combination would be given as front-line treatment to consecutive patients with BC, NSCLC, and MM. PATIENTS AND METHODS Patient Selection From December 1986, consecutive patients with recently diagnosed BM from BC, NSCLC, or MM not amenable to surgery and not pretreated with radiation therapy were enrolled in this prospective trial. They were from nine institutions affiliated with the GOIRC group. Patients considered not amenable to surgery were those with extracranial visceral disease, more than two brain metastases, or metastases in inoperable cerebral sites, or those who refused surgery. The eligibility criteria were as follows: diagnosis of brain metastases measurable or evaluable by computed tomography (CT) scan, normal bone marrow and renal function, no additional primary tumors, lack of severe cardiac disease, no prior treatment with PE, no prior treatment with any chemotherapy after the diagnosis of BM, and no prior radiation therapy to the brain. Prior to therapy, all patients underwent complete staging, which included clinical examination, chest Xray, bone isotope scan and/or X-ray, liver ultrasound or CT scan, complete blood count, and routine biochemistry. CT scan of the brain with contrast enhancement was made prior to chemotherapy, after the second course, and at every other course thereafter. Oral informed consent was obtained from each patient. Treatment The chemotherapy consisted of P (100 mg/m2 i.v. on Day 1, with hydration and mannitol forced diuresis) and E (100 mg/m2 i.v. on Days 1, 3, and 5 in Arm A and on Days 4, 6, and 8 in Arm B) every 3 weeks. Treatment was postponed by 1 week if a patient’s white blood cell (WBC) count was ,4000/mm3 and/or the platelet (PLT) count was ,100,000/mm3. Subsequently, the dose was reduced by 25% if the WBC was 3900–3000 or the PLT was 99,000–75,000 and by 50% if the WBC was 2900–2000 or the PLT was 74,000 –50,000; if the WBC was ,2000 or the PLT was ,50,000, treatment was postponed by 1 more week. Nephrotoxicity was monitored by repeating creatinine clearance on Day 18 of each course. If clearance creatinine decreased .25%, the treatment was postponed until recovery to normal levels. If hematologic or/and renal toxicities did not resolve after 4 weeks, the treatment was definitively stopped. Eighty-one of 107 patients (76%) initially received dexamethasone at a variable dose and continued it during the initial chemotherapy. The protocol did not define the standard dose and the timing of desamethasone treatment; therefore, they changed on the basis of clinical decisions, according to the extent of the edema and the severity and duration of the symptoms over time. In the absence of disease progression, death, or serious toxicity, chemotherapy was continued for six cycles and thereafter no systemic treatment was given until disease progression in the brain or elsewhere. At the end of the protocol chemotherapy, sequential radiation therapy was allowed. Response Evaluation During the protocol chemotherapy, patients underwent clinical examination before each cycle of chemotherapy, or earlier if needed. Chemotherapy was stopped if any serious toxicity occurred. The response of the tumor and the extent of edema were evaluated by CT scan using contrast enhancement every two cycles or at the time of clinical progression. Based on the CT scan findings, the Chemotherapy in Brain Metastases from Solid Tumors/Franciosi et al. TABLE 1 Demography of the 116 Randomized Patients 1601 TABLE 2 Characteristics of the 56 Patients with Brain Metastases from Breast Carcinoma Patient category No. of patients Total registered Not eligible Creatinine over 1.5 Absence of brain metastases Double primary, breast and thyroid gland Clinical diagnosis of brain metastases without histologic confirmation of the primary Colon carcinoma Small cell lung carcinoma No information available after randomization Evaluated for responsea Evaluated for survival 116 6 1 1 1 1 1 1 3 107 107 a Patients who had insufficient treatment or did not have their objective responses evaluated were considered nonresponders and included in the denominator of response evaluation. response in the brain was classified according to World Health Organization criteria16 and defined as follows: complete response (CR): complete disappearence of tumor and edema; partial response (PR): more than 50% reduction in the product of the largest perpendicular diameters on the CT slices with the largest metastatic lesions (regression of the edema was not considered, per se, to be a sign of tumor response); no change (NC): less than 50% reduction or less than 25% progression; progressive disease (PD): increase of more than 25% in the sites of the metastatic lesions or the occurrence of new lesions. Assessment of response at other metastatic sites, if any, was performed at the same time as BM evaluation. Statistical Analysis The aim of randomization was not to compare two slightly different schedules of chemotherapy (the change was clinically insignificant), but to force the investigators to report to the centralized randomization center all the patients being seen on an intentto-treat basis before the treatment started. The aim of this procedure (which is utilized by our group for most Phase II studies and is called “technical randomization”) is only to eliminate the possibility of selection bias in reporting the results of the most favorable patients, or to eliminate the reporting of “nonevaluable” cases. Due to the design of the study, the sample size was not based on any expected statistical difference between the two treatment arms. It was initially estimated that 100 evaluable patients had enrolled in the study. As said, the difference between the two treatment schedules (for Arm A and Arm B) was not expected to be clinically meaningful, and for this reason the results of treatment were assessed irrespectively of the chemotherapy schedule. Characteristic No. of patients % Total no. Median age, yrs (range) Menopausal status PrePostExtracranial disease Absent Present Type of brain metastasis Solitary Two More than two Prior systemic treatment None Chemotherapy Endocrine therapy Both Performance status (ECOG) 0 1 2 3 4 Neurologic function 1 2 3 4 Unknown 56 56 (41–75) 100% 8 48 14% 86% 11 45 20% 80% 20 14 22 36% 25% 39% 20 15 8 13 36% 27% 14% 23% 9 22 17 4 4 16% 39% 31% 7% 7% 21 15 15 1 4 37% 27% 27% 2% 7% ECOG: Eastern Cooperative Oncology Group. Time to progression (TTP), time to failure (TTF), duration of response (DR), and overall survival (OS) were measured from the date of randomization to one of the following dates: 1) the date of progression or death due to disease (TTP); 2) the date of progression, death, or suspension of treatment for any reason (protocol violation, toxicity, refusal, loss to follow-up, etc.) (TTF); 3) the date of progression or death due to disease but only for patients achieving CR or PR (DR); or 4) the date of death from any cause (OS). The minimum number of cycles requested for evaluation of response was two, unless clinical progression occurred was evident at least 1 month after the beginning of chemotherapy. RESULTS Patient Characteristics From December 1986 to July 1993, 116 patients were registered from the medical oncology units of Parma, Terni, Vicenza, Perugia, Roma, Reggio Emilia, Piacenza, and Lucca. Six patients were not eligible and three pa- 1602 CANCER April 1, 1999 / Volume 85 / Number 7 TABLE 3 Characteristics of 43 Patients with Brain Metastases from NSCLC TABLE 4 Characteristics of the 8 Patients with Brain Metastases from Malignant Melanoma Characteristic No. of patients % Total no. Median age, yrs (range) Gender Male Female Histologic subtype Adenocarcinoma Squamous cell Large cell Brochioloalveolar Extracranial disease Absent Present Type of brain metastases Solitary Two More than two Performance status (ECOG) 0 1 2 3 Neurologic function 1 2 3 4 Unknown 43 59 (28–75) 100% 32 11 74% 26% 21 16 5 1 49% 37% 12% 2% 7 36 16% 84% 19 7 17 44% 16% 40% 11 13 15 4 26% 30% 35% 9% 21 11 4 4 6 49% 26% 9% 9% 14% NSCLC: nonsmall cell lung carcinoma; ECOG: Eastern Cooperative Oncology Group. tients were excluded from the analysis because they were lost to follow-up shortly after registration. The demographic characteristics of all patients are reported in Table 1. Patients who received insufficient treatment or did not have their objective responses evaluated were considered nonresponders and included in the denominator of response evaluation. The characteristics of patients with BM from BC, NSCLC, and MM are reported in Tables 2, 3, and 4, respectively. Characteristic No. of patients Total no. Median age, yrs (range) Gender Male Female Extracranial disease Absent Present Type of brain metastasis Solitary Two More than two Prior systemic treatment None Chemotherapy Immunotherapy Performance status (ECOG) 0 1 2 3 Neurologic function 1 2 Unknown 8 51 (26–66) 2 6 1 7 3 2 3 3 4 1 3 2 2 1 3 4 1 ECOG: Eastern Cooperative Oncology Group. TABLE 5 Types of Response of Brain Metastases in the 107 Patients Considered for Analysis Type of response Breast carcinoma Lung carcinoma Melanoma Total no. No responsea Progressive disease No change Complete response Partial response Complete plus partial response 56 8 15 12 7 (13%) 14 21 (38%) 43 8 7 15 3 (7%) 10 13 (30%) 8 2 5 1 — — — a These patients received insufficient treatment or did not have their objective responses evaluated. Response Rates The types of response observed in patients with BC, NSCLC, and MM are reported in Table 5. Among the 56 patients with BC, 7 (13%) achieved CR and a total of 21 (38%) achieved CR or PR. Eight patients who were not evaluable because their treatment was insufficient or their responses were not evaluated were considered to have experienced treatment failure and included in this analysis. Patients received a median of 3 cycles, ranging from 1 to 8. Among the 43 patients with NSCLC, 3 (7%) achieved CR and a total of 13 (30%) achieved CR or PR. Eight patients who were not evaluable because their treatment was insufficient or their responses were not evaluated were considered to have experienced treatment failure and included in this analysis. Patients received a median of 3 cycles, ranging from 1 to 8. Among the 8 patients with MM, none achieved an objective response. Two patients received insufficient Chemotherapy in Brain Metastases from Solid Tumors/Franciosi et al. TABLE 6 Response in other Metastatic Sites in Patients Showing Objective Regression in the Brain Patients showing objective regression in the brain Extracerebral sites not present Extracerebral sites present but not evaluable Extracerebral sites present and evaluable Progression No change Partial response Complete response Breast Lung 21 6 4 11 2 3 3 3 13 2 3 8 — 2 5 1 treatment. Patients received a median of 2 cycles, ranging from 1 to 6. Table 6 reports the responses assessed in patients with metastases at other sites who showed an objective regression in BM. Among the 21 BC patients, 11 (52%) had evaluable extracerebral metastatic disease, and the types of response were CR in 3 patients, PR in 3, NC in 3, and PD in 2. Among the 13 NSCLC patients, 8 (62%) had evaluable extracerebral metastatic disease, and the types of response were CR in 1, PR in 5, and NC in 2. Time Parameters Table 7 shows the time parameters. The TTP was about 4 months for both BC and NSCLC. The median duration of response was about 8 months for both BC and NSCLC patients. The median survival was approximately 8 months for both BC and NSCLC patients. Approximately one-third and one-fourth of patients with BC and NSCLC, respectively, were still alive after 1 year. The TTP and survival of patients with MM were considerably shorter than those of patients with BC and NSCLC. Toxicity In general, the toxicity from this combination was similar to that observed in our previous studies of BC13 and NSCLC.17 In the current series, there were six important toxic events. Three patients (3%) died, 1 of gastric bleeding, 1 of gastric perforation and peritonitis, and 1 of neutropenic fever and diarrhea. Three patients had to suspend chemotherapy treatment because of atrial fibrillation in one case, severe tubular kidney damage in one case, and febrile neutropenia with pneumonitis in one case. DISCUSSION To our knowledge, this is the largest study and one of the largest trials in which chemotherapy was administered as a front-line treatment for patients 1603 with BM from solid tumors. Until now, two barriers have prevented a rigorous assessment of the activity of chemotherapy in this setting. The first is the concept that the BBB would not allow the watersoluble drugs to reach the tumor target. The second is the practically exclusive administration of lipid soluble drugs, mostly nitrosureas, which are known to be completely inactive in most solid tumors, such as breast carcinoma. In evaluating the activity of chemotherapy against BM, another problem centers around methodologic difficulties in assessing response. Indeed, the response evaluation might be influenced somewhat by the concomitant widespread administration of corticosteroids. However, it is unethical to negate the administration of corticosteroids that are known to provide symptomatic improvement by reducing cerebral edema. For this reason, it is difficult to distinguish the independent effect of corticosteroids from that of chemotherapy. In any case, the regression of the edema, which is the earliest and most dramatic effect of corticosteroid treatment, was not considered per se to certify a tumor response. Moreover, the proportion of patients treated with corticosteroids (76% overall) was similar between the responders and nonresponders. Therefore, the possibility currently remains that part of the reduction in tumor mass may be due to corticosteroid administration. In BM from BC, this chemotherapy induced CR in 13% of the patients and CR or PR in a total of 38%. These figures are lower than those reported in our previous trial13 of the same therapy, probably because the previous study was practically carried out as a single-institution study. On the other hand, the response rate achieved in this trial of BM was intermediate, between 58% for the untreated patients14 and 17% for the heavily pretreated patients reported by our group when this chemotherapy was administered to patients with extracranial disease.18 Indeed, even in this study a large proportion of patients had been pretreated with different chemotherapies. Therefore, it appears that this chemotherapy manifests the same range of activity in the brain and in other extracerebral disease sites. Moreover, these results confirm that the BBB is probably largely disrupted in brain metastases, which are supplied by their tumoral neoangiogenesis. A limited number of studies confirm the activity of other chemotherapies against brain metastases from BC. Rosner et al. reported 50% neurologic responses (CR in 10%) of 100 patients treated with different types of conventional chemotherapy (cyclophosphamide, 5-fluorouracil, and prednisone; cyclophosphamide, 5-fluorouracil, prednisone, methotrexate, and vincristine; or methotrexate, vincristine, and prednisone), but only a few patients were evaluated with CT scan.12 1604 CANCER April 1, 1999 / Volume 85 / Number 7 TABLE 7 Time Parameters Breast Time to progression Duration of response Survival % of patients still alive After 3 mos After 6 mos After 9 mos After 1 yr After 2 yrs After 3 yrs Lung Melanoma No. Median (range) No. Median (range) No. Median (range) 56 21 56 17 (0–2871) 33 (8–93) 31 (0–287) 43 13 43 17 (0–3921) 31 (9–3921) 32 (0–392) 8 — 8 11 (2–39) — 17 (2–48) 73% 56% 37% 32% 9% 3% Kolaric et al. reported 4 CR and 3 PR among 8 patients treated with P as a single agent who were evaluated only by neurologic examination and brain scintigraphy.11 More recently, Boogerd et al. confirmed that conventional chemotherapies (cyclophosphamide, methotrexate, and 5-fluorouracil or 5-fluorouracil, doxorubicin, and cyclophosphamide) induced a 54% response rate in 22 patients evaluated by CT scan.19 The median survival of approximately 8 months and the 32% 1-year survival of BC patients in this study compare favorably with a median survival of only 3–4 months and 1-year survival of 15% reported for patients treated with radiation and corticosteroids only.6 In patients with BM from NSCLC, this chemotherapy induced CR in 7% and CR or PR in at total of 30%. Even in this type of primary tumor, the response rate achieved in BM is similar to the response rate (23%) reported by our group in a randomized trial in which the PE combination was administered to 115 patients with advanced extracranial NSCLC.17 In the literature, there are only a few small studies assessing the effects of other front-line chemotherapies for patients with brain metastases from NSCLC. Malacarne et al., who administered carboplatin and etoposide to a small series of 18 patients, reported a 17% PR rate.20 The median survival of about 7 months and the 25% 1-year survival of the current study for patients with NSCLC compare favorably with the 3.3-month median survival reported administering radiation therapy and corticosteroids.21 The results of this chemotherapy for the few patients with BM from MM are discouraging. This may be due, in general, to the poor sensitivity of MM to chemotherapy and particularly to this combination. The protocol did not include specific recommendations of further therapies after protocol treatment, 73% 56% 45% 25% 10% 7% 50% 25% 13% — — and the individual decisions regarding further therapy were left to each investigator. However, sequential radiation therapy to the brain at the end of 6 cycles was mentioned as being allowed in the protocol. Unfortunately, no specific information was requested regarding this matter. In contrast, systemic therapies (mostly chemotherapy, but for the breast possibly endocrine therapy or both) were not allowed before the assessment of tumor progression in any disease site, according to the protocol. The possibility that these salvage treatments influenced the duration of overall survival cannot be ruled out, but they were clinically unavoidable in treating a certain proportion of the patients with BM, mostly because of the preterminal phase of the natural history of the disease. Regarding toxicity and side effects, a formal evaluation of the quality of life for patients was unfortunately not conducted in this study, as it was in most other studies performed in the 1980s. Apparently chemotherapy with this combination could be worse than about 4 weeks of radiation, although a direct comparison is not available. In any case, the toxicity from this combination was similar to that observed in our previous studies of breast carcinoma and NSCLC,13,17,18 and it is therefore feasible and generally tolerable. However, we observed a few cases of fatal gastric bleeding and intestinal perforation, which were probably due to the combined use of corticosteroids. This suggests that corticosteroid treatment combined with chemotherapy should be cautiously associated with gastroprotective agents and strictly monitored. Our study demonstrates that the most important determinant of responsiveness of BM from solid tumors is the intrinsic sensitivity of the neoplastic cells to the chemotherapeutic agents. Only in BM not yet vascularized, as in the adjuvant setting, would the Chemotherapy in Brain Metastases from Solid Tumors/Franciosi et al. water-soluble drugs be unable to overcome the stillintact BBB to reach high concentrations in the brain and to exert cytotoxic activity against metastatic tumor cells. In fact, the concentration of chemotherapeutic agents is usually very low or absent in cerebrospinal fluid or in normal brain tissue,22 and in adjacent areas of the normal brain it is inversely related to the distance from the tumor.23 On the other hand, when the BM are clinically detected, the BBB is often at least partially disrupted, and the water-soluble drugs can reach cytotoxic concentrations in the tumor tissue. Several reports have shown that the concentration of some chemotherapy agents is similar in intracerebral and extracerebral tumors.22 Platinum has been shown to reach potentially cytotoxic concentrations in intracerebral tumors similar to those in extracerebral tumors.23 Etoposide has been found in intracerebral tumors in variable and sometimes very high concentrations.24 The activity of the PE combination against BM from BC and NSCLC is of the same order as that achieved in extracerebral sites. We do not know whether this effect reflects a general sensitivity of BM to any potentially active chemotherapy or has to be considered a specific result of this combination. Indeed, there are too few studies in the literature in which other chemotherapies were administered to allow a general conclusions regarding a “normal” sensitivity of BM to any type of potentially active chemotherapy. A final consideration regards the relative role of chemotherapy and radiation therapy as treatment modalities in brain metastases from BC and NSCLC. Radiation therapy is a potentially effective treatment. However, the survival results achieved in this study, which compare favorably with those reported for radiation therapy only, suggest that further randomized studies are warranted to compare radiation therapy with chemotherapy or to compare radiation therapy with chemotherapy that is combined with or followed by radiation therapy. 6. REFERENCES 20. 1. 2. 3. 4. 5. Batzdorf U, Block KL, Selch MT. Neoplasms of the nervous system. In: Haskell CM. Cancer treatment. 3rd edition. Philadelphia, W. B. Saunders Company, 1990:436 –74. Di Stefano A, Yap HY, Hortobagy GN, Blumenshein GR. The natural history of breast cancer patients with brain metastases. Cancer 1979;44:1913– 8. Posner JB. Management of central nervous system metastases. Semin Oncol 1977;4:81–91. Rizzi A, Tandini M, Rocco G, Rossi G, Robustelli M, Radaelli F, et al. Lung cancer with a simple brain metastasis: therapeutic options. Tumori 1990;76:579 – 81. Balch CM, Milton GW. Diagnosis of metastatic melanoma at distant sites. In: Balch CM, Milton GW, editors. Cutaneous melanoma: clinical management and treatment results wordwide. Philadelphia: J. B. Lippincott, 1985:221. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 21. 22. 23. 24. 1605 Gelber R, Larson M, Borgelt B, Kramer S. Equivalence of radiation schedules for the palliative treatment of brain metastases in patients with favorable prognosis. Cancer 1981;48:1749 –53. Greig NH. Brain tumors and the blood tumor barrier. In: Neuwelt EA, editor. Implications of the blood brain barrier and its manipulation. Volume 2. New York: Plenum, 1989: 77–106. Greig NH. Chemotherapy of brain metastases: current status. Cancer Treat Rev 1984;11:157– 86. Shapiro W. Chemotherapy of metastatic central nervous system carcinoma. In: Weiss L, Gilbert H, Posner J, editors. Brain metastases. The Hague: Martinus Nijhoff, 1980. Wilson C. Single agent chemotherapy of brain tumors: a five-year review. Arch Neurol 1976;33:739 – 44. Kolaric K, Roth A, Jelicic I. Phase II clinical trial of cisdichlorodiammine platinum (cis-DDP) for antitumorigenic activity in metastatic brain tumor. J Cancer Res Clin Oncol 1982;104:287–93. Rosner D, Nemoto T, Warren WL. Chemotherapy-induced regression of brain metastases in breast carcinoma. Cancer 1986;58:832–9. Cocconi G, Lottici R, Bisagni GC, Bacchi M, Tonato M, Passalacqua R, et al. Combination therapy with platinum and etoposide of brain metastases from breast carcinoma. Cancer Invest 1990;8:327–34. Cocconi G, Bisagni GC, Bacchi M, Boni C, Bartolucci R, Ceci G, et al. Cisplatin and etoposide as first line chemothrapy for metastatic breast carcinoma: a prospective randomized trial of the Italian Oncology Group for Clinical Research. J Clin Oncol 1991;9:664 –9. Durand RF, Goldie JH. Interaction of etoposide and cisplatin in an in vitro tumor model. Cancer Treat Rep 1987;71:673–9. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer 1981;47:207–14. Crinò L, Clerici M, Figoli F, Carlini P, Ceci G, Cortesi E, et al. Chemotherapy of advanced non-small-cell lung cancer: a comparison of three active regimens. A randomized trial of the Italian Oncology Group for Clinical Research (GOIRC). Ann Oncol 1995;6:347–53. Ceci G, Bisagni GC, Cocconi G, Rodinò C, Belsanti V, Bertusi M, et al. Cisplatin and VP16 in metastatic breast carcinoma as a third line chemotherapy: a randomized study comparing low vs. high doses of cisplatin. Tumori 1995;81:241– 4. Boogerd W, Dalesio O, Bais EM, Van der Sande JJ. Response of brain metastases from breast cancer to systemic chemotherapy. Cancer 1992;69:972– 80. Malacarne P, Santini A, Maestri A. Response of brain metastases from lung cancer to systemic chemotherapy with carboplatin and etoposide. Oncology 1996;53:210 –3. Ryan GF, Ball DL, Smith JG. Treatment of brain metastases from primary lung cancer. Int J Radiat Oncol Biol Phys 1995;31:273– 8. Stewart DJ, Leavens M, Friedman J. Penetration of N-(phosphonacetyl)-L-aspartate into human central nervous system and intracerebral tumor. Cancer Res 1980;40:3163– 6. Stewart DJ, Leavens M, Maor M. Human central nervous system distribution of cis-diamminedichloroplatinum and use as a radiosensitizer in malignant brain tumors. Cancer Res 1982;42:2474 –9. Stewart DJ, Richard MT, Hugenholtz H. Penetration of VP-16 (etoposide) into human intracerebral and extracerebral tumors. J Neurooncol 1984;2:133–9.