Summary Although small cell lung cancer (SCLC) is very chemosensitive, cerebral metastases are treated with radiotherapy in the belief that they are protected from chemotherapy by the blood-brain barrier (BBB). The validity of this assumption has not been tested in clinical practice. In a randomised trial of treatment in 610 patients with SCLC, 19 patients who had symptomatic cerebral metastases at presentation were treated initially with chemotherapy, and cranial irradiation withheld. (Long et al., 1979). Indeed the increased permeability of tumour vessels to radiolabelled colloids and CT contrast media is fundamental to the radiological diagnosis of cerebral metastases. Despite this, when treating cerebral metastases it is often assumed that, because most cytotoxic drugs do not cross the intact BBB, chemotherapy will be ineffective.Even in SCLC, which is a highly chemosensitive tumour, chemotherapy has been largely ignored in the management of cerebral metastases, these patients being treated with steriods and cranial irradiation (Cox et al., 1980). There are reports of radiologically proven responses of cerebral metastases to systemic chemotherapy in SCLC (Kantarjian et al., 1984;Postmus et al., 1987; Kristiansen & Hansen, 1988) but the response rate to a single regimen is not known. However, the only systematic study of conventional chemotherapy for cerebral metastases (Rosner et al., 1986) showed that cerebral metastases from breast cancer have the same frequency of response as secondary deposits at other sites.Our aim was to assess in a prospective study the objective response rate of cerebral metastases in previously untreated SCLC patients who received uniform chemotherapy, rather than radiotherapy, as initial treatment.
MethodsBetween February 1982 and September 1985, 610 patients with histologically or cytologically confirmed SCLC entered a multicentre randomised chemotherapy trial (Spiro et al., 1989). They had no past history of malignancy and had not received previous radiotherapy or chemotherapy. Brain scans were not performed routinely, and only patients with symptoms or signs of cerebral metastases at presentation had a CT or radionuclide brain scan before starting treatment. Cerebral metastases were diagnosed by the presence of enhanced lesions on CT or areas of increased uptake on isotope brain scan, compatible with the clinical findings. In these patients, cranial irradiation was withheld and initial treatment was with chemotherapy. Patients with a severe neurological deficit received oral dexamethasone, but if possible the dose was reduced during the course of chemotherapy. Steroids were not used as anti-emetics. The chemotherapy was cyclophosphamide I g m-2 i.v. day 1, vincristine 2 mg i.v. day I and etoposide 100 mg tds p.o. days 1-3. If possible the CT scan was repeated before the second cycle of chemotherapy, and it was planned that all patients be evaluated with a further scan after four cycles of chemotherapy.The study was designed to assess the response of cerebral metastases to che...