High-dose chemotherapy: Is it standard management for any common solid tumor?

Abstract

High-dose chemotherapy with stem-cell support had as its basis the observation of dose-response relationships for many chemotherapeutic agents in laboratory models. The rationale to explore high-dose treatment in the clinic was further enhanced by several retrospective reviews in the 1980s which suggested delivered dose intensity of treatment was an important determinant of patient outcome. The availability of hematopoietic growth factors and technologic advances in the efficiency of stem-cell collection and administration have made the evaluation of exploring high-dose therapy safe and feasible. However, real questions remain regarding the apparently superior results of this treatment in the management of solid tumors. This paper reviews the results of high-dose chemotherapy in breast, ovarian and small cell lung cancers. Firstly the evidence for a dose-response relationship to chemotherapeutic agents in the 'standard' dosage range is examined. Secondly results of non-randomized and, where available, randomized trials of high-dose chemotherapy (HDCT) with stem-cell support are summarized and finally conclusions regarding the weight of the evidence for use of HDCT as 'standard' treatment are given. In none of these tumors is there sufficient evidence from randomized trials to consider HDCT a standard to be offered to all patients with a given stage of disease. The apparent benefit of HDCT seen in phase II trials could well be explained by such phenomena as stage shifts and patient selection. Many randomized trials in ovary and breast cancer are either ongoing or presented only as abstracts so final results must be awaited to quantify the benefit, if any of HDCT. It is acknowledged, however, that some practitioners already utilize this treatment. We speculate about the differences in philosophical approaches to cancer treatment which might contribute to early acceptance of novel therapies in the absence of adequate randomized data.