A human glioma spheroid model is used to investigate the efficacy of different light delivery schemes in 5-aminolevulinic acid (ALA)--mediated photodynamic therapy (PDT). The results provide the rationale for the development of an indwelling balloon applicator for optimizing light delivery. Human glioma spheroids were incubated in ALA (100 or 1000 microg /ml-1) for 4 hours and subjected to various light irradiation schemes. In one set of experiments, spheroid survival was monitored as a function of light fluence rate (5-200 mW cm-2). In all cases, spheroids were exposed to fluences of either 25 or 50 J cm-2. In a second study, the effects of repeated weekly PDT treatments, using sub-threshold fluences, were investigated. One group of spheroids was subjected to three treatments using fluences of 12, 12, and 25 J cm-2. Results were compared to spheroids receiving single treatments of either 12 or 25 J cm-2. A fluence rate of 25 mW cm-2 was used for all three groups of spheroids. In all cases, the effect of a given irradiation scheme was evaluated by monitoring spheroid growth. Low fluence rates produce greater cell kill than high fluence rates. The minimum effective fluence rate in human glioma spheroids is approximately 10 mW cm-2. Repeated weekly PDT treatments with sub-threshold fluences result in significant cell kill. In spheroids surviving the PDT treatments, growth is suppressed for the duration of the treatment period. The results of the in vitro studies support the development of an indwelling balloon applicator for the delivery of light doses in long term multi-fractionated PDT regimens.