Quantitation of metabolic changes in tumours may provide an objective measure of clinical and subclinical response to anticancer therapy. This pilot study assesses the value of quantitation of metabolic rate of glucose (MRGlu) measured in mmol min(-1) ml(-1) to assess early subclinical response to therapy in a relatively non-responsive tumour. Nine patients receiving the CRC Phase II study schedule of temozolomide were assessed with [18F]fluorodeoxyglucose ([18F]FDG) dynamic positron emission tomography (PET) scans prior to and 14 days after treatment with temozolomide given as 750-1000 mg m(-2) over 5 days every 28 days. Tumour MRGlu was calculated and compared with objective response at 8 weeks. Pretreatment MRGlu was higher in responders than non-responders. The responding patient group had a greater than 25% reduction in MRGlu in regions of high focal tumour uptake (HFU). Whole tumour changes in MRGlu did not correlate with response. Percentage change in HFU standardized uptake value (SUV) did discriminate the responding from the non-responding patients, but not as well as with MRGlu. Large differences also occurred in the normal brain SUV following treatment. Thus, MRGlu appeared to be a more sensitive discriminator of response than the simplified static SUV analysis. Changes in MRGlu may reflect the degree of cell kill following chemotherapy and so may provide an objective, quantitative subclinical measure of response to therapy.