Laser Doppler flowmetry is a simple method of determining, directly and continuously, tissue blood flow. However, its applicability to monitoring tumour blood flow interstitially during hyperthermia treatments is still being evaluated. The purposes of this study were to physically characterize the measurement probes, to evaluate potential sources of artifact with the interstitial use of the probes during hyperthermia treatment, and to obtain measurements in human tumours during hyperthermia sessions. The accuracy of the method in quantifying blood flow, velocity and volume during hyperthermia was found to be unaffected by heating the measurement probe to 42-46 degrees C or by exposing it to various intensities of 915 MHz microwave fields (10-40 W), or 1 MHz ultrasound fields. Catheter insertion methods were developed to place the flow probes interstitially in tumours. Tissue damage was confined to a distance of no greater than 0.12 mm away from the catheter tract, and physical evidence of vascular disruption was within a distance of 0.05 mm as measured in a rat tumour model. This degree of damage/disruption is unlikely to affect LDF measurements which represent blood flow averaged over a 1.0-1.5 mm radius from the probe tip. Concurrently, the device was used to monitor tumour blood flow parameters interstitially in human subjects during hyperthermia treatments given in combination with conventional radiotherapy. Blood-flow data from multiple sites of measurement showed marked heterogeneity within individual tumours (up to 55-fold differences) and between different tumours (greater than 100-fold differences). Measurements made by translating the probe along a tumour radius, beginning at the tumour core and advancing to the tumour edge, were consistent with a two-component tumour perfusion model (shell and core). Data are presented from one patient illustrating a persistent change in perfusion distribution during the hyperthermia treatment course, which occurred concomitantly with increases in thermal data. These results suggest that the technique might be of value in monitoring change in flow between treatments. Responses during hyperthermia treatment sessions were also investigated. Four temporal patterns of flow were observed, ranging from a steady increase in flow to a plateau level to a steady drop in flow during heating. These patterns were not well correlated with average temperature recorded at the site of flow measurement. Further study is needed to determine if this LDF technique is to be useful for evaluation of heat transfer by blood perfusion.