To determine the applicability and safety of a new canine model suitable for correlative magnetic resonance imaging (MRI) studies and morphological/pathophysiological examination over time after interstitial laser thermotherapy (ILTT) in brain tissue. A laser fibre (Diode Laser 830 nm) with an integrated temperature feedback system was inserted into the right frontal white matter in 18 dogs using frameless navigation technique. MRI thermometry (phase mapping i.e. chemical shift of the proton resonance frequency) during interstitial heating was compared to simultaneously recorded interstitial fiberoptic temperature readings on the border of the lesion. To study brain capillary function in response to ILTT over time quantitative autoradiography was performed investigating the unidirectional blood-to-tissue transport of carbon-14-labelled alpha amino-isobutyric acid (transfer constant K of AIB) 12, 36 hours, 7, 14 days, 4 weeks and 3 months after ILTT. All laser procedures were well tolerated, laser and temperature fibres could be adequately placed in the right frontal lobe in all animals. In 5 animals MRI-based temperature quantification correlated strongly to invasive temperature measurements. In the remaining animals the temperature fibre was located in the area of susceptibility artifacts, therefore, no temperature correlation was possible. The laser lesions consisted of a central area of calcified necrosis which was surrounded by an area of reactive brain tissue with increased permeability. Quantitative autoradiography indicated a thin and spherical blood brain barrier lesion. The magnitude of K of AIB increased from 12 hours to 14 days after ILTT and decreased thereafter. The mean value of K of AIB was 19 times (2 times) that of normal white matter (cortex), respectively. ILTT causes transient, highly localised areas of increased capillary permeability surrounding the laser lesion. Phase contrast imaging for MRI thermomonitoring can currently not be used for reliable temperature readings in vivo. The suggested new canine model proved to be safe, accurate, easy to use, and provides clinical, radiographic, pathological and physiological correlations.