Accurate localization of laser light within a tumour lessens the need for selective tumour retention of the photosensitizer. The aim of this study was to investigate different routes of photosensitizer administration for interstitial photodynamic therapy (IPDT) of the liver. Sprague-Dawley rats were photosensitized with HPD 5 mg kg−1 intravascularly at 48 h or by regional administration 60 min prior to light delivery or by interstitial injection (0.04 mg, 0.15 ml) directly into the hepatic parenchyma at 10 and 60 min prior to light delivery. Thirty-two joules of light from a helium-neon (HeNe) laser were delivered interstitially into the median lobe of the liver via a 200-μm optical fibre. Four days after light delivery the liver was harvested, sectioned and stained with haematoxylin and eosin (HE). The maximum cross-sectional area of photodamage was estimated for each photosensitizer administration route in six livers. Both conventional PDT and interstitial routes of administration of the photosensitizer showed comparable areas (±s.e.m.) of bioactivity (8.32±2.03 mm2 and 9.5±1.44 mm2) that were greater than those for control livers treated with light only (1.89±0.39 mm2,p<0.01). The maximum area of biological effect was noticed in livers regionally photosensitized by the portal vein or hepatic artery 60 min prior to light delivery (intraportal vein 13.32±1.52 mm2 and intrahepatic artery 14.21±4.19 mm2,p<0.01). These results suggest that for IPDT, regional administration of a photosensitizer may achieve the greatest biological effect. This route may be the most appropriate route for interstitial PDT using a selective light delivery system within the liver.