S_mary Photodynamic therapy (PDT) has the potential to destroy small tumours with safe healing of adjacent normal tissue. This study looks at the effects of PDT on the normal pancreas and adjacent tissues in hamsters using the photosensitiser meso-tetrahydroxyphenylchlorin (mTHPC). Pharmacokinetic studies used fluorescence microscopy on sections of pancreas, stomach and duodenum 1 h to 6 days after mTHPC. Highest levels of sensitiser were seen in the gastric and duodenal mucosa and in the acinar pancreas after 2-4 days. For PDT, light at 652 nm was delivered by placing a 0.2 mm diameter bare-ended fibre against the tissue. An energy of 50 J was used 2 or 4 days after 0.1 or 0.3 mg kg-l mTHPC and animals killed I to 7 days later.Maximum necrosis was seen 3 days after PDT with lesions up to 4 mm in pancreas, 4.5 mm in duodenum and 2.5 mm in stomach. By fractionating the light dose, the lesion size could be increased by 30%. The main complication was free or sealed duodenal perforation (avoided by shielding the duodenum). Partial, reversible bile duct obstruction was seen occasionally. There was no macroscopic damage to the bile ducts or major blood vessels. Apart from the duodenum, all lesions healed safely. In this animal model, only the duodenum was at risk of serious, irreversible damage. Treatment is likely to be safer in the much thicker human duodenum. mTHPC is a powerful photosensitiser and suitable for fiurther study for tumours in the region of the pancreas although care is required near the duodenum.Keywords: photodynamic therapy; pancreas; duodenum; stomach; bile duct Photodynamic therapy (PDT) involves the local activation of a preadministered photosensitiser by light of a wavelength matched to the absorption characteristics of the photosensitiser. The activated photosensitiser gives rise to the production of cytotoxic singlet oxygen (Weishaupt, 1976). In many publications it has been shown that it is relatively easy to destroy small volumes of a wide variety of tumours with PDT (Li et al., 1990. However, what matters to a patient is whether the entire tumour volume can be destroyed without unacceptable damage to adjacent normal tissues. This means that it is essential to understand what happens in the region where tumour is invading normal areas.Until now little work has been done on this aspect (Bown, 1990). Although much of the interest in PDT has centred around the possibility of selective destruction of tumours, this aspect is almost always over emphasised, and truly selective destruction of cancers is virtually impossible (Barr et al., 1990. Many normal tissues heal mainly by regeneration after PDT, but for tissues like muscle, there is only partial restoration of function (Chevretton et al., 1992). A previous report (Schroder et al., 1988) showed that PDT will produce necrosis in a chemically induced pancreatic cancer in hamsters using dihaematoporphyrin ether (DHE) but at the price of duodenal perforation. The most studied photosensitiser, haematoporphyrin derivative (HpD) and purified fraction...