Usually t.he persistence of ventricular activity in Tc 99m DTPA cisternography is determined using 24 hour images in our country, However, the 24 hour image is regarded as an insufficient data by many investigators. They prefer 48 hour image, The aim of the present study was to evaluate the feasibility and usefulness of 48 hour delayed imaging in Tc-99m DTPA cisternography. We performed 48 hour cisternography using Tc 99m DTPA in 38 pat,ients with known hyclrocephalus, 37 111 MHq (1-3mCi) of Tr, 99m DTPA were injected by lumbar punct.ure. Anterior and both lateral images were obtained at 2, 6, 24 and 48 hours. The classification of hydrocephalus was done at 24 and 48 hours by two independent nuclear medicine physicians. L he 48 hour images were interpretable in all cases. In seven patient,s, the clearance of ven-tricular activity was noted only on 48 hours images. Therefore, their classification was changed from type V to type III. Sixteen of 38 patients were underwent shun(operations. Clinical improvement was noted in 13 patients(1/1 of type II, 3/6 of type Ill, 8/9 of type 1V ). One case who showed a changing pattern from IV to lll a showed no clinical improvement after shunt operation. In conclusion, the 48 hour delayed imaging was feasible and useful technique in Tc 99m DTPA cisternography.