Diseases involving hip may influence on the anatomy of the proximal femur, especially on the endosteal canal shape. The simple radiographic comparative study was performed between normal and diseased hips. For the diseasd hip group, 74 hips ol' femoral head AVN(avascular necrosis) (average age: 49.3 years old, patients were all male) who had limped for average l9 months were selected and 77 male hips (average age: 48.9 years old) who had no prior hip disease were selected for the normal hip group. The external shape of both group showed no significant differences except for the femoral head dimensions. Statistically significant differences were ohserved hetween two groups in the proximal endosteal canal shape and cortical thickness. The femur of the diseased hip group had wider canal from the level of distal to lesser trochanter to the level of isthmus, with a significant reduction in the metaphyseal index (the ratio hetween the canal width proximal to lesser trochanter and distal to lesser trochanter) and the canal flare index (p<0.001). Theoretically this study suggests that the optimal fit and fill to the proximal endostcal canals of normal and diseased hips cannot be obtained by oneshape cementless femoral stcms.