The accuracy of a radiographic absorptiometry (RA) technique called digital image processing (DIP), discriminative ability of RA for osteoporotic fracture, and the relationship between RA and dual X-ray absorptiometry (DXA) of the spine and forearm were evaluated. We measured 16 cadaver hands, 32 healthy non-black premenopausal women, 39 healthy non-black postmenopausal women, and 35 non-black osteoporotic postmenopausal females. The overall correlation between the ash weights of the entire metacarpal and the DIP values was excellent (r = 0.954, P < 0.001, SEE = 0.14, CV = 6.4%). Short-term precision error of DIP was 3.5%. Age-related bone loss determined by DIP is comparable to that of spinal and forearm DXA: annual BMD decreases were 0.46% for DIP, 0.45% for forearm, and 0.32% for the spine. DIP of the 2nd metacarpal shows a gradient of risk for spinal fracture only slightly below that of forearm DXA, but substantially below that of spinal DXA. Age-adjusted odds ratios were 1.81 for RA, 2.45 for spinal DXA, and 1.94 for forearm DXA.