The current study was undertaken to evaluate the clinical utility of DVA, a system for imaging the lateral spine on the Lunar Prodigy densitometer. DVA images were obtained and bone density of the lumbar spine and proximal femur measured in 297 subjects (272 women), aged 64+/-13 years. The images were classified as: normal (N) if no fractures were detected and all vertebrae between T6 and L4 were visualized, fracture (F) if any vertebra had a fracture (defined as 25% or more reduction in the vertebral height) even if some of the other vertebrae could not be visualized, and un-interpretable (U) if at least one of the vertebra between T6 and L4 could not be classified and no fractures were detected in the visualized vertebrae. A subset of 66 patients also had standard radiographs of the thoracic and lumbar spine. Compared to radiographs, DVA had a 95% sensitivity to detect fractures and 82% specificity (to exclude them). Among all 297 subjects studied, DVAs were interpretable in 87%. They were classified as N in 204 (68%), F in 55 (19%) and U in 38 (13%). The reasons for un-interpretability were: scoliosis, scapular or rib shadow, severe arthritic changes and multiple vertebral compression fracture with severe spinal deformities. Only 11% of F subjects gave a history of a vertebral fracture, and only 56% of F subjects met the BMD criteria for osteoporosis (T score <-2.5). These results indicate that adding DVA, a low radiation and relatively low cost "point of service" procedure, to BMD measurement provides the clinician with a more comprehensive fracture risk assessment than that afforded by clinical evaluation and BMD measurement alone.