A consecutive case retrospective chart and radiographic review. To determine the incidence of nine radiographic dystrophic features acquired during the process of modulation, and to analyze the statistical correlation of these acquired dystrophic features with clinical progression of a spinal deformity. In patients with neurofibromatosis, spinal deformities with seemingly few initial dystrophic features have shown a tendency to acquire dystrophic changes during long-term follow-up periods. Similarly, deformities with dystrophic changes can acquire further dystrophic features. This phenomenon is termed "modulation," a feature unique to spinal deformities in neurofibromatosis. These dystrophic changes may evolve slowly or aggressively, and may spread to other regions as well. A review was done of the clinical records, photographs, radiographs, and other imaging studies of 457 patients referred between 1982 and 1995 with the diagnosis of neurofibromatosis Type 1. One hundred twenty-eight patients were diagnosed with a spinal deformity. Ninety-one patients who had a complete set of clinical and radiographic data were included in the study. Location and type of curve as well as the extent of spinal deformity were studied for their effect on the tendency for modulation. Initial spinal radiographs were analyzed for nine radiographic dystrophic features: rib penciling, vertebral rotation, posterior vertebral scalloping, anterior vertebral scalloping, lateral vertebral scalloping, vertebral wedging, spindling of the transverse process, widened interpedicular distance, and enlarged intervertebral foramina. Subsequent radiographs were analyzed critically for evolution, progression, or spread of these features. Correlation of acquisition in these dystrophic features with clinical progression in the spinal deformity, as measured in increments of scoliosis and kyphosis, was analyzed. In 81% of patients with spinal deformity diagnosed before 7 years of age and in 25% of patients with such a diagnosis after 7 years of age, evidence of modulation was observed. Location, side, and extent of the deformity and patient gender did not influence the propensity of the deformity to modulate. Correlation of modulation with clinical progression of the deformity showed rib penciling to be the only singular factor statistically influencing risk of progression. Of the deformities that acquired three or more penciled ribs, 87% showed significant clinical progression. No other radiographic dystrophic feature individually influenced progression. However, when three or more of the dystrophic skeletal features were acquired, the risk of progression reached statistical significance in 85% of patients. Spinal deformities in patients with neurofibromatosis 1 should be regarded as deformities in evolution. One should resist assigning these evolving deformities to either the dystrophic or nondystrophic end of the spectrum without considering the possibility of modulation across the spectrum. A spinal deformity that develops before 7 years of age should be followed closely for evolving dystrophic features (i.e., modulation). When a curve acquires either three penciled ribs or a combination of three dystrophic features, clinical progression is almost a certainty.