Abstract: With the deep integration of artificial intelligence and medicine, predictive models have demonstrated unique advantages in the early detection and personalized treatment strategies of osteoporosis, which can improve the prognosis of patients with osteoporosis and reduce the incidence of osteoporotic fractures. This study proposed a multimodal integrated model framework combining radiomics features with biomarkers and summarized the novel indicators included in recent research on predictive models for osteoporosis.