Osteoporosis is a silent disease with low screening rates in many developing countries. This study aimed to evaluate the feasibility of using an artificial intelligence (AI)-based system to screen osteoporosis from pelvic and hip radiographs in Vietnam. We conducted a cross-sectional study at a tertiary medical center in Central Vietnam in 2023. A total of 2000 consecutive pelvic and hip radiographs from patients aged ≥ 40 years were collected. After excluding poor-quality images, 1987 radiographs were analyzed using an AI-based software designed to estimate bone mineral density (BMD) from plain radiographs and derive T-scores. Osteoporosis was defined as a T-score ≤ −2.5. Patient characteristics, radiographic findings, and risk factors for osteoporosis were analyzed. Among 1987 patients (mean age 66.4 ± 15.1 years; 41.3% men), osteoporosis was identified in 872 patients (43.9%). The prevalence increased with age and was higher in women than in men (58.7% vs 22.8%, P < 0.001). Osteoporosis was associated with femoral neck (OR = 3.8, 95% CI: 2.7–5.2) and intertrochanteric fractures (OR = 7.0, 95% CI: 4.5–11.0). Patients with lower T-scores had a higher risk of hip fractures, especially those with T-scores ≤ −3.0 (OR = 11.5, 95% CI: 5.5–24.5). AI-based analysis of pelvic and hip radiographs is a feasible and effective tool for osteoporosis screening in Vietnam. The prevalence of osteoporosis in this hospital-based setting was high, particularly among elderly women. AI-assisted screening may offer an accessible strategy for early detection of osteoporosis in resource-limited settings.
Nguyen et al. (Sun,) studied this question.