Table 1 Inclusion and exclusion criteria based on the PICOS framework

Population (P)

Studies assessing patients with suspected or diagnosed osteoporosis through panoramic radiographs.

Studies of patients with suspected or diagnosed osteoporosis based on images of the hip, femur, or lumbar spine.

Studies fewer than 10 participants.

Intervention (I)

Research on deep learning models for osteoporosis diagnosis

Research utilizing non-deep learning techniques (e.g., traditional machine learning and statistical methods)

Comparator (C)

Studies evaluating deep learning models against traditional diagnostic methods, including dual-energy X-ray absorptiometry (DXA), mandibular cortical index (MCI), and expert radiologist assessments.

Outcome (O)

● Studies reporting diagnostic accuracy metrics, including sensitivity, specificity, and area under the curve (AUC).

● Studies that provide data for calculating these metrics.

Studies that do not report diagnostic accuracy or lack sufficient data for metric calculation.

Study Design (S)

● Original peer-reviewed studies utilizing DL algorithms for osteoporosis prediction or diagnosis.

● published in English language

Review, editorials, commentaries, letters to the editor, case series, case reports, conference abstract and preprint articles