Development and validation of a multimodal interpretable machine learning model for the identification of osteoporosis in patients with type 2 diabetes mellitus: a multicenter retrospective study - Takeaways - MDSpire

Development and validation of a multimodal interpretable machine learning model for the identification of osteoporosis in patients with type 2 diabetes mellitus: a multicenter retrospective study

  • By

  • Jihao Cheng

  • Chuanjiang Liu

  • Mengyin Gu

  • Dongying Su

  • Jingyun Liao

  • July 1, 2026

  • 0 min

Share

  • 1

    Osteoporosis is a common but underdiagnosed complication in type 2 diabetes mellitus (T2DM), increasing fracture and mortality risks.

  • 2

    A multimodal machine learning model was developed using data from 1,002 T2DM patients to predict osteoporosis risk.

  • 3

    The eXtreme Gradient Boosting model achieved an AUC of 0.877 in training and 0.911 in external validation for osteoporosis detection.

  • 4

    Key predictors identified include age, hemoglobin, neutrophil count, uric acid, lymphocyte-to-HDL ratio, skeletal muscle index, and metabolic score.

  • 5

    The model's interpretability was enhanced using SHapley Additive exPlanations (SHAP), revealing significant contributions from age, hemoglobin, and uric acid.

Original Source(s)

Development and validation of a multimodal interpretable machine learning model for the identification of osteoporosis in patients with type 2 diabetes mellitus: a multicenter retrospective study

  • by Jihao Cheng, Chuanjiang Liu, Mengyin Gu, Dongying Su, Jingyun Liao

  • July 1, 2026

Related Content

An interpretable deep concatenated architecture for osteoporosis detection using enhanced knee radiographs

Unveiling the efficacy predictors and potential mechanisms of Semen Cuscutae against osteoporosis via machine learning and meta-analysis: a preclinical study

Biological Retinal Age Predicts Fractures

New study says AI-based retinal imaging has potential to be used as a screening tool for osteoporosis risk

  • June 22, 2026

  • 3 min