Cost-Effectiveness of Opportunistic Osteoporosis Screening Using Chest Radiographs With Deep Learning in the United States_ASBMR2025

Scientific Information: 

Osteoporosis is frequently underdiagnosed due to the limitations of traditional screening methods, resulting in missed opportunities for early intervention. Recent advancements in deep learning models using chest radiographs show promise for opportunistic osteoporosis screening, particularly in middle-aged and older individuals. By enabling automated screening in clinical settings, these models could help reduce the burden of osteoporotic fractures. This study aims to assess the cost-effectiveness of this approach in US women aged 50 and older, providing valuable insights for decision-makers on its potential implementation. An economic model, incorporating both a decision tree and a microsimulation Markov model, estimated the cost per quality-adjusted life year (QALY) gained ($2024) for screening via chest radiographs with deep learning, followed by treatment, versus no screening and treatment. The patient pathways were based on the sensitivity and specificity of Al-enhanced radiographs. Real-world medication persistence, along with realistic assumptions for the probability of DXA examination following screening detection and treatment initiation rates, were incorporated. Women with osteoporosis were stratified into high risk (receiving alendronate monotherapy for five years) and very high risk (receiving an 18-month anabolic treatment with abaloparatide followed by five years of alendronate). Parameter uncertainty was analyzed through sensitivity analyses. The opportunistic screening strategy led to improved health outcomes, generating an additional 1.5 QALYs per 1,000 women screened and preventing 2.8 fractures, though it resulted in higher treatment costs. The cost per QALY gained for opportunistic screening was estimated at $72,085, which remains below the US cost-effectiveness threshold of $100,000 per QALY, indicating its cost-effectiveness. The cost-effectiveness remained favorable as long as Al tool costs were under $62 per patient. Further improvements in cost- effectiveness could be achieved by optimizing follow-up care, treatment initiation, and medication adherence. Specifically, reducing medication non-persistence by 50% improved the ICER to $28,663, and achieving full medication adherence further reduced it to $16,414. In conclusion, this study underscores the cost- effectiveness and public health significance of Al-driven screening, demonstrating its potential to improve early detection and address unmet diagnostic needs in osteoporosis care. 

Title: 

Cost-Effectiveness of Opportunistic Osteoporosis Screening Using Chest Radiographs With Deep Learning in the United States 

Submitter's E-mail Address: 

m.hiligsmann@maastrichtuniversity.nl 

Category 1: 

Clinical: Osteoporosis – Assessment 

Category 2: 

Clinical: Osteoporosis - Treatment 

CareerLevel: 

Senior Investigator 

Copyright Transfer Agreement: 

I agree to the COPYRIGHT TRANSFER AGREEMENT as shown above and have obtained written permission from all other contributors to execute this Agreement on their behalf 

Pre-Meeting Symposium: 

I do NOT want to submit my abstract to any Pre-Meeting symposium. 

Keywords: 

Cost-effectiveness, Opportunistic screening and Osteoporosis 

Preferred Presentation Format: 

Oral or Poster Presentation 

Submitted to pre-print journal or server: 

No 

Recording Permission: 

I confirm that if my abstract is selected for an oral presentation, ASBMR has permission to record my presentation for educational purposes and for viewing by ASBMR members and paid annual meeting registrants. 

Funding from an organization with a proprietary or financial interest: 

Yes 

Funding from the National Institutes of Health (NIH) or other government agency: 

No 

IRB Approval: NotApplicable 

IACUC Approval: 

NotApplicable 

HIPPA Compliant: 

NotApplicable 

Copyright Permission(s): Yes 

Opportunity for Debate: 

Yes 

IRB Approval: 

NotApplicable 

Payment Agreement: 

Yes 

Off-Label Uses and Limitations of Data: 

Yes 

Approve of media/medical journalists capturing photography during presentation: 

Yes 

Principal Investigator 

Mickael Hiligsmann 

Email: m.hiligsmann@maastrichtuniversity.nl -- Will not be published 

Maastricht University 

Netherlands 

Any relevant financial relationships? Yes 

Organization Name 

Relationship 

Angelini Pharma 

Grant/Research Support 

radiomics.bio 

Grant/Research Support 

Signed on 04/14/2025 by Mickael Hiligsmann 

Presenting Author 

Stuart L. Silverman 

Email: stuarts@bhillsra.com -- Will not be published 

Clinical Professor of Medicine, Cedars-Sinai Medical Center 

USA 

Any relevant financial relationships? Yes 

Organization Name 

Relationship 

Promedius 

Consultant 

Amgen 

Consultant 

Consultant 

Radius Health 

Signed on 04/14/2025 by Stuart Silverman 

Author 

Jean-Yves L. Reginster 

Email: jyreginster@ulg.ac.be -- Will not be published 

College of Sciences, King Saud University 

Saudi Arabia 

Any relevant financial relationships? Yes 

Organization Name 

Promedius 

Promedius 

Relationship 

Grant/Research Support 

Consultant 

Signed on 04/14/2025 by Jean-Yves Reginster 

Sponsor 

Stuart L. Silverman 

Email: stuarts@bhillsra.com -- Will not be published 

Clinical Professor of Medicine, Cedars-Sinai Medical Center 

USA 

" 

Sponsor's Response: Has not responded 

If necessary, you can make changes to your abstract submission. 

To access your submission in the future, use the direct link to your abstract submission from one of the automatic confirmation emails that were sent to you during the submission. 

Or point your browser to /asbmr/reminder.cgi to have that URL mailed to you again. Your username/password are 3744/235040. 

Any changes that you make will be reflected instantly in what is seen by the reviewers. You DO NOT need to go through all of the submission steps in order to change one thing. If you want to change the title, for example, just click "Title" in the abstract control panel and submit the new title. 

When you have completed your submission, you may close this browser window. 

Tell us what you think of the abstract submission process 

Home Page