Knee scans may help rule out osteoporosis in surgery candidates
The model scored 0.885 on a measure of distinguishing osteoporosis from other bone-density categories in external validation; the authors considered the findings hypothesis-generating.
Paper by Shaojian Shi, Zhijie Wang, Yongqiang Zheng et al.
Our read
Biomarkers- Studied:
- Accepted, not yet published: cross-sectional study of 310 older adults awaiting total knee replacement
- What changed:
- The knee-scan model distinguished osteoporosis from other bone-density categories with an area under the curve of 0.885
- The catch:
- Bony growths were associated with more missed osteoporosis cases without structural corrections
Editor of record: Boris Djordjevic · AgingResearch.News
Older adults awaiting total knee replacement had knee computed tomography (CT) scans and reference bone-density tests within one month of each other in a retrospective diagnostic-accuracy study. The analysis included 822 patients for model development and internal validation and 310 for external validation.
The best-performing model used manually outlined scan regions to distinguish normal bone density, low bone density short of osteoporosis, and osteoporosis, a condition that weakens bones. In external validation, its score for distinguishing osteoporosis from the other categories—the area under the curve—was 0.885. Bony growths around the joint were associated with more missed osteoporosis cases; adding structural measurements improved detection. Automated outlining performed numerically close to manual outlining but did not meet the preset standard for being acceptably close.
Why it matters
Osteoporosis often goes unrecognized in older knee-replacement candidates. Scans already taken for surgery may offer a way to prioritize further bone-density assessment.
Caveats
Bone measurements around the knee explained about one-third of the variation in central bone density, limiting performance. The retrospective findings in untreated knee-replacement candidates remained hypothesis-generating.
The paper
Shaojian Shi, Zhijie Wang, Yongqiang Zheng et al.
Frontiers in Medicine, Nov 11, 2026, CC BY, Accepted, not yet published
Comments
Be the first to comment.