Predicting femoral neck strength from bone mineral data: A structural approach

Thomas J. Beck, Christopher B. Ruff, Karen E. Warden, William W. Scott, Gopala U. Rao

Research output: Contribution to journalArticle

Abstract

An interactive computer program was developed to derive femoral neck geometry from raw bone mineral image data for an estimate of hip strength using single plane engineering stress analysis. The program, which we call Hip Strength Analysis (HSA), was developed as an attempt to improve the predictive value of hip bone mineral data for osteoporosis fracture risk assessment. We report a series of experiments with an aluminum phantom and with cadaver femora, designed to test the accuracy of derived geometric measurements and strength estimates. Using data acquired with both Lunar DP3 (DPA) and Hologic QDR-1000 (x-ray) scanners, HSA computed femoral neck cross-sectional areas (CSA) and cross- sectional moments of inertia (CSMI) on an aluminum phantom were in excellent agreement with actual values (r >.99). Using necks of 22 cadaver specimens were in good general agreement with literature values. HSA computed cross- sectional properties of three of these specimens were compared with measurements derived from sequential CT cross-sectional images. Discrepancy between the two methods averaged less than 10% along the length of the femoral neck. Finally, breaking strengths of 20 of the femora were measured with a materials testing system, showing better agreement with HSA predicted strength (r =.89, percent standard of the estimate (%SEE) = 21%) than femoral neck bone mineral density (r =.79, %SEE = 28%).

Original languageEnglish (US)
Pages (from-to)6-18
Number of pages13
JournalInvestigative radiology
Volume25
Issue number1
DOIs
StatePublished - Jan 1990

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Fingerprint Dive into the research topics of 'Predicting femoral neck strength from bone mineral data: A structural approach'. Together they form a unique fingerprint.

  • Cite this