Bone Abstracts (2016) 5 OC1.6 | DOI: 10.1530/boneabs.5.OC1.6

Standard QCT shows substantial underestimation of treatment effects on bone mineral density: impact of spatial resolution

Stefan Reinhold1,2, Timo Damm1, Jaime Peña1, Reinhard Barkmann1 & Claus-C Glüer1


1Sektion Biomedizinische Bildgebung, Klinik für Radiologie und Neuroradiologie, Christian-Albrechts-Universität, Kiel, Germany; 2Institut für Informatik, Christian-Albrechts-Universität, Kiel, Germany.


Quantitative Computed Tomography (QCT) can be performed with different protocols. We investigated whether selection of spatial resolution, reconstruction method, and scanner type have an impact on measured treatment effects for different bone compartments.

Methods: In a multi-center study, a group of 50 men, age 55.1 (range 25–78) undergoing 18 months of either osteoanabolic (n=25) or antiresorptive treatment (n=25) were scanned with QCT (120 kVp, 100 mAs, slice thickness 2.5–3 mm, in-plane pixel size 0.6 mm) and High Resolution QCT (HR-QCT: 120 kVp, 360 mAs, slice thickness 0.6–0.8 mm, reconstruction increment 0.3–0.4 mm, in-plane pixel size 0.16–0.19 mm). The protocol specified measurement at L1 for QCT and T12 for HR-QCT but due to fractures of T12 in seven patients HR-QCT and QCT were both obtained at L1 (thus impact of matching vertebral could be tested). We compared integral, cortical and trabecular bone mineral density (BMD) i) at baseline and ii) their changes during treatment (ΔBMD) as observed by QCT and HR-QCT, adjusting for vertebral level mismatch. All results are expressed as mean±SEM.

Results: The table reveals large underestimation of treatment effects of 34–69% for standard QCT. The magnitude of underestimation was positively corrected with baseline BMD (P<0.0001 for all three compartments) and varied by scanner type/reconstruction kernel. Vertebral level mismatch had no additional independent significant influence.

Bone CompartmentHR-QCT BMD baselineQCT BMD baselineHR-QCT ΔBMDQCT ΔBMDError BMDError ΔBMD
Trabecular71.7±4.680.3±4.410.1±2.36.6±2.3+12%−35%
Cortical271.2±8.1171.4±.535.1±3.310.9±2.4−37%−69%
Integral134.4±4.8127.9±4.016.9±2.711.1±2.4−5%−34%

Conclusion: Spatial resolution has a strong effect on CT-based BMD evaluations. HR-QCT reveals 45% (trabecular, integral) to 2.9 times (cortical) larger treatment effects for BMD and true treatment effects may be even larger due to spatial resolution limits of HR-QCT. This has major implications for the interpretation of QCT-based finite element modeling and for standardization of QCT studies.

Article tools

My recent searches

No recent searches.