Bone Abstracts

Histomorphometric analysis is an essential technique to measure bone formation and resorption parameters. Here we present three novel open source image analysis packages that allow the rapid semi-automated analysis of histomorphometric bone resorption, osteoid, and calcein double labelling parameters.

Mice were injected with calcein 5 and 2 days before killing. Tibia and vertebrae were fixed in formalin, embedded in methylmethacrylate and the blocks sectioned at 5 μm. For measurement of trabecular architecture and osteoid, sections were stained with van Gieson/von Kossa, and for analysis of resorption parameters with Aniline Blue and TRAP. Calcein Blue was used as a fluorescent bone stain that does not affect calcein labels. Software was developed in Java using Netbeans and ImageJ as an image analysis library.

The resulting three software packages, osteoid Histo, Trap Histo and calcein Histo, use a relative simple wizard like user interface to guide users through the analysis. The Trap Histo and Osteoid Histo programs identify bone, osteoclasts, and osteoid by colour thresholding in combination with object filtering tools. Semi-automated detection of the calcein labels resulted in a substantial decrease in analysis time required compared to manual drawing of the labels. However, labels cut at an oblique angle were not well recognised and required manual editing. The use of this semi-automated software lead to good reproducibility of measurements with intra-observer CV in the range of 1–4%, and inter-observer CV (three observers) in the range of 2–8% for resorption parameters. Measurements of bone architecture using the osteoid Histo program were highly reproducible as well with inter-observer CV% of 2–4% for BV/TV, BS/BV, Tb.Th and O.Th. However, OS/BS and OV/BV showed higher variability with CV% between 8–15%.

In conclusion, the method for performing bone histomorphometry described here is relatively easy and reproducible. For well stained, good quality sections, analysis can be performed in <5 min/section.