Bone mineral density measurements in postmortem computed tomography: An open-source approach

Abstract

Objectives

In quantitative postmortem computed tomography (qPMCT) the presence of putrefaction gas in tissues can obscure measurements such as bone mineral density (BMD). Quantitative CT analysis procedures adopted directly from clinical CT may not be designed to compensate for intracorporeal gas, which require additional measures for PM-imaging applications. Thus, a solid unbiased procedure for volumetric BMD analysis in PMCT of the deceased presenting with intracorporeal gas is desirable.

Materials and methods

We tested three different analysis procedures (AP1-3) for BMD analysis of the lumbar vertebrae (L1-3). Data in this retrospective study was based on synchronous PMCT acquisition with a solid five-phase Cann-Genant phantom from routine forensic examinations of 154 individuals distributed into three putrefaction groups: “None” (n = 95), “Mild” (n = 54), and “Moderate” (n = 10). AP1 was based on commercially available software (“Mindways”), which required the operator to subjectively place region of interest (ROIs) in areas without gas. The open-source software (“FIJI”) was used for AP2 and AP3 and enabled comparison of objectively placed ROIs with AP1. In AP3, threshold-filtering was applied to remove the signal from gas (in AP2 data) prior to BMD analysis.

Results

AP1 provided higher BMD values than AP2-3 due to subjective placement of ROIs in denser cortical areas. AP2 yielded the lowest BMD measurements with most variation, while AP3 yielded BMD measurements comparable to in vivo values published in clinical studies. AP3 provided greater interobserver correlation.

Conclusion

AP3 provided a simple open-source software-based approach to PMCT BMD analysis that allows for precise BMD measurements in PMCT.