Analysis of trabecular bone microstructure in osteoporotic femoral heads in human patients: in vivo study using multidetector row computed tomography

This study was performed in accordance with the principles of the Declaration of Helsinki
and was approved by the ethics committee at our institution. This study was a prospective
case series that was approved by the Ethics Committee of Okanami General Hospital
and we obtained written informed consent from participants. Between April 2012 and
August 2012, we recruited ten consecutive patients who had sustained intertrochanteric
fracture and required MDCT imaging. All patients were diagnosed by plain radiography,
and MDCT was performed for preoperative evaluation of fracture type on the injured
side; we used images of the uninjured side for this study.

The patients comprised two men and eight women with a median age of 85.1 years (range,
73–96 years). Patients with previous hip fracture or surgery, osteoarthritis of the
hip, malignant tumor in any part of the body, or receiving bone modifying medication
were excluded.

Before surgery, an MDCT scan was performed with an Aquilion 64 CT scanner (Toshiba,
Tokyo, Japan) using a standard protocol (120 kV, 250 mA, collimation of 0.5 mm, and
reconstruction index of 0.3 mm) to evaluate bone quality. The scans were performed
under the following conditions: field of view of 200 mm and pixel matrix of 512?×?512.
For the morphometric analysis, specific regions of interest (ROIs) were defined within
the femoral head (Fig. 1). All ROIs were located in the center of the lateral view. On the coronal view of
the femoral head, ROI 4–6 were defined as the femoral neck area. Each area in the
neck was on the line perpendicular to the neck axis in the femoral neck isthmus. ROI
5 was in the center of the neck, ROI 4 was in the superior neck, and ROI 6 was in
the inferior neck. ROI 4 and 6 were positioned 5 mm deep to cortical bone. ROI 1–3
were defined as the femoral head apex area. ROI 1–3 were designed to be located as
extensions of ROI 4–6, parallel to the femoral neck axis. Each area in the apex was
5 mm deep to subchondral bone. We defined ROI 1 as superior, ROI 2 as central, and
ROI 3 as inferior. Each ROI had a cylindrical shape, with a diameter of 5 mm and a
depth of 10 mm.

Fig. 1. Regions of interest (ROI) in the femoral head. Line X: Femoral neck axis. Line Y:
The line perpendicular to the neck axis in the femoral neck isthmus. ROI 1–3 were
defined as the apical area. ROI 4–6 were defined as the neck area. ROI 1–3 were located
on the extension of ROI 4–6, parallel to the femoral neck axis

After MDCT, imaging data were transferred to a workstation, and the trabecular microstructure
parameters were measured using three-dimensional (3D) image analysis software (TRI/3D-BON;
RATOC System Engineering Co., Tokyo, Japan). To establish the intraobserver reliability
for measuring each parameter, two experienced orthopedic surgeons (M.M. and Y.Sa.)
inputted all ROIs manually under 3D coordinates on this software, and then each parameter
was measured automatically according to the software program. Grayscale images were
segmented using a median filter to remove noise with a fixed threshold to extract
mineralized bone components. We used a discriminant analysis method of image thresholding
based on the density histogram of a selected ROI to ensure consistent image thresholding
across all subjects studied. Isolated small particles in the marrow space and isolated
small holes in bone were removed with a cluster-labeling algorithm to remove the small
noise in the binary extraction. The measurement parameters calculated in 3D were the
bone volume fraction, which indicates bone volume/total volume (BV/TV, %), trabecular
thickness (Tb.Th, ?m) (Fig. 2), trabecular separation (Tb.Sp, ?m) (Fig. 2), and structure model index (SMI). The SMI is used to evaluate whether trabecular
bone is rod-like or plate-like; a smaller value indicates a more plate-like structure
20], 21]. It has been established that good bone quality includes a higher BV/TV, higher Tb.Th,
lower Tb.Sp, and lower SMI 13], 16].

Fig. 2. Trabecular microstructural parameters (reprinted from the literature with permission)
18]. The black arrow indicates the trabecular thickness (Tb.Th, ?m), and the white arrow indicates the trabecular separation (Tb.Sp, ?m)

Statistical analyses were performed using EZR software (Saitama Medical Center, Jichi
Medical University, Saitama, Japan), which is the graphical user interface for R (The
R Foundation for Statistical Computing, Vienna, Austria) 22]. More precisely, it is a modified version of R commander designed to add statistical
functions frequently used in biostatistics. To minimize the effect of confounders,
we standardized all data as follows: the discriminant analysis method was used to
compare each ROI, all data were divided by the average for each individual as we used
the optimal threshold value for each individual, and all data were standardized as
the corrected ratio. The data are presented as mean?±?standard deviation (SD). The
trabecular microstructure parameters among ROIs were statistically evaluated by analysis
of variance (ANOVA) and Tukey’s test. Inter- and intra-class correlation coefficients
were used to assess inter- and intra-observer reliability. The significance level
was set at p??0.05.