Ellen B. Fung - Bone Health Assessment in Pediatrics: Guidelines for Clinical Practice

Accuracy and precision are easily understood using a target analogy shown in Fig.. Here all five arrows are grouped together within the bulls-eye, indicating a high degree of both accuracy and precision.

When bone densitometry is used to quantify the bone density for the purpose of diagnosing osteoporosis or predicting fracture risk, it is imperative that the measurement be accurate. On the other hand, when bone densitometry is used to follow changes in bone density over time, precision becomes paramount. Strictly speaking, the initial accuracy of the measurement is no longer of major concern. It is only necessary that the measurement be reproducible or precise since it is the change between measurements that is of interest. Bone densitometry has the potential to be the most precise quantitative measurement technique in clinical medicine. The precision that is actually obtained however is highly dependent upon the skills of the technologist. Precision itself can be quantified in a precision study, as discussed in , it is preferable to use the root-mean-square standard deviation (RMS-SD) or root-mean-square coefficient of variation (RMS-CV) to express precision rather than the arithmetic mean or average standard deviation (SD) or coefficient of variation (CV). It is not always clear whether the manufacturers precision value is being expressed as the RMS or arithmetic average. In general, the arithmetic mean SD or CV will be better than the RMS-SD or RMS-CV. Manufacturers also do not usually state the average bone density of the population in the precision study or the exact number of people and number of scans per person, making the comparison of such values with values obtained at clinical facilities difficult.

The distinction between weight bearing and non-weight bearing is reasonably intuitive. The lower extremities are weight bearing as is the cervical, thoracic, and lumbar spine. Often forgotten, although it is the most sensitive weight-bearing bone, is the calcaneus or os calcis. Portions of the pelvis are considered weight bearing as well. The remainder of the skeleton is considered non-weight bearing.

The axial skeleton includes the skull, ribs, sternum, and spine as shown in Fig.. The scapulae and the pelvis are therefore part of the appendicular skeleton. The proximal femur is also obviously part of the appendicular skeleton, although it is often mistakenly included in the axial skeleton. Contributing to this confusion is the current practice of including dual-energy X-ray bone density studies of the proximal femur under the CPT code 77080 used for DXA spine bone density studies.

The characterization of skeletal sites as either central or peripheral is unique to densitometry. Central sites are the thoracic and lumbar spine in either the PA or lateral projection and the proximal femur. By extension, those densitometers that have the capability of measuring the spine and proximal femur are called central densitometers. As a matter of convention, this designation is generally not applied to QCT (quantitative Computed Tomography), even though spine bone density measurements are made with QCT. Peripheral sites are the commonly measured distal appendicular sites such as the calcaneus, tibia, metacarpals, phalanges, and forearm. Again, by extension, densitometers that measure only these sites are called peripheral densitometers. Some central devices also have the capability of measuring peripheral sites. Nevertheless, they retain their designation as central bone densitometers. The central and peripheral skeleton is illustrated in Fig..

The characterization of a site as predominantly cortical or trabecular bone is important in densitometry. Some disease states show a predilection for one type of bone over the other, making this an important consideration in the selection of the site to measure when a particular disease is present or suspected. Similarly, the response to certain therapies is greater at sites that are predominantly trabecular because of the greater metabolic rate of trabecular bone. There are also circumstances in which a physician desires to assess the bone density at both a predominantly cortical and predominantly trabecular site in order to have a more complete evaluation of a patients bone mineral status.