Steps should be taken to make sure that a DXA device is of sufficiently high quality and that its accuracy and precision are according to manufacturer guidelines. This is influenced by several parameters, such as skill and training of the technician performing the scan, site of measurement, and clinical features of patients (Damilakis and Guglielmi 2010). To determine the accuracy of a DXA densitometer, a phantom containing tissue-mimicking materials of known and different densities is scanned to ensure that measured values are truly assessed for all quantities under investigation (Diessel et al. 2000). The phantom represents the typical density range and size of normal human spine. The accuracy error of DXA bone densitometers is better than 10 % (Kanis et al. 1994). Phantom measurements have to be performed every day or at least three times a week. Results are entered into the quality control (QC) database to enable compensation of calibration changes (Guglielmi et al. 2012). If there are significant shifts or drifts in phantom data, the scanner should be serviced before scanning any patients.

One of the purposes of the DXA scan image is to see if the patient is positioned correctly, and the technologist should verify this condition before the patient leaves the testing unit. In fact, improper patient positioning and scan analysis may result in poor precision or poor reproducibility, both of which will limit the ability of the machine to measure small changes over time (Lenchik et al. 1998). Initially during DXA scan acquisition, the technologist uses the image to confirm adequate positioning and to ensure appropriate placement of ROIs, including appropriate intervertebral designations. The image should be assessed for patient motion because motion produces unpredictable alteration in bone mineral content measurements (Cawkwell 1998) (Fig. 6.5). Patients with scoliosis cannot be positioned with the spine straight on the table, and often with severe scoliosis, there are degenerative changes that invalidates the spine measurement (Fig. 6.6).

Identification of overlying hardware or retained barium may obscure the osseous structures, making BMD measurements at a site invalid (Fig. 6.7). The correct and appropriate undressing of patients before DXA scan is fundamental – in the spine, as in all other DXA examinations. Correct numbering of lumbar vertebral levels and correct ROI placement are especially important when performing serial BMD measurements to minimize scan result variability (Fig. 6.8). Staron et al. (1999) reported that incorrect placement of the intervertebral disk spaces on the image for ROI placement was the most common operator-dependent error in spine BMD measurements. The most common anatomy – five lumbar vertebrae and the last set of ribs on T12 vertebra – is found in about 85 % of people; about 15 % of the population has four or six lumbar vertebrae or the last set of ribs on T11 or L1 vertebrae. When in doubt, it is best to count from the bottom up, assuming that the L4–L5 interspace is at the pelvic brim (Fig. 6.9). There must be sufficient soft tissue on both sides of the spine; otherwise, BMD will be underestimated. All evaluable vertebrae should be used, but vertebrae that are affected by local structural change should be deleted from the analysis. Most agree that decisions can be based on two vertebrae. If all vertebrae are affected, the spine should be reported as “invalid,” with no BMD number or T-score given.

In the hip, analytic pitfalls relate to placement of femoral ROIs and particularly those influencing the neck region. The default hip analysis includes a midline that must be placed correctly for the other sites to be identified correctly. The preferred position for the rectangular femoral neck box differs for the different manufacturers. For all, the femoral neck ROI should include only the femoral neck; it should not extend medially to include part of the ischium, nor should it extend inferolaterally to include part of the trochanter (Fig. 6.10). Newer machines will automatically “ignore” the ischium within the femoral neck box, or it can be “painted out” manually, if necessary (Yu et al. 1995).

Correct identification of the patient is important for obvious reasons, but also important are the correct date of birth, sex, and ethnicity, all of which are used to calculate T- and Z-scores. In particular situations, the choice of the correct reference standard according to race and sex (e.g., transsexual patients) may be a matter of discussion and need to take into consideration the patient’s values in multiple reference systems. The use of the same machine for DXA consecutive evaluations in the follow-up of a patient is also another important point to take into consideration. It is important for patients to repeat BMD measurements on the same machine at any time in their follow-up or at least on a machine from the same manufacturer. The error between machines or in converting measurements from one manufacturer’s standard to another can introduce errors large enough to wipe out the sensitivity of the measurements.

No clear evidence shows the time lapse required between two subsequent surveys to show significant changes of the BMD values. Proper execution time of the different monitoring phases is another possible source of technical-methodological mistake. Because of limitations in the precision of testing, a minimum of 2 years may be needed to reliably measure a chnge in BMD (U.S. Preventive Services Task Force Recommendation Statement 2011).

As previously mentioned, to further improve the accuracy and to optimize the use of bone densitometry, a team of expert radiologists and technologists should be constituted. Some states also require specific training in bone densitometry. Hence, radiologists should be fully experienced in both physiopathology and treatment of osteoporosis and bone metabolic diseases and performance of all radiologic techniques employed for the diagnosis of the disease and bone densitometry diagnostic activity. On the other hand, technologists should have sufficient knowledge and experience of the bone densitometry equipments such as patient positioning and performance of quality control procedures (Khan et al. 2004).

Primary care physicians are usually not directly involved in the performance and interpretation of DXA scans but should be familiar with the information on DXA scan reports and how it applies to patient management (Watts 2004). Mistakes in interpretation and reporting are the most difficult to correct, because they occur in the final steps of the process, usually with no one to identify the errors before an incorrect report is sent. It is logical that such errors can be minimized by initial training and continuing education in the field of bone densitometry (Lewiecki et al. 2006). The DXA image is assessed for visible disease processes. In addition to noting the presence of these abnormal conditions in the DXA report, it is important to determine how these processes affect the BMD measurement.

Certain conditions can artificially elevate the BMD, obscuring osteoporosis and leading to underestimation of fracture risk (Barden et al. 2003). These include degenerative disk disease, compression fractures (Jacobson et al. 2000) (Fig. 6.11), scoliosis, postsurgical defects, overlying atherosclerotic calcifications, implantable devices such as stents and vena cava filters, soft tissue calcifications (Fig. 6.12), retained radiopaque contrast agents, lumbar hardware, vertebroplasty cement, and external objects such as piercings, bra clips, and high-density objects on clothing (Theodorou and Theodorou 2002; Bazzocchi et al. 2012).

In the lumbar spine, almost all artifacts and local structural change will spuriously elevate BMD. This is especially true for spinal degenerative change (spondylosis), which can elevate spine BMD by 2, 3, or more T-score units (Yu et al. 1995). The affected vertebral levels should be excluded from the ROI, or another BMD measurement site should be selected. Lateral DXA may be used to reduce the effect of sclerotic facet osteoarthrosis (Fig. 6.13) or overlying vascular calcifications on the BMD measurement because this area would be excluded from the ROI (Fig. 6.14). In the spine, absent bone (after surgery such as laminectomy or spina bifida) will spuriously lower the BMD. In the hip, sclerotic conditions such as bone islands (enostosis) may increase the BMD measurement, if they are included in the ROI. Significant osteoarthrosis may increase the BMD measurement, if cortical thickening or buttressing extends into the femoral neck ROI. Similarly, Paget disease, calcific tendinitis (calcium hydroxyapatite deposition), and vascular calcifications may increase the BMD measurement, if included in the ROI. If a disease process is affecting the BMD measurement in a particular ROI, another ROI or, alternatively, another site may be used. Similarly, BMD at the proximal femur can be altered by the degree of internal rotation, placement of soft tissue gluteal silicon implants (Hauache et al. 2000) (Fig. 6.15), and overlying soft tissue thickness or fat.

Other disease conditions such as avascular necrosis of the femoral head and developmental dysplasia of the hip may be visible on the DXA image but will not affect the BMD measurement because these conditions are outside the ROI boundaries. It is essential, however, that a suspected abnormality on the DXA image be evaluated further with correlative imaging to confirm a diagnosis and to exclude potential aggressive or malignant causes (Jacobson et al. 2000). It is important to suggest that all alternative or corrective methods and changes in technical acquisition as well as considerations related to particular elements (e.g., excluded sites or limitations) should be described in the reports.

Today, quantitative ultrasound (QUS) represents an effective alternative or integrative method for studying bone tissue and managing patients presenting with osteoporosis or other metabolic bone diseases (Guglielmi et al. 2011). Although QUS does not produce an image showing bone structure, it is used to measure quantitative parameters and assess tissue properties. Interest in quantitative US methods can be attributed primarily to the fact that they involve no radiation exposure. In addition, US-dedicated scanners offer the advantages of small size, quick and simple measurements, and low cost compared with axial DXA or quantitative CT scanners. These advantages explain the motivation for using QUS for the assessment of osteoporosis, assuming that performance is comparable. The primary disadvantage of QUS is its lack of sensitivity (Moyad 2003), making it inappropriate for long-term monitoring of osteoporosis or response to drug therapy. In addition to equipment drift, in vivo precision is also affected by external factors (Njeh et al. 2000). Currently, US assessment is used as a screening tool, with confirmation of diagnosis by means of DXA evaluation.

QUS devices generate pulsed acoustic waves with a frequency between 500 kHz and 1.25 MHz according to the manufacturer, which is considerably lower than the frequencies commonly used in US imaging. US waves are affected not only by the amount of material but also by its elasticity and structure. As a consequence, QUS may yield additional information about bone properties and fragility. The QUS method involves generating ultrasound impulses that are transmitted transversally or longitudinally through the bone being studied. The ultrasound wave is produced in the form of a sinusoid impulse by special piezoelectric probes and is detected once it has passed through the medium. There are two distinct probes (emitting and receiving) and the skeletal segment for evaluation is placed between them. The first ultrasound parameters employed for characterizing the bone tissue are speed of sound (SoS) and broadband ultrasound attenuation (BUA). More complex parameters have been developed from combination of the aforementioned amplitude dependent speed of sound (AD-SoS), stiffness, and quantitative ultrasound index (QUI) (Gluer 1997). In the diagnosis of osteoporosis, these latter parameters have proved to be more useful in identifying subjects with low bone mineral density and therefore at high risk for fracture (Hans et al. 1996; Guglielmi et al. 2003).