Common errors in dual-energy x-ray absorptiometry (DXA) examinations: insights from a cross-sectional study in Northeastern Brazil

Errores comunes en las exploraciones de absorciometría dual de rayos X (DXA): hallazgos de un estudio transversal en el noreste de Brasil

Luiz Henrique Rosendo

Universidade Federal de Campina Grande, Campina Grande , Brasil

Lucas Silva lucas.silva@famed.ufal.br

Universidade Federal de Alagoas, Maceió , Brasil

Cátia Gondim

Universidade Federal de Campina Grande, Campina Grande , Brasil

Osteoporosis is a metabolic bone disease characterized by decreased bone strength, predisposing individuals to fractures, even from low-impact events such as falls from standing height [1]. The most common fracture sites include the distal radius, vertebrae, and proximal femur, with particularly severe consequences among older adults due to the associated morbidity and mortality [2]. Early detection and effective management are critical to mitigating the disease’s clinical burden.

Dual-energy X-ray absorptiometry (DXA) is the gold standard for diagnosing osteoporosis, providing quantitative measures such as bone mineral density (BMD), T-scores, and Z-scores, which guide clinical decisions [3]. However, DXA results can be significantly compromised by technical errors during image acquisition and processing [4]. Improper patient positioning, failure to exclude bone artifacts (e.g., osteophytes or prosthetic implants), and incorrect vertebral selection can result in misinterpretation of BMD values, leading to either overestimation or underestimation of bone loss [5].

These technical errors are often linked to the operator-dependent nature of DXA and highlight the importance of consistent image acquisition protocols and operator training [6]. Incorrect diagnoses may delay treatment initiation or lead to unnecessary therapeutic interventions, potentially increasing healthcare costs and patient burden [7]. Furthermore, DXA plays a crucial role in monitoring treatment responses, underscoring the need for precision at the initial diagnosis.

This study aims to evaluate the profile and prevalence of technical errors in DXA exams conducted at a referral hospital. By identifying common pitfalls, we aim to propose recommendations for minimizing diagnostic errors and optimizing the clinical management of osteoporosis.

The demographic profile of the patients in this study aligns with known osteoporosis risk factors, reinforcing the need for precise DXA assessments. The mean age of 65.6 years and the predominance of female patients reflect the well-established higher prevalence of osteoporosis in postmenopausal women due to hormonal changes affecting bone density [9]. Additionally, the mean BMI of 26.4 kg/m² suggests that many patients fall within the overweight range, which can influence DXA accuracy, as higher soft tissue composition may lead to overestimated bone mineral density values [10]. Ethnicity is another critical variable, as 91% of the sample identified as white, a group with generally lower bone mass compared to individuals of African descent, who tend to have higher peak bone density [11].

Additionally, this work still reveals a high prevalence of technical errors in DXA exams, with significant implications for the diagnosis and management of osteoporosis. The most frequent error observed was the presence of osteophytes, affecting 64% of the exams. Osteophytes can falsely elevate BMD readings, leading to the underdiagnosis of osteoporosis and inappropriate clinical management [7]. Studies have shown that undetected osteophytes can obscure the severity of bone loss, potentially delaying the initiation of preventive therapies for fractures [12].

Another frequent issue was inadequate femoral rotation, present in 45% of cases, which led to the inclusion of the lesser trochanter in the image, thereby falsely elevating BMD measurements [13]. Errors in the selection and exclusion of vertebrae were identified in 24% of cases. Improper exclusion of abnormal vertebrae, as highlighted in the literature, can result in inaccurate diagnoses of osteopenia or osteoporosis, depending on the affected vertebrae [13, 14]. Additionally, incorrect vertebral cataloging, observed in 16% of exams, can further distort diagnostic outcomes, particularly when L1 is misidentified or confused with T12 [15].

The presence of scoliosis, noted in 20% of cases, represents another critical diagnostic challenge. Scoliosis alters spinal curvature and bone density measurements, increasing the likelihood of false-negative or false-positive results, as previously documented in clinical studies [16, 17]. Furthermore, incorrect ROI selection, affecting 20% of the exams, underscores the importance of proper operator training and adherence to standardized imaging protocols [18].

Comparisons with international studies provide valuable context. For instance, the error rate in vertebral cataloging in this study (16%) was lower than the 38% reported in Ecuador [19] but similar to findings from Turkey, where error rates ranged from 10.5% to 65.5% depending on the institution [20]. This variability highlights the critical role of institutional protocols and operator experience in minimizing technical errors [6].

Errors in DXA scans can lead to significant misclassification, particularly underestimating osteoporosis diagnoses. In the analyzed data, exams with technical errors showed an increased rate of normal diagnoses and reduced rates of osteopenia and osteoporosis, highlighting the risk of false-negative results. Similar findings have been reported, with studies indicating that technical errors in DXA occur frequently and can result in misdiagnoses, potentially delaying appropriate treatment [21]. These inaccuracies underscore the importance of rigorous quality control and standardized protocols to ensure accurate bone density assessment, in this context secondary reviews improve diagnostic sensitivity by addressing errors in image acquisition and interpretation [17].

This study is limited by its non-probabilistic sampling, which restricts the generalization of findings. Additionally, the single-researcher analysis may have introduced observer bias, despite adherence to protocols. Future research should involve larger, multicenter samples and automated quality control to minimize bias and improve reliability.

The authors have no conflict of interest to declare. The authors declared that this study has received no financial support.

redalyc-journal-id: 6920