Performance of Progressive Generations of GPT on an Exam Designed for Certifying Physicians as Certified Clinical Densitometrists.

BACKGROUND: Artificial intelligence (AI) large language models (LLMs) such as ChatGPT have demonstrated the ability to pass standardized exams. These models are not trained for a specific task, but instead trained to predict sequences of text from large corpora of documents sourced from the internet. It has been shown that even models trained on this general task can pass exams in a variety of domain-specific fields, including the United States Medical Licensing Examination. We asked if large language models would perform as well on a much narrower subdomain tests designed for medical specialists. Furthermore, we wanted to better understand how progressive generations of GPT (generative pre-trained transformer) models may be evolving in the completeness and sophistication of their responses even while generational training remains general. In this study, we evaluated the performance of two versions of GPT (GPT 3 and 4) on their ability to pass the certification exam given to physicians to work as osteoporosis specialists and become a certified clinical densitometrists. The CCD exam has a possible score range of 150 to 400. To pass, you need a score of 300. METHODS: A 100-question multiple-choice practice exam was obtained from a 3rd party exam preparation website that mimics the accredited certification tests given by the ISCD (International Society for Clinical Densitometry). The exam was administered to two versions of GPT, the free version (GPT Playground) and ChatGPT+, which are based on GPT-3 and GPT-4, respectively (OpenAI, San Francisco, CA). The systems were prompted with the exam questions verbatim. If the response was purely textual and did not specify which of the multiple-choice answers to select, the authors matched the text to the closest answer. Each exam was graded and an estimated ISCD score was provided from the exam website. In addition, each response was evaluated by a rheumatologist CCD and ranked for accuracy using a 5-level scale. The two GPT versions were compared in terms of response accuracy and length. RESULTS: The average response length was 11.6 ±19 words for GPT-3 and 50.0±43.6 words for GPT-4. GPT-3 answered 62 questions correctly resulting in a failing ISCD score of 289. However, GPT-4 answered 82 questions correctly with a passing score of 342. GPT-3 scored highest on the "Overview of Low Bone Mass and Osteoporosis" category (72 % correct) while GPT-4 scored well above 80 % accuracy on all categories except "Imaging Technology in Bone Health" (65 % correct). Regarding subjective accuracy, GPT-3 answered 23 questions with nonsensical or totally wrong responses while GPT-4 had no responses in that category. CONCLUSION: If this had been an actual certification exam, GPT-4 would now have a CCD suffix to its name even after being trained using general internet knowledge. Clearly, more goes into physician training than can be captured in this exam. However, GPT algorithms may prove to be valuable physician aids in the diagnoses and monitoring of osteoporosis and other diseases.

An update on methods for assessing bone quality and health in Cystic fibrosis.

With increasing life expectancy in people with Cystic fibrosis (CF), the focus of clinical care has shifted to management and prevention of non-pulmonary comorbidities. CF related bone disease, defined by low bone mineral density (BMD), is prevalent across all age groups and acknowledges the increased fractures rates that negatively impact lung function and quality of life. Dual energy X-ray absorptiometry (DXA) measurement of bone mineral content (BMC) and "areal" BMD (aBMD) is recommended for identifying and monitoring bone health in children and adults due to its low cost, low radiation exposure, and widespread availability. Recent studies in children and adolescents with chronic illness focus on adjustment of BMC and aBMD measurements for height due to the effects of short stature and delayed maturation on bone size. Expanded reference databases for alternate imaging sites such as the ultradistal radius and hip present opportunities for research and long-term monitoring. As the two-dimensional nature of DXA imposes limitations, we highlight other imaging modalities including peripheral quantitative computed tomography QCT (pQCT), magnetic resonance imaging, and quantitative ultrasound (QUS). These tools, while primarily used in a research setting, can impart information on true volumetric bone density and bone microarchitecture as well as contribute to fracture assessment and prediction. Due to the high morbidity and mortality associated with vertebral and hip fracture, we will present on vertebral fracture assessment (VFA) in both children and adults as well as applied analyses including hip structural analysis (HSA), trabecular bone score (TBS), and fracture risk assessment (FRAX) for high risk groups. Questions remain on the future clinical applicability and accessibility of these assessment and prediction tools, longitudinal monitoring through adolescence and adulthood, and how outcome measures may guide bone modifying therapies.

The differences in the relationship between obstructive sleep apnea severity and trabecular bone score in men and women with type 2 diabetes.

AIMS: Type 2 diabetes mellitus (T2DM) and obstructive sleep apnea (OSA) may adversely affect bone. Gender is a well-established factor influencing bone health. We investigated the impact of OSA on bone mineral density (BMD) and trabecular bone score (TBS) in T2DM. METHODS: Eighty-one T2DM patients [33 men and 48 women] participated. OSA was diagnosed using an overnight monitor, with its severity assessed by an apnea hypopnia index (pAHI). The measurements of hypoxia, including the percentage of total sleep time in which oxygen saturation remains below 90% (pT90), the oxygen desaturation index (pODI) and minimum O2 (min O2), were reported. Lumbar spine (L1-4) and femoral neck (FN) BMD were measured using dual-energy X-ray absorptiometry (DXA). TBS was computed from DXA images. RESULTS: Sixty-five patients (80.2%) had OSA. pAHI, pT90, pODI and min O2 were not correlated to L1-4 BMD, FN BMD or TBS in all participants by multiple regression analyses adjusting for age, gender and BMI. However, an interaction between gender and pAHI, and gender and pODI were significantly associated with TBS (b = 0.003, p = 0.034 and b = 0.004, p = 0.046, respectively). We therefore reassessed an association between pAHI or pODI and TBS separately between men and women. After adjusting for age and BMI, more severe OSA (higher pAHI) and higher pODI significantly associated with lower TBS (b = -0.002, p = 0.034 and b = -0.003, p = 0.021, respectively) in men. On the other hand, higher pAHI non-significantly associated with better trabecular microarchitecture as indicated by higher TBS (b = 0.002, p = 0.059) in women. When considered only postmenopausal (n = 33), higher pAHI and higher pODI were significantly associated with higher TBS (b = 0.004, p = 0.003 and b = 0.004, p = 0.008, respectively). CONCLUSIONS: In T2DM patients, there is a complex interrelationship among OSA severity, gender and TBS. More severe OSA predicted lower TBS in men, but predicted higher TBS in postmenopausal women.

Evaluation of bone mineralization in former preterm born children: Phalangeal quantitative ultrasound cannot replace dual-energy X-ray absorptiometry.

BACKGROUND: Preterm infants are at risk of impaired bone health in later life. Dual-energy X-ray absorptiometry-scan (DXA) is the gold standard to determine bone mineralization. Phalangeal quantitative ultrasound (pQUS) is an alternative technique that is inexpensive, easy to use and radiation-free. The aim of this study was to investigate whether both techniques reveal equivalent results. MATERIALS AND METHODS: Sixty former preterm infants (31 boys; 29 girls) received a DXA and pQUS at age 9 to 10 years. DXA measured bone mineral content (BMC) and bone mineral density (BMD) for total body and lumbar spine (L1-4), while pQUS measured the amplitude dependent speed of sound (AD-SoS) and bone transit time (BTT) at metacarpals II-IV providing continuous values and Z-scores based on age and sex. Four statistical methods evaluated the association between both techniques: Pearson's correlation coefficients, partial correlation coefficients adjusted for gestational age, height and BMI, Bland-Altman analysis and cross tabulation. RESULTS: Both techniques showed a statistically significant weak correlation for continuous values as well as Z-scores (0.291-0.462, p < 0.05). Boys had significant and relatively high correlations (0.468-0.585, p < 0.05). In comparison, the correlations for girls were not significant. Correlation coefficients further decreased while calculating the partial correlations. The Bland-Altman plots showed poor agreement. Sensitivity ranged from 33% to 92% and specificity from 16% to 68%. Positive and negative predictive values ranged from 4% to 38% and 82% to 97%, respectively. CONCLUSIONS: We found statistically significant weak correlations and poor agreement between DXA and pQUS measurements. DXA is not equivalent to pQUS and therefore not replaceable by this technique in former preterm born children at the age of 9 to 10 years.

The history of bone densitometry.

Bone densitometry (dual energy x-ray absorptiometry-DXA) is a vital medical tool needed for the diagnosis of osteoporosis in non-fractured patients; predicting future fracture risk; and monitoring bone mineral density (BMD) in untreated or treated patients. The history of the pivotal international society involved in the science and clinical interpretation of DXA, the International Society for Clinical Densitometry (ISCD) is defined in this manuscript. Since DXA and Osteoporosis management are intimately linked, the ISCD has over the years developed strong bonds with both the National Osteoporosis Foundation (NOF) and the International Osteoporosis Foundation (IOF). The positive impact that ISCD has led in the proper performance and clinical interpretation of bone mass measurements has been enormous worldwide.