Identification Of Vertebral Fractures Research Articles

Deep learning-based automated high-accuracy location and identification of fresh vertebral compression fractures from spinal radiographs: a multicenter cohort study

BackgroundDigital radiography (DR) is a common and widely available examination. However, spinal DR cannot detect bone marrow edema, therefore, determining vertebral compression fractures (VCFs), especially fresh VCFs, remains challenging for clinicians.MethodsWe trained, validated, and externally tested the deep residual network (DRN) model that automated the detection and identification of fresh VCFs from spinal DR images. A total of 1,747 participants from five institutions were enrolled in this study and divided into the training cohort, validation cohort and external test cohorts (YHDH and BMUH cohorts). We evaluated the performance of DRN model based on the area under the receiver operating characteristic curve (AUC), feature attention maps, sensitivity, specificity, and accuracy. We compared it with five other deep learning models and validated and tested the model internally and externally and explored whether it remains highly accurate for an external test cohort. In addition, the influence of old VCFs on the performance of the DRN model was assessed.ResultsThe AUC was 0.99, 0.89, and 0.88 in the validation, YHDH, and BMUH cohorts, respectively, for the DRN model for detecting and discriminating fresh VCFs. The accuracies were 81.45% and 72.90%, sensitivities were 84.75% and 91.43%, and specificities were 80.25% and 63.89% in the YHDH and BMUH cohorts, respectively. The DRN model generated correct activation on the fresh VCFs and accurate peak responses on the area of the target vertebral body parts and demonstrated better feature representation learning and classification performance. The AUC was 0.90 (95% confidence interval [CI] 0.84–0.95) and 0.84 (95% CI 0.72–0.93) in the non-old VCFs and old VCFs groups, respectively, in the YHDH cohort (p = 0.067). The AUC was 0.89 (95% CI 0.84–0.94) and 0.85 (95% CI 0.72–0.95) in the non-old VCFs and old VCFs groups, respectively, in the BMUH cohort (p = 0.051).ConclusionIn present study, we developed the DRN model for automated diagnosis and identification of fresh VCFs from spinal DR images. The DRN model can provide interpretable attention maps to support the excellent prediction results, which is the key that most clinicians care about when using the model to assist decision-making.

Opportunistic Identification of Vertebral Compression Fractures on CT Scans of the Chest and Abdomen, Using an AI Algorithm, in a Real-Life Setting.

This study evaluated the performance of a vertebral fracture detection algorithm (HealthVCF) in a real-life setting and assessed the impact on treatment and diagnostic workflow. HealthVCF was used to identify moderate and severe vertebral compression fractures (VCF) at a Danish hospital. Around 10,000 CT scans were processed by the HealthVCF and CT scans positive for VCF formed both the baseline and 6-months follow-up cohort. To determine performance of the algorithm 1000 CT scans were evaluated by specialized radiographers to determine performance of the algorithm. Sensitivity was 0.68 (CI 0.581-0.776) and specificity 0.91 (CI 0.89-0.928). At 6-months follow-up, 18% of the 538 patients in the retrospective cohort were dead, 78 patients had been referred for a DXA scan, while 25 patients had been diagnosed with osteoporosis. A higher mortality rate was seen in patients not known with osteoporosis at baseline compared to patients known with osteoporosis at baseline, 12.8% versus 22.6% (p = 0.003). Patients receiving bisphosphonates had a lower mortality rate (9.6%) compared to the rest of the population (20.9%) (p = 0.003). HealthVCF demonstrated a poorer performance than expected, and the tested version is not generalizable to the Danish population. Based on its specificity, the HealthVCF can be used as a tool to prioritize resources in opportunistic identification of VCF's. Implementing such a tool on its own only resulted in a small number of new diagnoses of osteoporosis and referrals to DXA scans during a 6-month follow-up period. To increase efficiency, the HealthVCF should be integrated with Fracture Liaison Services (FLS).

Development of a shape-based algorithm for identification of asymptomatic vertebral compression fractures: A proof-of-principle study

ObjectivesVertebral fracture is both common and serious among adults, yet it often goes undiagnosed. This study aimed to develop a shape-based algorithm (SBA) for the automatic identification of vertebral fractures. MethodsThe study included 144 participants (50 individuals with a fracture and 94 without a fracture) whose plain thoracolumbar spine X-rays were taken. Clinical diagnosis of vertebral fracture (grade 0 to 3) was made by rheumatologists using Genant's semiquantitative method. The SBA algorithm was developed to determine the ratio of vertebral body height loss. Based on the ratio, SBA classifies a vertebra into 4 classes: 0 = normal, 1 = mild fracture, 2 = moderate fracture, 3 = severe fracture). The concordance between clinical diagnosis and SBA-based classification was assessed at both person and vertebra levels. ResultsAt the person level, the SBA achieved a sensitivity of 100% and specificity of 62% (95% CI, 51%–72%). At the vertebra level, the SBA achieved a sensitivity of 84% (95% CI, 72%–93%), and a specificity of 88% (95% CI, 85%–90%). On average, the SBA took 0.3 s to assess each X-ray. ConclusionsThe SBA developed here is a fast and efficient tool that can be used to systematically screen for asymptomatic vertebral fractures and reduce the workload of healthcare professionals.

Machine Learning and Deep Learning in Spinal Injury: A Narrative Review of Algorithms in Diagnosis and Prognosis

Spinal injuries, including cervical and thoracolumbar fractures, continue to be a major public health concern. Recent advancements in machine learning and deep learning technologies offer exciting prospects for improving both diagnostic and prognostic approaches in spinal injury care. This narrative review systematically explores the practical utility of these computational methods, with a focus on their application in imaging techniques such as computed tomography (CT) and magnetic resonance imaging (MRI), as well as in structured clinical data. Of the 39 studies included, 34 were focused on diagnostic applications, chiefly using deep learning to carry out tasks like vertebral fracture identification, differentiation between benign and malignant fractures, and AO fracture classification. The remaining five were prognostic, using machine learning to analyze parameters for predicting outcomes such as vertebral collapse and future fracture risk. This review highlights the potential benefit of machine learning and deep learning in spinal injury care, especially their roles in enhancing diagnostic capabilities, detailed fracture characterization, risk assessments, and individualized treatment planning.

Reporting of paediatric osteoporotic vertebral fractures in Duchenne muscular dystrophy and potential impact on clinical management: the need for standardised and structured reporting

BackgroundIn boys with Duchenne muscular dystrophy (DMD), initiation of bisphosphonate is recommended upon identification of moderate or severe vertebral fractures, even if asymptomatic. Clear radiological reporting is important for consistency of clinical interpretation and management.ObjectivesTo audit radiology reports of spine imaging for vertebral fracture assessment in DMD, and assess potential impact on diagnosis and management.Materials and MethodsLateral thoracolumbar spine imaging (71 lateral spine radiographs and 13 lateral dual energy absorptiometry spine image) in 84 boys with DMD performed across two centres. Anonymised radiology reports by paediatric radiologists were circulated to two neuromuscular clinicians and two endocrinologists. Clinicians determined if there was vertebral fracture, no vertebral fracture, or unclear interpretation. Endocrinologists also determined if bisphosphonate was indicated. A single observer (a clinician with expertise in vertebral fracture assessment) performed vertebral fracture assessment in 37 images and re-reported using a structured format. Structured reports were re-circulated to the four clinicians to re-evaluate the degree of concordance in clinical diagnosis of vertebral fracture and treatment decisions with bisphosphonate.ResultsThe term “fracture” was used in 25/84 (30%) radiology reports and only in 8/43 (19%) with description of vertebral body abnormalities. Fracture grading was included in 7/43 (16%) radiology reports. Diagnostic concordance by the clinicians was noted in 36/84 (43%). Unclear interpretation was noted in 22% to 51% based on radiology reports. No unclear interpretation was noted with structured reports. Complete diagnostic (37/37, 100%) and treatment (37/37, 100%) concordance was noted with the structured reports, whereas complete diagnostic and treatment concordance was noted in only 16/37 (43%) and 17/37 (46%) of the radiology reports, respectively.ConclusionOnly a third of radiology reports of spine imaging in DMD explicitly used the terminology “fracture”. Grading was only noted in a small percentage. Variability in diagnostic interpretation by clinicians may lead to differing management plans. As identification of vertebral fracture is a trigger for treatment, developing reporting guidelines for paediatric vertebral fracture assessment will improve care. A structured template should be introduced for radiological reporting of paediatric vertebral fracture assessment.

THU433 Spontaneous Vertebral Fractures In The Setting Of Prolonged CRRT

Abstract Disclosure: F. Akanbi: None. M. Kutahyalioglu: None. P. Gogineni: None. S. Raja: None. Introduction: Bone resorption and hypercalcemia due to prolonged immobilization are known complications in critically ill patients. This is more pronounced in patients with high bone turnover rates such as in secondary hyperparathyroidism. In this case report, the chelation of calcium by citrate masked the immobilization hypercalcemia, resulting in marked bone loss and spontaneous vertebral fractures in the context of normal calcium levels. Case report: 23 year old man with no past medical history was seen on account of concerns for calcium loss from his bones as he was on CRRT, for 4 months and the amount of calcium in the extracorporeal circuit was more than expected. He had presented with new onset systolic heart failure (EF 10%) secondary to fulminant post-COVID-19 myocarditis, and renal failure due to acute tubular necrosis. CRRT with citrate dextrose anticoagulation was initiated at presentation. He had received pulse dose prednisone for presumed post-viral myocarditis. He was awaiting heart and kidney transplantation. Greater than average calcium clearance during CRRT prompted workup for hypercalcemia. He denied personal or family history of hypercalcemia, renal stones, fractures, dental issues or hyperparathyroidism. He reported eating adequate amounts of dairy products. Labs showed corrected calcium 9.3 (8.6-10.3 mg/dl), elevated PTH 189 (10-65 pg/ml), low serum protein and albumin, and low 25OHVD that was 14 (25-100 ng/mL) on presentation and actively being replaced with ergocalciferol. Phosphorus, magnesium, liver and thyroid function were normal. The patient had complained of back pain after being immobile for 3 months. MRI showed subacute vertebral fractures of T12 and L1. Thereafter, spine X-ray showed multiple vertebral fractures of L2-L5. Beta CTX was 4153 (120-946 pg/ml) and DXA scan revealed low bone density for age with Z score -3.7 in the femoral neck, invalid in the spine due to the fractures, felt to be the source of high extracorporeal calcium concentration. A dose of Denosumab was given for treatment with improvement of calcium clearance during CRRT. Discussion: Acute fractures and infrequently, prolonged immobilization cause hypercalcemia with suppression of PTH. In our patient who was critically ill and immobilized for 4 months, the use of CRRT maintained normocalcemic blood levels, and renal failure with vitamin D deficiency caused secondary hyperparathyroidism. Only the declining requirements for calcium replacement during CRRT prompted workup for hypercalcemia that lead to identification of significant vertebral fractures and treatment. Declining need for calcium replacement during CRRT should prompt investigation and treatment of disorders of hypercalcemia and metabolic bone disorders in this patient population. Presentation: Thursday, June 15, 2023

241 ‘Please consider bone assessment’: improving the identification and management of Vertebral Fractures by a ‘Fracture Liaison Service’

Abstract Background Osteoporotic fractures are a major public health problem due to an ageing population causing profound disability and reduced quality of life. Vertebral Fractures (VFs) are the second most common osteoporotic fracture with nearly 1 in 4 women having one in their lifetime. Furthermore, a VF is a significant warning sign for another fracture in the following year—known as ‘imminent fracture risk’. Despite this early opportunity for clinical intervention with bone protective medication, VFs are frequently under-diagnosed and under-treated. This observational study aimed to look at the point prevalence of VFs picked up on radiology reports and compare with those identified by the ‘Fracture Liaison Service’ (FLS). Methods The PACs office ran a search for all MRI spine, CT-TAP and a plain radiographic film reports from January 2021 until April 2023 with the terms ‘vertebrae’ and ‘compression’ in the text. Inclusion criteria was all patients >65 without major trauma with a radiologist defined osteoporotic VF. Using excel, we cross-referenced the proportion of these already know to the FLS. Results There were 338 osteoporotic VFs diagnosed on radiological imaging between 2021 & 2023. Of these, 101 had been identified by the FLS. The median age was 77 years old. There were 228 (67.2%) females and 111 (32.8%) males. Conclusion This study helped to highlight that over 67.2% of the osteoporotic VFs are not being identified by the FLS. Therefore, is it assumed that timely intervention and management of their imminent fracture risk may not have been undertaken. This data is being used to drive a quality improvement project with the radiology department where any VF in women >70 will be reported to include—‘please consider bone assessment’.

External validation of a convolutional neural network algorithm for opportunistically detecting vertebral fractures in routine CT scans.

To evaluate the performance of a previously trained Convolutional Neural Network (CNN) model to automatically detect vertebral fractures (VFs) in CT scans in an external validation cohort. Two Chinese studies and clinical data were used to retrospectively select CT scans of the chest, abdomen and thoracolumbar spine in men and women aged ≥50 years. The CT scans were assessed using the semiquantitative (SQ) Genant classification for prevalent VFs in a process blinded to clinical information. The performance of the CNN model was evaluated against reference standard readings by the area under the receiver operating characteristics curve (AUROC), accuracy, Cohen's kappa, sensitivity, and specificity. A total of 4,810 subjects were included, with a median age of 62 years (IQR 56-67), of which 2,654 (55.2%) were females. The scans were acquired between January 2013 and January 2019 on 16 different CT scanners from three different manufacturers. 2,773 (57.7%) were abdominal CTs. A total of 628 scans (13.1%) had ≥1 VF (grade 2-3), representing 899 fractured vertebrae out of a total of 48,584 (1.9%) visualized vertebral bodies. The CNN's performance in identifying scans with ≥1 moderate or severe fractures achieved an AUROC of 0.94 (95% CI: 0.93-0.95), accuracy of 93% (95% CI: 93%-94%), kappa of 0.75 (95% CI: 0.72-0.77), a sensitivity of 94% (95% CI: 92-96%) and a specificity of 93% (95% CI: 93-94%). The algorithm demonstrated excellent performance in the identification of vertebral fractures in a cohort of chest and abdominal CT scans of Chinese patients ≥50 years.

Effectiveness of fracture liaison services in osteoporosis.

In response to the gradual decline in the number of prescriptions for anti-osteoporosis medication (AOM) following fragility fractures, fracture liaison services (FLSs) have been set up around the world with the aim of filling this treatment gap. Several studies have already reported the benefits of such organizations, particularly in reducing fracture risk, mortality rates and healthcare costs, and literature on FLSs has increased at a steady pace over time. A narrative review was conducted on the latest available findings on the effectiveness of FLSs. Various approaches to implementing an effective FLS program are discussed. FLS programs have enhanced the management of osteoporosis-related fractures. However, several studies have highlighted that not all FLSs are necessarily effective in reducing subsequent fracture risk and mortality. Long-term AOM persistence and monitoring are another critical issue in FLS programs. A few studies have reported that FLSs are associated with an improvement in AOM persistence, regardless of the type of AOM. Practitioners in the FLS setting need to be aware of the impact of recency of fracture and fracture recurrence rates, and the need for timely interventions. The administration of zoledronic acid in an in-patient setting may improve AOM treatment rates in patients, who often encounter obstacles to outpatient follow-up. Introducing 'vertebral fracture identification services' in FLS programs is also an option. However, doing so leads to an increase in workload and this would need to be considered by any FLS that is considering introducing such a service. Evidence suggests that digital technologies can support (i) multidisciplinary teams in providing the best possible patient care based on current evidence, and (ii) patient self-management. However, as the methodological quality of many of the studies evaluating these technologies was poor, their validity of their results is limited. Further research should focus on the optimal implementation of post-fracture care using automated systems, and standardized reporting of patient's characteristics and outcome measures using key performance indicators.

Red flags to screen for vertebral fracture in people presenting with low back pain.

Red flags to screen for vertebral fracture in people presenting with low back pain.

Active identification of vertebral fracture in the FLS model of care

SummaryThe identification of vertebral fracture is a key point in an FLS. We have analyzed the characteristics of 570 patients according to the route of identification (referral by other doctors, emergency registry or through VFA), concluding that promoting referral by other doctors with a training campaign is effective.PurposeVertebral fractures (VF) are associated with increased risk of further VFs. Our objective was to analyze the characteristics of patients with VF seen in a Fracture Liaison Service (FLS).MethodsAn observational study was carried out on patients with VF referred to the outpatient metabolic clinic (OMC) after a training campaign, identified in the emergency registry, and captured by VF assessment with bone densitometry (DXA-VFA) in patients with non-VFs. Patients with traumatic VF or VF > 1 year, infiltrative or neoplastic disease were excluded. The number and severity of VFs (Genant) were analyzed. Treatment initiation in the first 6 months after baseline visit was reviewed.ResultsOverall, 570 patients were included, mean age 73. The most common route for identifying VF was through referral to OMC (303 cases), followed by the emergency registry (198) and DXA-VFA (69). Osteoporosis by DXA was found in 312 (58%) patients and 259 (45%) had ≥ 2 VFs. The rate of grade 3 VFs was highest among patients on the emergency registry. Those identified through OMC had a higher number of VFs, a higher rate of osteoporosis, more risk factors and greater treatment initiation. Patients with VFs detected by DXA-VFA were mostly women with a single VF and had a lower rate of osteoporosis by DXA.ConclusionsWe present the distribution of VFs by the route of identification in an FLS. Promoting referral by other doctors with a training campaign may help in the quality improvement of the FLS-based model of care.

Diagnosis, Follow-Up and Therapy for Secondary Osteoporosis in Vulnerable Children: A Narrative Review

By definition, children constitute a vulnerable population, especially when they are chronically ill and/or disabled. A characteristic of chronically ill and disabled children is that they also suffer from indirect effects of their disease, such as immobilization, chronic inflammation, reduced time outdoors in the sun, osteotoxic effects of disease-targeted therapy (like glucocorticoids), and poor nutrition. All these factors may lead to bone fragility due to secondary osteoporosis, a co-morbidity that may be overlooked in the context of serious underlying diseases. The ultimate goal of osteoporosis diagnosis and monitoring in this setting is the early identification, prevention, and treatment of low-trauma long bone and vertebral fractures; indeed, vertebral fractures are a frequently under-diagnosed manifestation of overt bone fragility in this context. Efforts to prevent first-ever fractures are also meritorious, including encouragement of weight-bearing activities, optimization of nutritional status, including calcium and vitamin D supplementation, and the diagnosis and treatment of delayed growth and puberty; however, these conservative measures may be insufficient in those at high risk. Numerous natural history studies have shown that vertebral fractures are more common than non-vertebral (i.e., long bone) fractures in at-risk children. Not surprisingly, the cornerstone of secondary osteoporosis monitoring is lateral spine imaging for the early detection of vertebral collapse. Although dual-energy x-ray absorptiometry (DXA) is the gold standard to measure bone mineral density, digital X-ray radiogrammetry may be used as a surrogate measure of bone strength if dual-energy x-ray absorptiometry is not available. In the event that preventive measures fail, treatment with bisphosphonates may be appropriate. Typically, treatment with intravenous bisphosphonates is reserved for children with overt bone fragility and limited potential for spontaneous recovery. However, there is increasing attention to very high-risk children, such as boys with Duchenne muscular dystrophy, who may benefit from bisphosphonate therapy prior to first-ever fractures (given their high fracture frequency and essentially absent potential for spontaneous recovery). This article provides a contemporary overview of the definition and diagnosis of osteoporosis in children with chronic illness, along with the approach to monitoring those at risk and the evidence for currently recommended intervention strategies.

Should vertebral fracture assessment be performed in Fracture Liaison Service patients with non-vertebral fracture?

Vertebral fractures are often unappreciated. It was recently advocated that all Fracture Liaison Service (FLS) patients have densitometric VFA performed. We evaluated the likelihood of vertebral fracture identification with VFA in patients with prior fracture using the Manitoba Bone Density database. METHODS : VFA was performed in patients with T-scores below - 1.5 and age 70 + (or younger with height loss or glucocorticoid use) obtaining bone densitometry in Manitoba from 2010 to 2018. Those with prior clinical vertebral fracture, pathologic fracture, or uninterpretable VFA were excluded. Vertebral fractures were identified using the modified ABQ method. Health records were assessed for non-vertebral fracture (excluding head, neck, hand, foot) diagnosis codes unassociated with trauma prior to DXA. Multivariable odds ratios (ORs) for vertebral fracture were estimated without and with adjustment for age, sex, body mass index, ethnicity, area of residence, income level, comorbidity score, diabetes mellitus, falls in the last year, glucocorticoid use, and lowest BMD T-score. The study cohort consisted of 12,756 patients (94.4% women) with mean (SD) age 75.9 (6.8) years. Vertebral fractures were identified in 1925 (15.1%) overall. Vertebral fractures were significantly more likely (descending order) in those with prior pelvis, hip, humerus, other sites, and forearm, but not ankle fracture. There was modest attenuation with covariate adjustment but statistical significance was maintained. Prior hip, humerus, pelvis, forearm, and other fractures are associated with an increased likelihood of previously undiagnosed vertebral fracture, information useful for risk stratification and monitoring. These data support recommending VFA in FLS patients who are age 70 + with low BMD.

Application of tomosynthesis for vertebral compression fracture diagnosis and bone healing assessment in fracture liaison services.

Early identification of vertebral compression fractures (VCFs) is crucial for successful secondary fracture prevention. Tomosynthesis, a low-dose tomographic imaging technique, may facilitate the evaluation and long-term follow-up of VCFs in patients with osteoporosis. Herein, we compared the performances of plain radiography and tomosynthesis for VCF diagnosis and healing assessment in patients enrolled in fracture liaison services in our hospital. Forty-nine patients with new VCFs at the T10–L5 levels were prospectively recruited between August 2018 and May 2020; all patients underwent thoracolumbar plain radiography and tomosynthesis. We evaluated the accuracy of the VCF diagnosis, image quality, and VCFs healing process. Tomosynthesis identified 90 levels of VCF in 49 patients, while plain radiography revealed only 87.8% (79/90) of them. There were 44.9% (22/49) patients with neglected chronic VCFs as seen on tomosynthesis. Tomosynthesis images had improved VCF diagnostic accuracy up to 12.2% and showed significantly more anatomic details than plain radiography. For diagnosis of VCFs, the performance of plain radiographs was poorer than that of tomosynthesis images (plain radiographs: sensitivity 84%, specificity 93.5%, false positive rate 6.5%, and false negative rate 16%; tomosynthesis: sensitivity 93.2%, specificity 100%, false positive rate 0%, and false negative 6.8%), using magnetic resonance imaging (MRI) as gold standard. The Kappa coefficient between Tomosynthesis and MRI is 0.956 while between radiography and MRI is 0.704. Tomosynthesis showed significantly more anatomic details than plain radiography and all the examiners revealed a clear preference for tomosynthesis. Tomosynthesis scored 3.3 times higher on the fracture healing assessment at the 3-month follow-up than plain radiographs. Tomosynthesis is a promising tool for VCF screening and diagnosis in patients with osteoporosis and for monitoring fracture healing status at a low radiation dose and cost.

Development of a manufacturer-independent convolutional neural network for the automated identification of vertebral compression fractures in vertebral fracture assessment images using active learning.

Convolutional neural networks (CNNs) can identify vertebral compression fractures in GE vertebral fracture assessment (VFA) images with high balanced accuracy, but performance against Hologic VFAs is unknown. To obtain good classification performance, supervised machine learning requires balanced and labeled training data. Active learning is an iterative data annotation process with the ability to reduce the cost of labeling medical image data and reduce class imbalance. To train CNNs to identify vertebral fractures in Hologic VFAs using an active learning approach, and evaluate the ability of CNNs to generalize to both Hologic and GE VFA images. VFAs were obtained from the OsteoLaus Study (labeled Hologic Discovery A, n=2726), the Manitoba Bone Mineral Density Program (labeled GE Prodigy and iDXA, n=12,742), and the Canadian Longitudinal Study on Aging (CLSA, unlabeled Hologic Discovery A, n=17,190). Unlabeled CLSA VFAs were split into five equal-sized partitions (n=3438) and reviewed sequentially using active learning. Based on predicted fracture probability, 17.6% (n=3032) of the unlabeled VFAs were selected for expert review using the modified algorithm-based qualitative (mABQ) method. CNNs were simultaneously trained on Hologic, GE dual-energy and GE single-energy VFAs. Two ensemble CNNs were constructed using the maximum and mean predicted probability from six separately trained CNNs that differed due to stochastic variation. CNNs were evaluated against the OsteoLaus validation set (n=408) during the active learning process; ensemble performance was measured against the OsteoLaus test set (n=819). The baseline CNN, prior to active learning, achieved 55.0% sensitivity, 97.9% specificity, 57.9% positive predictive value (PPV), F1-score 56.4%. Through active learning, 2942 CLSA Hologic VFAs (492 fractures) were added to the training data-increasing the proportion of Hologic VFAs with fractures from 4.2% to 12.5%. With active learning, CNN performance improved to 80.0% sensitivity, 99.7% specificity, 94.1% PPV, F1-score 86.5%. The CNN maximum ensemble achieved 91.9% sensitivity (100% for grade 3 and 95.5% for grade 2 fractures), 99.0% specificity, 81.0% PPV, F1-score 86.1%. Simultaneously training on a composite dataset consisting of both Hologic and GE VFAs allowed for the development of a single manufacturer-independent CNN that generalized to both scanner types with good classification performance. Active learning can reduce class imbalance and produce an effective medical image classifier while only labeling a subset of available unlabeled image data-thereby reducing the time and cost required to train a machine learning model.

Prevention and Treatment of Glucocorticoid-Induced Osteoporosis in Adults: Consensus Recommendations From the Belgian Bone Club.

Glucocorticoids are effective immunomodulatory drugs used for many inflammatory disorders as well as in transplant recipients. However, both iatrogenic and endogenous glucocorticoid excess are also associated with several side effects including an increased risk of osteoporosis and fractures. Glucocorticoid-induced osteoporosis (GIOP) is a common secondary cause of osteoporosis in adults. Despite availability of clear evidence and international guidelines for the prevention of GIOP, a large treatment gap remains. In this narrative review, the Belgian Bone Club (BBC) updates its 2006 consensus recommendations for the prevention and treatment of GIOP in adults. The pathophysiology of GIOP is multifactorial. The BBC strongly advises non-pharmacological measures including physical exercise, smoking cessation and avoidance of alcohol abuse in all adults at risk for osteoporosis. Glucocorticoids are associated with impaired intestinal calcium absorption; the BBC therefore strongly recommend sufficient calcium intake and avoidance of vitamin D deficiency. We recommend assessment of fracture risk, taking age, sex, menopausal status, prior fractures, glucocorticoid dose, other clinical risk factors and bone mineral density into account. Placebo-controlled randomized controlled trials have demonstrated the efficacy of alendronate, risedronate, zoledronate, denosumab and teriparatide in GIOP. We suggest monitoring by dual-energy X-ray absorptiometry (DXA) and vertebral fracture identification one year after glucocorticoid initiation. The trabecular bone score might be considered during DXA monitoring. Extended femur scans might be considered at the time of DXA imaging in glucocorticoid users on long-term (≥ 3 years) antiresorptive therapy. Bone turnover markers may be considered for monitoring treatment with anti-resorptive or osteoanabolic drugs in GIOP. Although the pathophysiology of solid organ and hematopoietic stem cell transplantation-induced osteoporosis extends beyond GIOP alone, the BBC recommends similar evaluation, prevention, treatment and follow-up principles in these patients. Efforts to close the treatment gap in GIOP and implement available effective fracture prevention strategies into clinical practice in primary, secondary and tertiary care are urgently needed.

Incidental Identification of Vertebral Fragility Fractures by Chest CT in COVID-19-Infected Individuals

IntroductionIt is critical to identify asymptomatic vertebral compression fractures (VCFs) as soon as possible in order to avoid subsequent fragility fractures. The purpose of the study was to see how many vertebral compression fractures there were in patients admitted to the COVID-19 pneumonia unit in a single tertiary care hospital who underwent chest computed tomography (CT) scans.Materials and methodsSagittal reconstruction of the thoracic spine was done in around 504 patients and classified into mild, moderate, and severe categories, and we compared it with the radiological reports of the same.ResultsIn our study, the median age was 53 years (range: 31-91 years); 63% were men and 37% were women. Of the 504 patients, 76 (15%) had at least one vertebral compression fracture (VCF); 53 (10.2%) had one VCF, and 23 (4.8%) had multiple VCF, with 50 having mild fractures, 15 having moderate fractures, and 11 having severe fractures. Males (13.87%) and females (14.72%) had the same proportion of VCF (p = 0.83). Only 10% of the patients with VCFs we identified had a description in their report (eight patients).ConclusionThe reporting of VCF is insufficient. VCF detection should be included in the search patterns of radiologists and physicians, regardless of the primary reason for performing chest CT. Although many patients are unable to come to the hospital during pandemic/epidemic, careful evaluation and inclusion of mild fractures in reports, as well as an explanation of the risk of subsequent fractures and treatment accordingly, would completely eliminate the risk of subsequent fractures.

Understanding current UK practice for the incidental identification of vertebral fragility fractures from CT scans: an expert elicitation study

BackgroundThere is an emerging interest in using automated approaches to enable the incidental identification of vertebral fragility fractures (VFFs) on existing medical images visualising the spine.AimTo quantify values, and the degree of uncertainty associated with them, for the incidental identification of VFFs from computed tomography (CT) scans in current practice.MethodsAn expert elicitation exercise was conducted to generate point estimates and measures of uncertainty for four values representing the probability of: VFF being correctly reported by the radiologist; the absence of VFF being correctly assessed by the radiologist; being referred for management when a VFF is identified; having a dual-energy X-ray absorptiometry (DXA) scan after general practitioner (GP) referral. Data from a sample of seven experts in the diagnosis and management of people with VFFs were pooled using mathematical aggregation.ResultsThe estimated mean values for each probability parameter were: VFF being correctly reported by the radiologist = 0.25 (standard deviation (SD): 0.21); absence of VFF being correctly assessed by the radiologist = 0.89 (0.10); being referred for management when a VFF is identified by the radiologist = 0.15 (0.12); having a DXA scan after GP referral = 0.66 (0.28).DiscussionThese estimates could be used to facilitate the subsequent early economic evaluation of potential new approaches to improve the health outcomes of people with VFFs.ConclusionIn the absence of epidemiological studies, this study produced point estimates and measures of uncertainty for key parameters needed to describe current pathways for the incidental diagnosis of VFFs.

Inter-observer agreement of vertebral fracture assessment with dual-energy x-ray absorptiometry equipment.

To investigate the time and effort needed to perform vertebral morphometry, as well as inter-observer agreement for identification of vertebral fractures on vertebral fracture assessment (VFA) images. Ninety-six images were retrospectively selected, and three radiographers independently performed semi-automatic 6-point morphometry. Fractures were identified and graded using the Genant classification. Time needed to annotate each image was recorded, and reader fatigue was assessed using a modified Simulator Sickness Questionnaire (SSQ). Inter-observer agreement was assessed per-patient and per-vertebra for detecting fractures of all grades (grades 1-3) and for grade 2 and 3 fractures using the kappa statistic. Variability in measured vertebral height was evaluated using the intraclass correlation coefficient (ICC). Per-patient agreement was 0.59 for grades 1-3 fracture detection, and 0.65 for grades 2-3 only. Agreement for per-vertebra fracture classification was 0.92. Vertebral height measurements had an ICC of 0.96. Time needed to annotate VFA images ranged between 91 and 540s, with a mean annotation time of 259s. Mean SSQ scores were significantly lower at the start of a reading session (1.29; 95% CI: 0.81-1.77) compared to the end of a session (3.25; 95% CI: 2.60-3.90; p < 0.001). Agreement for detection of patients with vertebral fractures was only moderate, and vertebral morphometry requires substantial time investment. This indicates that there is a potential benefit for automating VFA, both in improving inter-observer agreement and in decreasing reading time and burden on readers.

Study on the Application of MRI in Forensic Identification of Vertebral Compression Fracture

Study on the Application of MRI in Forensic Identification of Vertebral Compression Fracture