Experience with artificial intelligence algorithms for the diagnosis of vertebral compression fractures based on computed tomography: from testing to practical evaluation

Study design Retrospective observational study. Purpose To identify the role of percutaneous vertebral biopsy in histopathological diagnosis of vertebral compression fractures and to identify the frequency of unexpected malignancy in vertebral compression fractures. Overview of literature Vertebral compression fractures are common in the Indian population. Magnetic resonance imaging and nuclear imaging have some limitations in the diagnosis of definitive pathology of vertebral compression fractures. Therefore, histological confirmation is necessary for definitive diagnosis and to plan appropriate management for patient. Methods A retrospective observational study was conducted involving 84 patients who underwent percutaneous vertebral biopsy between 2010 and 2014. We performed C-arm guided percutaneous transpedicular core vertebral biopsy of vertebral compression fractures under combination of local anesthesia and intravenous conscious sedation. Results Sufficient biopsy material was obtained in 79 of the 84 cases. In the other five cases, biopsy material was not sufficient for reporting. Out of the 79 cases, osteoporotic pathology was detected in 69 patients, malignancy was detected in 8 patients and no pathology was found in 2 patients. Two patients with distant metastases to vertebra were identified. Primary spinal malignancy was detected in 6 patients (1 unsuspected plasmacytoma, 5 diagnosed malignancy preoperatively). So, the frequency of unsuspected malignancy of this study was 1.19% (1/84). None of the patients had any complications. Conclusions C-arm guided percutaneous transpedicular vertebral biopsy is useful in obtaining definitive histopathological diagnosis of vertebral compression fractures, especially in differentiating malignant and non-malignant vertebral compression fractures and helping plan appropriate management of patients. The rate of unexpected malignancy in vertebral compression fracture was 1.19%.

To retrospectively assess the prevalence and clinical outcomes of unreported vertebral compression fractures at abdominal computed tomography (CT). This HIPAA-compliant study had institutional review board approval; the need for informed consent was waived for this retrospective analysis. A total of 2041 consecutive adult patients (1640 women, 401 men; age range, 19-94 years) underwent both abdominal multidetector CT and dual-energy x-ray absorptiometry (DXA) within 6 months of each other between 2000 and 2007, before sagittal CT reconstructions were obtained routinely. Transverse (axial) and retrospective sagittal multidetector CT reconstructions were reviewed for the presence of moderate or severe vertebral body compression fractures of the lower thoracic and lumbar spine by using the Genant visual semiquantitative method. Twenty-six patients were excluded for evidence of pathologic fracture or for technical factors limiting compression fracture detection. Electronic medical records were reviewed for patients with moderate or severe compression fractures to determine whether the fracture was reported at prospective CT interpretation, was known previously, or was diagnosed subsequently. Correlation was made with central DXA T scores. Statistical analysis was performed with the Student t test and Fisher exact test. At least one moderate or severe vertebral body compression fracture was identified retrospectively at CT in 97 patients (mean age, 70.8 years). Fractures involved one level in 67 and multiple levels in 30 patients, for a total of 141 fractures. In 81 (84%) patients, prospective CT diagnosis was not made. Patients in whom fractures were reported prospectively were significantly older and were more likely to have a severe compression fracture (P < .05). In 52 (64%) patients with an unreported fracture, the vertebral compression fracture was not known clinically. In 18 patients, subsequent diagnosis of a compression fracture was determined by means of another imaging study (median interval, 7 months). At DXA, 39 (48%) of 81 patients with unreported vertebral body compression fractures had a nonosteoporotic T score (greater than -2.5). Most clinically important vertebral body compression fractures in nontrauma patients at risk for low bone mineral density may go unreported at abdominal multidetector CT if sagittal reconstructions are not routinely evaluated.

Atraumatic vertebral compression fractures (VCFs) are commonly encountered in clinical practice and often represent a diagnostic challenge. MRI plays a major role in the differential diagnosis of benign and malignant VCFs, due to its high contrast resolution and the possibility to obtain quantitative and functional data with the employment of advanced sequences. Computer-aided diagnosis systems are also applied on MRI images for this purpose, showing promising results. In this setting, aim of this pictorial review is to elucidate the role of MRI in the differential diagnosis of VCFs with a specific focus on advanced and post-processing imaging techniques.

Objective To investigate the risk factors of degenerative scoliosis patients with osteoporotic vertebral compression fracture. Methods The data of 150 cases of degenerative scoliosis were retrospectively reviewed from October 2007 to October 2015. According to the occurrence of vertebral compressive fractures, the 150 patients were divided into the experiment group(fracture) and control group(non-fracture). Visual analogue scale(VAS) was used to assess the back pain, and the bone mineral density were measured in patients by using dual-energy X-ray absorptiometry, Cobb angle was measured to evaluate the severity of scoliosis. The occurrence of osteopontin in scoliosis were observed and analyzed. Logistic analysis was used to investigate the risk factors of vertebral compression fractures. Results ①In experiment group, there were T11 vertebral compression fractures in 4 cases, T12 vertebral compression fractures in 13 cases, L1 vertebral compression fractures in 16 cases, and both T12 and L1 vertebral fracture in 5 cases. ②The gender composition of the two groups had no significant difference(χ2=0.253, P=0.615). Compared with the control group, the average age of the experiment group was lower(t=4.392, P<0.01), VAS score was higher(t=14.350, P<0.01), more severe osteoporosis(t=13.296, P<0.01), a higher incidence of osteopontin(χ2=11.824, P=0.001). The Cobb angle of the two groups had no significant difference(t=1.749, P=0.082). After analysis of Logistic regression, trauma history(OR=1.57, 95% CI: 1.09-2.11), bone bridge formation(OR=3.46, 95% CI: 2.10-5.38) and bone osteoporosis(OR=2.48, 95% CI: 1.58-4.36) will increase the chances of patients with fractures. Conclusions Osteoporotic vertebral compression fracture usually occur in thoraco-lumbar region in patients with degenerative scoliosis, and its risk factor are trauma, history, osteoporosis and bone bridge formation. Key words: Osteoporotic fractures; Vertebral compression frature; Degenerative scoliosis; Risk factors

Vertebral osteoporotic fractures are a worldwide problem and can cause significant morbidity. retrospective analysis reviewing functional outcome of 70 patients who underwent balloon-kyphoplasty(BK). Inclusion criteria, a-patients above 60 years of age b-symptomatic patients who failed conservative treatment c-radiological diagnosis of vertebral compression fracture(VCF). Primary outcome was evaluation of functional outcome in mid- (1-3 years) and long term (&gt;3 years). Secondary outcomes a-analysis of the functional outcome of patients with severe disability or worse in mid- and long-term b-comparison of functional outcome between mid-term and long term follow-up c-correlate number of levels operated on with functional outcome. There were 70 patients with average age of 74 years. Lumbar and thoracic VCFs were included. Average follow-up was 2.7 years. Twenty-eight patients had long-term follow-up of ≥ 3 years and the remaining 42 had mid-term follow-up of &gt; 1 year and &lt; 3 years. All patients experienced clinical and statistically significant improvement in their Oswestry Disability Index score. This was the case in mid-term and long-term follow-up. This was also the case for patients presenting with severe disability. There was no correlation between the number of levels and functional outcome. 1. Osteoporotic vertebral compression fractures are fragility fractures and can be associated with significant morbidity. 2. Surgical treatment can help improve functional outcome on failure of conservative treatment. 3. Balloon kyphoplasty can significantly improve pa-tients' functional outcome in mid-term and long-term follow-up including patients presenting with severe disability or worse. 4. This study is of value in consenting kyphoplasty patients as they can expect an improvement in their disability in the mid-to-long term.

Background: Discriminating pathological from osteoporotic vertebral compression ‌a fracture is sometimes challenging.‌Aim: to create scoring system using several MRI signs, trying to correctly ‌differentiate osteoporotic from malignant compression vertebral fractures.‌Patients and method: a prospective case control analytic study was performed in ‌MRI unite of Al-Imamain Al-Kadhemain Medical City/ Baghdad/Iraq, 68 patients (24 ‌male and 44 female) were included, all were presenting with acute back pain and ‌vertebral compression fracture on x ray. MRI was done within 2 months from ‌symptom onset and follow up for at least 4-6 weeks. A 9 MRI signs were evaluated in ‌terms of sensitivity, specificity and accuracy in diagnosis of malignant fractures. A ‌stepwise analysis was used to create a scoring system to differentiate malignant from ‌osteoporotic vertebral compression fractures.‌Results: Malignant vertebral fracture was diagnosed in 13 patients and osteoporotic ‌vertebral fracture 55 patients. All used MRI signs had a variable specificity and ‌sensitivity for diagnosis of malignant fractures but all are statistically significant ‌except one. Mean score for patients with MVF was 3.5 (range of 3-6), while Mean ‌score in patients with OVF was 2.1 (range of 0-4). The optimum score cutoff value ‌was 4 with 61.5% sensitivity, 81.8% specificity, 79.4% accuracy and statistically ‌significant (P-value=0.012) for the discrimination of malignant from osteoporotic ‌vertebral compression fractures.‌Conclusion: Convex posterior vertebral border, asymmetrical diffuse or rounded ‌abnormal bone marrow signal, involvement of the pedicle, and para-vertebral soft ‌tissue mass are highly suggestive MRI sign of malignant fracture. Post contrast ‌enhancement is not a specific sign. Fluid cleft sign is suggestive of osteoporotic ‌fracture. This simple scoring system could be useful tool for the differentiation of ‌malignant and osteoporotic vertebral fractures with the best accurate cutoff value is ‌‌4. ‌Keywords: Magnetic resonance imaging, malignant vertebral compression fracture, ‌Osteoporotic vertebral compression fracture. ‌

Little is known about the natural course of pain from vertebral compression fractures (VCF). In this study we evaluated the pain course in conservatively treated patients with back pain and a VCF on the spine radiograph. Between May 2007 and November 2008, 169 patients with back pain referred by the general practitioner for spine radiographs and with a VCF were requested to participate in this follow-up study. Base line questionnaires about visual analogue scale (VAS) score, type of treatment and use of osteoporosis medication were filled in by 82 patients. Questionnaires were repeated at 6 weeks, and at 3, 6 and 12 months. Significant pain relief was defined as a decrease in VAS score of 50% or more from baseline. At baseline, mean VAS score in 82 patients was 6.9 (SD 2.0). Significant pain relief at 12 months was reported by 44 patients (54%) while in 38 patients (46%) pain relief was insufficient. No predictors for pain relief could be identified. Patients with insufficient pain relief at 12 months used significantly more analgesics and in these patients physiotherapy did better than other types of therapy. More than half of conservatively treated patients with back pain and VCF had sufficient pain relief at 12 months with most pain decrease in the first 3 months. However, a substantial proportion of patients still reported disabling pain. There were no predictors for the development of chronic pain. Patients with continuing pain ≥3 months after diagnosis of VCF may be candidates for vertebroplasty.

Objective To determine the risk factors for new vertebral fractures following percutaneous vertebroplasty (PVP) for osteoporotic vertebral compression fractures (OVCF). Methods This retrospective study included 703 patients (133 males and 570 females; 59-88 years, mean 76 years) treated with PVP for OVCF from January 2006 to June 2015. Involved vertebrae included 5 T6,3 T7,19 T8,16 T9,14 T10,46 T11,218 T12,241 L1,137 L2,108 L3,55 L4 and 37 L5. A total of 80 patients were detected with further vertebral fractures (new fracture group) and 623 patients with no new vertebral fractures (control group). Possible risk factors for new vertebral fractures were noted, including age, gender, body mass index (BMI), bone mineral density, involved vertebral segment, amount of cement injection, intradiscal cement leak, and anterior-posterior ratio of the fractured vertebrae. And the fracture group were subdivided into adjacent vertebral fracture group (n=48) and non-adjacent vertebral fracture group (n=32). Results Eighty (11%) of the 703 patients had subsequent vertebral fractures after PVP. Bone mineral density differed significantly between new fracture and control groups [(-3.31±0.74)SD vs. (-2.98±0.86)SD ] (P 0.05). There were no significant differences in all the factors between adjacent vertebral fracture and non-adjacent vertebral fracture groups (P>0.05). Conclusions Severe osteoporosis is the main risk factor of new vertebral fractures after OVCF managed with PVP. Therefore, treatment of osteoporosis is the key point to prevent postoperative new vertebral fractures. Key words: Spinal fractures; Vertebroplasty; Bone density

Objective To determine the effects of unilateral and bilateral vertebral transpedicular vertebroplasty and kyphoplasty for the treatment of severe osteoporotic vertebral compression fracture. Methods The clinical data of 63 patients with severe osteoporotic vertebral compression fracture, who underwent vertebroplasty or kyphoplasty in our hospital between January 2013 and August 2016, were retrospectively analyzed. The operative time, intraoperative fluoroscopy time, average vertebral bone cement injection volume, bone cement leakage, preoperative and postoperative visual analogue scale (VAS) scores, pain remission rate, and SF-36 health survey questionnaire were recorded. The changes of the vertebral anterior wall, center, and posterior wall vertical height and vertebral kyphosis angle were compared at the baseline, at 3d after the operation, and at the last follow-up. The patients were followed up for 1 year after the operaiton and the time of new vertebral fracture was recorded. Results For the treatment of severe osteoporotic vertebral compression fracture, unilateral transpedicular vertebroplasty was least costly on the operative time and fluoroscopy time. Its bone cement volume per section was less than that of the kyphoplasty, which was the same as the bilateral transpedicular vertebroplasty. The unilateral and bilateral transpedicular vertebroplasty showed more bone cement leakage and the new vertebral fracture time was shorter. The VAS score, physical health comprehensive score and mental health comprehensive score after the operation in the 3 groups were better than those before the operation. Both of the unilateral and bilateral transpedicular vertebroplasty could effectively relieve pain symptoms. The anterior wall, center and posterior wall vertical height significantly increased after the operation compared with those before the operation. The vertebral kyphosis angle in the 3 groups was significantly improved (all P<0.05). Kyphoplasty significantly favored the recovery of anterior wall, center and posterior wall vertical height and improvement of vertebral kyphosis angle. Conclusion For the treatment of severe osteoporotic vertebral compression fracture, vertebroplasty and kyphoplasty may effectively relieve pain symptoms, restore the vertebral body structure and vertebral height. The effect of unilateral and bilateral vertebroplasty is generally the same, with simple operation, short operative time and less cost, whereas the cement leakage rate and the recovery of vertebral height are less than those of kyphoplasty. Key words: Vertebroplasty; Osteoporosis; Compressive fractures

SummaryThe purpose of this study was to determine the real-world incidence and predictors of additional vertebroplasty or balloon kyphoplasty after initial vertebral augmentation, as a proxy for subsequent symptomatic vertebral fracture. Of patients, 15.5% underwent subsequent vertebral augmentation. The patient’s comorbidities are strongly associated with risk of subsequent treatment.PurposeTo determine the real-world incidence and predictors of additional vertebroplasty or balloon kyphoplasty after initial vertebral augmentation, as a proxy for subsequent symptomatic and disabling vertebral fracture.MethodsWe conducted a retrospective cohort study using commercial insurance claims data (Optum’s de-identified Clinformatics® Data Mart Database). Adult patients who underwent subsequent treatment for vertebral fracture within 24 months of initial balloon kyphoplasty (BKP) or vertebroplasty (VP) were classified into “subsequent treatment” or “no subsequent treatment” cohorts. Survival analysis was applied to investigate the effect of risk factors on subsequent treatment.ResultsBetween 1 January 2008 and 30 June 2020, a total of 32,513 adult patients underwent a BKP/VP procedure following a diagnosis of vertebral compression fracture in the preceding 12 months. Five thousand thirty-five patients (15.5%) underwent a subsequent BKP/VP treatment within 2 years; 90% had a single fracture level treated. An increased hazard of subsequent treatment was associated with a number of fractures treated at initial BKP/VP (≥ 4 levels, adjusted hazard ratio (AHR) 1.68 (95% CI 1.24–2.26); steroid use, AHR 1.9 (95% CI 1.31–1.48); Elixhauser Comorbidity Index ≥ 4, AHR 1.44 (95% CI 1.17–1.77); and multiple myeloma, AHR 1.31 (95% CI 1.13–1.53)). Age < 70 years was associated with reduced hazard of subsequent treatment (AHR 0.81, 95% CI 0.74–0.89).ConclusionsOne in seven patients underwent subsequent treatment for vertebral fracture after initial vertebral augmentation. Baseline patient characteristics were associated with increased risk of subsequent fracture within 2 years, suggesting that a patient’s natural history is strongly associated with risk of subsequent treatment rather than the initial surgical procedure itself.

X-ray diagnostics of compression fractures of bodies of the thoracic vertebrae in children not seldom causes great difficulties due to the fact that even in healthy children vertebrae have a number of features, particularly, the wedge shape. For the purpose of differential diagnosis there was performed chest X-ray examination of the thoracic vertebrae in healthy children and cases after the compression damage. The wedge index and the disk coefficient for various segments of the thoracic spine were calculated. The diagnostic efficiency of radionuclide studies was estimated to reaches 79%. The useof highly informative modern medical techniques (CT and MRI) in the diagnosis of vertebral compression fractures in children allows accurately and timely make the diagnosis and determine the condition of the surrounding tissues. Based on the analysis of medical records, radiographs, identification of options of radionuclide, CT and MRI studies, there was elaborated an algorithm for the diagnosis of compression fractures of vertebrae, which allowed optimize the diagnostic process. There are determined advantages of this algorithm.

Background: The causes of subsequent vertebral fractures after kyphoplasty are debated. It is reported that most new vertebral fractures after kyphoplasty develop in adjacent vertebrae. Objectives: We explored whether kyphoplasty increases the incidence of adjacent vertebral fractures and identified risk factors for new vertebral compression fractures (VCFs) after kyphoplasty. Study Design: Retrospective study. Setting: Inpatient population of a single center. Methods: We studied 356 patients treated with kyphoplasty from January 2008 to March 2012. Among those patients, there were 35 new VCFs after kyphoplasty. Subsequently, these patients were divided into 2 groups: an “adjacent fracture” group and a “nonadjacent fracture” group. In addition, all patients treated with kyphoplasty were further assigned to either a “new fracture” group or a “no fracture” group. Results: The occurrence of new VCFs in the “nonadjacent fracture” group was significantly higher than that in the “adjacent fracture” group. The average bone mineral density (BMD) of the spine was -3.95 in the “new fracture” group and -2.86 in the “no fracture” group. The risk of new vertebral fracture increased as the bone mineral density decreased (P &lt; 0.05). The morbidity of women was significantly higher in the “new fracture” group (94.29%) than in the “no fracture” group (77.88%) (P = 0.025). Limitations: Retrospective study at a single center. Conclusion: New VCFs after kyphoplasty occurred most often in nonadjacent vertebrae. VCFs after kyphoplasty were common in patients with low bone mineral density and in women, suggesting that osteoporosis is an underlying mechanism. Institutional Review: This study was approved by the institutional review board. Key words: Percutaneous kyphoplasty, vertebral compression fractures, bone mineral density, polymethylmethacrylate, adjacent vertebral fracture

Objective To investigate the effect of bisphosphonate medication (zoledronic acid,aclasta) on spinal fusion for osteoporotic patients through radiographic,clinical,and biological assessments.Methods A total of 79 patients with osteoporosis who were candidates for single-level posterior lumbar interbody fusion was randomly assigned to the experimental group (zoledronic acid injection,5mg,on the third day after surgery) or the control group (the same amount of saline injection,on the third day after surgery).Functional radiography and CT scans were used to evaluate fusion status.Bridging bone formation was graded into 3 categories:Grade A (bridging bone through bilateral vertebral),Grade B (bridging bone through a unilateral vertebral),or Grade C (incomplete bony bridging).The incidence of vertebral compression fractures occurring after surgery was assessed by means of MR imaging.A solid fusion was defined as less than 5° of angular motion in flexion-extension radiographs and the presence of Grade A or B bridging bone.Bone metabolic markers (β-C-terminal telopeptide of type Ⅰ collagen,β-CTX; and N-terminal propeptide of type Ⅰ collagen,PINP) were measured to investigate the biological effects of zoledronic acid on spinal fusion.Bone mineral density of femoral neck was measured by the dual X-ray absorptiometry.Clinical outcome was evaluated by means of the Oswestry Disability Index (ODI).Results Grade A or B bridging bone was more frequently observed in the experimental group at 3,6,and 9 months postoperatively (all P ＜ 0.05,respectively,Mann-Whitney U-test).At 12-months postoperative follow-up,bridging bone and solid fusion were not significantly different.No vertebral fractures were observed in the experimental group,whereas 6 patients in the control group showed vertebral compression fractures(P ＜ 0.05,Mann-Whitney U-test).Biochemical analysis of bone turnover demonstrated that zoledronic acid inhibited bone resorption from the early phase of the fusion process and also suppressed bone formation.Poor clinical results in the control group were demonstrated by ODI.Conclusions Osteoporosis patients undergoing spinal fusion who take bisphosphonates throughout the postoperative period was recommended. Key words: Imidazoles/therapeutic use; Phosphonic acids/therapeutic use; Osteoporosis/surgery; Lumbar vertebrae ; Fracture fixation, internal ; Spinal fusion ; Randomized controlled trial

Radiologists interpreting body CT scans may be the first to identify osteoporosis and associated vertebral fractures. This study correlates L1 trabecular attenuation measurements with prevalent vertebral body fractures in older adults undergoing routine CT. Mean L1 trabecular attenuation was measured at thoracoabdominal CT in 1966 consecutive adults (983 men and 983 women) 65 years and older (mean age, 74.1 ± 6.6 [SD] years). Sagittal reconstructions and lateral scouts were analyzed for moderate or severe thoracolumbar vertebral compression fractures according to the Genant semiquantitative assessment method. The diagnostic performance of L1 attenuation for the evaluation of prevalent vertebral fractures was assessed, including ROC curve analysis. A total of 162 (8.2%) individuals (mean age, 78.3 years; 66 men, 96 women) had at least one moderate or severe vertebral fracture. The mean L1 attenuation was 70.2 HU among patients with a prevalent fracture, whereas it was 132.3 HU among patients without fracture (p < 0.001). The prevalence of moderate or severe vertebral compression fractures was 32.5% when L1 attenuation was ≤ 90 HU. Prevalence increased to 49.2% with L1 attenuation of ≤ 50 HU. ROC curve analysis determined an optimal threshold of 90 HU (sensitivity = 86.9%, specificity = 83.9%), with a corresponding AUC of 0.895. The odds ratio of having a moderate or severe vertebral compression fracture was 31.9 for L1 attenuation ≤ 90 HU (95% CI, 20.2-50.5; p < 0.001). Patients with moderate or severe vertebral compression fractures have significantly lower L1 attenuation values than patients who do not. L1 attenuation ≤ 90 HU may represent an optimal threshold for determining risk for osteoporotic vertebral fractures.

We examined the relationship between bone mineral density, bone strength, skeletal muscle mass, grip strength, and skin autofluorescence (SAF) in women with osteoporotic vertebral compression fractures (VCF). A total of 1039 women (mean age 73.3 years) were included in our study. These included 222 cases of VCF (mean 77.8 years) and 817 controls (mean 72.0 years). Lumbar and femur BMD were measured for all participants using dual-energy X-ray absorptiometry (DXA). Bone strength surrogates, such as cross-sectional area (CSA) of the proximal femur, were evaluated using Advanced Hip Assessment software. SAF was measured with an autofluorescence reader. We used a bioelectrical impedance analyzer (BIA) to analyze body composition, including appendicular skeletal muscle mass index (SMI; appendicular lean mass (kg)/(height (m))2. We measured bone density, geometric parameters related to bone strength, skeletal muscle mass, grip strength, and SAF in both groups. We also examined factors related to vertebral fracture using multiple logistic regression analysis. Women with vertebral fractures had lower SMI (5.55 vs 5.76 kg/m2, p = 0.0006), smaller femoral cross-sectional area (97.20 vs 100.09, p = 0.014), lower grip strength (16.81 vs 19.16 kg, p < 0.0001), and increased skin autofluorescence (2.38 vs 2.25, p = 0.0002) compared to women without fractures. The prevalence of sarcopenia (SMI < 5.75) was 63.51% in VCF subjects and 52.02% in controls, revealing a high prevalence in VCF (p = 0.002). Skeletal muscle mass and grip strength were not significantly different between patients with acute and old VCF, suggesting that low skeletal muscle mass and muscle weakness may exist before fracture. From the multiple logistic regression analysis, lower femoral density (p = 0.0021), CSA (p = 0.0166), leg muscle mass (p = 0.0127), and left arm grip strength (p = 0.0255) were risk factors for vertebral compression fractures; all were negatively correlated with increased vertebral fractures. Lower limb muscle mass and grip loss may be closely related to the onset of vertebral compression fracture.