Detecting Distal Radial Fractures from Wrist Radiographs Using a Deep Convolutional Neural Network with an Accuracy Comparable to Hand Orthopedic Surgeons.

In recent years, fracture image diagnosis using a convolutional neural network (CNN) has been reported. The purpose of the present study was to evaluate the ability of CNN to diagnose distal radius fractures (DRFs) using frontal and lateral wrist radiographs. We included 503 cases of DRF diagnosed by plain radiographs and 289 cases without fracture. We implemented the CNN model using Keras and Tensorflow. Frontal and lateral views of wrist radiographs were manually cropped and trained separately. Fine-tuning was performed using EfficientNets. The diagnostic ability of CNN was evaluated using 150 images with and without fractures from anteroposterior and lateral radiographs. The CNN model diagnosed DRF based on three views: frontal view, lateral view, and both frontal and lateral view. We determined the sensitivity, specificity, and accuracy of the CNN model, plotted a receiver operating characteristic (ROC) curve, and calculated the area under the ROC curve (AUC). We further compared performances between the CNN and three hand orthopedic surgeons. EfficientNet-B2 in the frontal view and EfficientNet-B4 in the lateral view showed highest accuracy on the validation dataset, and these models were used for combined views. The accuracy, sensitivity, and specificity of the CNN based on both anteroposterior and lateral radiographs were 99.3, 98.7, and 100, respectively. The accuracy of the CNN was equal to or better than that of three orthopedic surgeons. The AUC of the CNN on the combined views was 0.993. The CNN model exhibited high accuracy in the diagnosis of distal radius fracture with a plain radiograph.

ISSLS PRIZE IN CLINICAL SCIENCE 2019: clinical importance of trunk muscle mass for low back pain, spinal balance, and quality of life-a multicenter cross-sectional study.

STUDY DESIGN: A multicenter cross-sectional study. OBJECTIVES: To clarify the relationship of trunk muscle mass with low back pain, spinal sagittal balance, and quality of life. Few reports have investigated the relationship of trunk muscle mass with lumbar spine function and spinal balance, and the clinical significance of trunk muscle mass remains unclear. METHODS: Patients attending spinal outpatient clinics at 10 different medical institutions were enrolled in this study. Patient demographics, trunk muscle mass and appendicular skeletal muscle mass (ASM) measured by bioelectrical impedance analysis (BIA), body mass index (BMI), Charlson Comorbidity Index (CCI), the Oswestry Disability Index (ODI), visual analog scale (VAS) for low back pain, sagittal vertical axis (SVA), and EuroQol 5 Dimension (EQ5D) score were investigated. Multivariate nonlinear regression analysis was used to investigate the association of trunk muscle mass with the ODI, VAS score, SVA, and EQ5D score. RESULTS: Of 2551 eligible patients, 1738 (mean age 70.2 ± 11.0 years; 781 men and 957 women) were enrolled. Trunk muscle mass was significantly correlated with the ODI, VAS score, SVA, and EQ5D score (P < 0.001) when adjusted for age, sex, BMI, ASM, CCI, and history of lumbar surgery. Patient deterioration was associated with a decrease in trunk muscle mass, and the deterioration accelerated from approximately 23 kg. CONCLUSIONS: Trunk muscle mass was significantly associated with the ODI, VAS score, SVA, and EQ5D score. Trunk muscle mass may assume an important role to elucidate and treat lumbar spinal dysfunction and spinal imbalance. These slides can be retrieved under Electronic Supplementary Material.

Efficacy and Safety of Beta-Tricalcium Phosphate/Polylactic-Co-Glycolic Acid for Implantation of Bone Defects.

INTRODUCTION: Bone defects are often observed after surgery for fractures and bone tumors. Their treatment is technically difficult and sometimes results in negative clinical and economic outcomes. To repair bone defects, a bone graft is implanted by selecting a transplant material from an autologous or artificial bone. Each method has its advantages and disadvantages. Compared to the gold standard of autologous bone graft, bone graft substitutes are not limited by the amount of harvested graft and avoid complications at the donor site. ORB-03 is a new cotton-like bone graft substitute composed of beta-tricalcium phosphate (ß-TCP) and a bioabsorbable polymer, polylactic-co-glycolic acid (PLGA). ORB-03 is easy to mold and can fill various bone defect shapes, and its three-dimensional microfiber scaffold can enhance the differentiation of osteoblasts and promote osteogenesis. We investigated the efficacy, ease of handling, and safety of ORB-03 as a bone graft substitute. A multicenter, open-label, single-group study was conducted at six institutions. METHODS: Between July 2018 and August 2019, 60 patients with bone defects caused by fracture, benign tumors, or an iliac donor site from bone harvesting were enrolled in this study; 54 patients were finally included for the safety analysis and 48 patients for the image analysis. During surgery, ORB-03 was mixed with the patient's blood and molded into a bone defect. To evaluate the efficacy of ORB-03, radiography and computed tomography (CT) were performed at intervals until 24 weeks after surgery. RESULTS: The effective rate and its accurate bilateral 95% confidence interval (CI) were calculated based on the efficacy criteria at 24 weeks postoperatively. The ease with which ORB-03 could be handled in surgery was evaluated. Adverse events that occurred after surgery were evaluated, and those associated with ORB-03 were examined. Bone fusion was good in all cases, and the radiography and CT effective rates were 100.0% and 91.5%, respectively. Handling was easy in all cases. There were four adverse events, none of which were clinically problematic. CONCLUSIONS: ORB-03 was found to be easy to handle, safe, and effective as a bone graft substitute for bone defects.

The unfolded protein response is a major mechanism by which LRP1 regulates Schwann cell survival after injury.

In peripheral nerve injury, Schwann cells (SCs) must survive to exert a continuing and essential role in successful nerve regeneration. Herein, we show that peripheral nerve injury is associated with activation of endoplasmic reticulum (ER) stress and the adaptive unfolded protein response (UPR). The UPR culminates in expression of C/EBP homology protein (CHOP), a proapoptotic transcription factor in SCs, unless counteracted by LDL receptor-related protein-1 (LRP1), which serves as a major activator of phosphatidylinositol 3-kinase (PI3K). Sciatic nerve crush injury in rats induced expression of the ER chaperone GRP78/BIP, reflecting an early, corrective phase of the UPR. However, when LRP1 signaling was inhibited with receptor-associated protein, PI3K activity was decreased and CHOP protein expression increased, particularly in myelinating SCs. In cultured SCs, the PKR-like ER kinase target eIF2α was phosphorylated and CHOP was induced by (1) inhibiting PI3K, (2) treating the cells with tumor necrosis factor-α (TNF-α), or (3) genetic silencing of LRP1. CHOP gene deletion in SCs decreased cell death in response to TNF-α. Furthermore, the effects of TNF-α on phosphorylated eIF2α, CHOP, and SC death were blocked by adding LRP1 ligands that augment LRP1-dependent cell signaling to PI3K. Collectively, our results support a model in which UPR-activated signaling pathways represent a major challenge to SC survival in nerve injury. LRP1 functions as a potent activator of PI3K in SCs and, by this mechanism, limits SC apoptosis resulting from increased CHOP expression in nerve injury.

Neuroprotective therapy using granulocyte colony-stimulating factor for acute spinal cord injury: a phase I/IIa clinical trial.

OBJECTIVE: Granulocyte colony-stimulating factor (G-CSF) is a cytokine that is clinically used to treat neutropenia. G-CSF also has non-hematopoietic functions and could potentially be used to treat neuronal injury. To confirm the safety and feasibility of G-CSF administration for acute spinal cord injury (SCI), we have initiated a phase I/IIa clinical trial of neuroprotective therapy using G-CSF. METHODS: The trial included a total of 16 SCI patients within 48 h of onset. In the first step, G-CSF (5 µg/kg/day) was intravenously administered for 5 consecutive days to 5 patients. In the second step, G-CSF (10 µg/kg/day) was similarly administered to 11 patients. We evaluated motor and sensory functions of patients using the American Spinal Cord Injury Association (ASIA) score and ASIA impairment scale (AIS) grade. RESULTS: In all 16 patients, neurological improvement was obtained after G-CSF administration. AIS grade increased by one step in 9 of 16 patients. A significant increase in ASIA motor scores was detected 1 day after injection (P < 0.01), and both light touch and pin prick scores improved 2 days after injection (P < 0.05) in the 10 µg group. No severe adverse effects were observed after G-CSF injection. CONCLUSION: These results indicate that intravenous administration of G-CSF (10 µg/kg/day) for 5 days is essentially safe, and suggest that some neurological recovery may occur in most patients. We suggest that G-CSF administration could be therapeutic for patients with acute SCI.

Determining the short-term neurological prognosis for acute cervical spinal cord injury using machine learning.

It is challenging to predict neurological outcomes of acute spinal cord injury (SCI) considering issues such as spinal shock and injury heterogeneity. Deep learning-based radiomics (DLR) were developed to quantify the radiographic characteristics automatically using a convolutional neural network (CNN), and to potentially allow the prognostic stratification of patients. We aimed to determine the functional prognosis of patients with cervical SCI using machine learning approach based on MRI and to assess the ability to predict the neurological outcomes. We retrospectively analyzed the medical records of SCI patients (n=215) who had undergone MRI and had an American Spinal cord Injury Association Impairment Scale (AIS) assessment at 1 month after injury, enrolled with a total of 294 MR images. Sagittal T2-weighted MR images were used for the CNN training and validation. The deep learning framework TensorFlow was used to construct the CNN architecture. After we calculated the probability of the AIS grade using the DLR, we built the identification model based upon the random forest using 3 features: the probability of each AIS grade obtained by the DLR method, age, and the initial AIS grade at admission. We performed a statistical evaluation between the actual and predicted AIS. The accuracy, precision, recall and f1 score of the ensemble model based on the DLR and RF were 0.714, 0.590, 0.565 and 0.567, respectively. The present study demonstrates that prediction of the short-term neurological outcomes for acute cervical spinal cord injury based on MRI using machine learning is feasible.

Regulation of cytokine expression by Schwann cells in response to α2-macroglobulin binding to LRP1.

Binding of activated α(2)-macroglobulin (α(2)M) to LDL receptor-related protein-1 (LRP1) in Schwann cells activates ERK/MAP kinase and Akt and thereby promotes cell survival and migration. The goal of this study was to determine whether α(2)M binding to LRP1 regulates expression of cytokines and chemokines. To assess the LRP1 response selectively, we studied primary cultures of rat Schwann cells. In a screening assay that detects 84 gene products, monocyte chemoattractant protein-1 (MCP-1/CCL2) mRNA expression was increased more than 13-fold in Schwann cells treated with activated α(2)M. The effects of α(2)M on MCP-1 expression were selective, because expression of the general proinflammatory cytokine tumor necrosis factor-α (TNF-α) was not induced. We confirmed that α(2)M selectively induces expression of MCP-1 and not TNF-α in single-target qPCR assays. MCP-1 protein accumulated at increased levels in conditioned medium of α(2)M-treated cells. LRP1 was necessary for induction of MCP-1 expression, as determined in experiments with the LRP1 antagonist receptor-associated protein, a mutated form of full-length α(2)M that does not bind LRP1, and in studies with Schwann cells in which LRP1 was silenced. Inhibiting ERK/MAP kinase activation blocked expression of MCP-1. These studies support a model in which LRP1 regulates multiple aspects of Schwann cell physiology in the response to PNS injury.

Gender-specific analysis for the association between trunk muscle mass and spinal pathologies.

We investigated the relationship between trunk muscle mass and spinal pathologies by gender. This multicenter cross-sectional study included patients aged ≥ 30 years who visited a spinal outpatient clinic. Trunk and appendicular muscle mass were measured using bioelectrical impedance analysis. The Oswestry Disability Index (ODI), visual analog scale (VAS) score for low back pain, sagittal vertical axis (SVA), and EuroQol 5 Dimension (EQ5D) score were investigated to evaluate spinal pathology. The association between trunk muscle mass and these parameters was analyzed by gender using a non-linear regression model adjusted for patients' demographics. We investigated the association between age and trunk muscle mass. We included 781 men and 957 women. Trunk muscle mass differed significantly between men and women, although it decreased with age after age 70 in both genders. Lower trunk muscle mass was significantly associated with ODI, SVA, and EQ5D score deterioration in both genders; its association with VAS was significant only in men. Most parameters deteriorated when trunk muscle mass was < 26 kg in men and < 19 kg in women. Lower trunk muscle mass was associated with lumbar disability, spinal imbalance, and poor quality of life in both genders, with significant difference in muscle mass.

Lumbopelvic Fixation for Sacral Insufficiency Fracture Presenting with Sphincter Dysfunction.

Sacral insufficiency fractures (SIFs) are common in the elderly. In patients with SIF, objective neurological abnormalities such as sphincter dysfunction or leg paresthesia are uncommon. We present a case of SIF accompanied by spinopelvic dissociation with late neurological compromise treated by spinopelvic fixation. A 61-year-old woman presented to our hospital with low back pain without obvious trauma history. She had a past history of eosinophilic granulomatosis with polyangiitis and treatment with steroids. Her low back pain became worse, and she started to have radiating left posterior thigh pain and motor weakness in the left ankle and both great toes with symptoms of stress urinary incontinence, constipation, and loss of anal sensation. Magnetic resonance imaging revealed an H-shaped sacrum fracture. We attributed the neurological symptoms to unstable SIF and performed lumbopelvic fixation. After the surgery, her leg pain and symptoms of stress urinary incontinence improved markedly, as did anal sensation. At a 6-month follow-up, the patient reported no low back pain and she was walking independently without pelvic complaints. CT showed bone union was achieved. Even minimally displaced SIF in patients with osteoporosis can be a cause of bowel and bladder disturbance. Lumbopelvic fixation is a treatment option for SIF with spinopelvic dissociation presenting neurological deficit.

Occipital Condyle Fracture: A Case Report of a Typically Stable Fracture That Required Surgical Treatment.

An occipital condyle fracture (OCF) is a relatively rare trauma that is now increasingly diagnosed because of the wide availability of computed tomography. For nondisplaced OCFs, conservative treatment is generally recommended, and there is no previous report of a nondisplaced OCF requiring surgery. We report a patient who had a nondisplaced OCF with craniocervical misalignment (a condyle-C1 interval > 2.0 mm) and C1-C2 translation treated with a halo vest and occipitocervical fusion surgery. An 87-year-old Asian woman fell from a 4-meter height and hit her head. She was transferred to our emergency room. Computed tomography revealed a nondisplaced impaction OCF with a 2.5 mm occipital condyle-C1 interval and a 5 mm C1-C2 translation. The fracture pattern was considered stable. However, since craniocervical misalignment and C1-C2 translation were present, the patient was placed in a halo device, and we reduced the occipitoatlantoaxial joint, adjusting the halo ring position preoperatively. Confirming reduction of the atlantooccipital facet joint and the atlantoaxial joint by computed tomography, we performed an occipitocervical fusion. This is the first report of a nondisplaced OCF with craniocervical misalignment and C1-C2 translation that required surgical treatment. Clinicians should be aware of craniocervical misalignment and atlantoaxial instability even in Tuli type 1 OCFs.

Perioperative Complications in 155 Patients Who Underwent Oblique Lateral Interbody Fusion Surgery: Perspectives and Indications From a Retrospective, Multicenter Survey.

STUDY DESIGN: A retrospective multicenter survey. OBJECTIVE: To investigate the perioperative complications of oblique lateral interbody fusion (OLIF) surgery. SUMMARY OF BACKGROUND DATA: OLIF has been widely performed to achieve minimally invasive, rigid lumbar lateral interbody fusion. The associated perioperative complications are not yet well described. METHODS: The participants were patients who underwent OLIF surgery under the diagnosis of degenerative lumbar diseases between April 2013 and May 2015 at 11 affiliated medical institutions. The collected data were classified into intraoperative and early-stage postoperative (≤1 mo) complications. The intraoperative complications were then subcategorized into organ damage (neural, vertebral, vascular, and others) and other complications, mainly related to instrumental failure. The collected data were also divided and analyzed based on whether the surgeon was certified to perform the surgery and the incidence of complications in the early (April 2013-March 2014) and late stages (April 2014-May 2015) of OLIF introduction. RESULTS: In the 155 included patients, 75 complications were reported (incidence rate, 48.3%). The most common complication was endplate fracture/subsidence (18.7%), followed by transient psoas weakness and thigh numbness (13.5%) and segmental artery injury (2.6%). Almost all these complications were transient, except for three patients who had permanent damage: one had ureteral injury and two had neurological injury. Postoperative complications included surgical site infection (1.9%) and reoperation (1.9%). Whether the primary operator was experienced did not affect the incidence of complications. Regarding the introductory stage, the incidence of complications was 50% in the early stage and 38% in the late stage. CONCLUSION: The overall incidence of perioperative complications of OLIF surgery reached 48.3%, of which only 1.9% resulted in permanent damage. Our analysis based on surgeon experience indicated that the OLIF procedure could be performed without increasing incidence of complications, under the guidance of experienced supervisors. LEVEL OF EVIDENCE: 3.

Low-density lipoprotein receptor related protein-1 (LRP1)-dependent cell signaling promotes neurotrophic activity in embryonic sensory neurons.

Developing sensory neurons require neurotrophic support for survival, neurite outgrowth and myelination. The low-density lipoprotein receptor-related protein-1 (LRP1) transactivates Trk receptors and thereby functions as a putative neurotrophin. Herein, we show that LRP1 is abundantly expressed in developing dorsal root ganglia (DRG) and that LRP1-dependent cell signaling supports survival, neurite extension and receptivity to Schwann cells even in the absence of neurotrophins. Cultured embryonic DRG neurons (E15) were treated with previously characterized LRP1 ligands, LRP1-receptor binding domain of α2-macroglobulin (RBD), hemopexin domain of MMP-9 (PEX) or controls (GST) for two weeks. These structurally diverse LRP1 ligands significantly activated and sustained extracellular signal-regulated kinases (ERK1/2) 5-fold (p<0.05), increased expression of growth-associated protein-43(GAP43) 15-fold (P<0.01), and increased neurite outgrowth 20-fold (P<0.01). Primary sensory neurons treated with LRP1 ligands survived > 2 weeks in vitro, to an extent equaling NGF, a finding associated with canonical signaling mechanisms and blockade of caspase-3 cleavage. LRP1 ligand-induced survival and sprouting were blocked by co-incubation with the LRP1 antagonist, receptor associated protein (RAP), whereas RAP had no effect on NGF-induced activity. Site directed mutagenesis of the LRP1 ligand, RBD, in which Lys(1370) and Lys(1374) are converted to alanine to preclude LRP1 binding, were ineffective in promoting cell signaling, survival or inducing neurite extension in primary sensory neurons, confirming LRP1 specificity. Furthermore, LRP1-induced neurite sprouting was mediated by Src-family kinase (SFK) activation, suggesting transactivation of Trk receptors. Co-cultures of primary embryonic neurons and Schwann cells showed that LRP1 agonists promoted axonal receptivity to myelination to Schwann cells. Collectively, these findings identify LRP1 as a novel and perhaps essential trophic molecule for sensory neuronal survival and development.

Neuroprotective therapy using granulocyte colony-stimulating factor for patients with worsening symptoms of thoracic myelopathy: a multicenter prospective controlled trial.

STUDY DESIGN: An open-labeled multicenter prospective controlled clinical trial. OBJECTIVE: To confirm the feasibility of granulocyte colony-stimulating factor (G-CSF) administration for patients with thoracic myelopathy. SUMMARY OF BACKGROUND DATA: Although G-CSF is best known as an important cytokine commonly used to treat neutropenia, it also has nonhematopoietic functions. Previous experimental studies have shown that G-CSF can enhance tissue regeneration of several organs, such as the heart and the brain. We previously reported that G-CSF promotes functional recovery after spinal cord injury in rodents. On the basis of those findings, we started a clinical trial of neuroprotective therapy, using G-CSF for patients with worsening symptoms of thoracic myelopathy. METHODS: Patients whose Japanese Orthopaedic Association (JOA) score for thoracic myelopathy had decreased 2 points or more during a recent 1-month period were eligible for entry. After giving informed consent, patients were assigned to G-CSF and control groups. The G-CSF group (n = 10) received G-CSF 10 µg/kg per day intravenously for 5 consecutive days. The control group (n = 14) received similar treatments as the G-CSF group except for G-CSF administration. The primary outcome was JOA recovery rate at 1 month after G-CSF administration or initial treatment. RESULTS: There was greater improvement in neurological functioning between baseline and 1-month follow-up in the G-CSF group (JOA recovery rate: 29.1 ± 20.5%) than in the control group (JOA recovery rate: 1.1 ± 4.2%) (P < 0.01). No serious adverse events occurred during or after the G-CSF administration. CONCLUSION: The results provide evidence that G-CSF administration caused neurological recovery in patients with worsening symptoms of thoracic compression myelopathy.

Simulated surgery for a patient with neurofibromatosis type-1 who had severe cervicothoracic kyphoscoliosis and an anomalous vertebral artery.

STUDY DESIGN: Case report. OBJECTIVE: To describe the usefulness of simulated surgery for evaluation of a patient with neurofibromatosis type-1 (NF-1) who had severe cervicothoracic kyphoscoliosis and an anomalous vertebral artery (VA). SUMMARY OF BACKGROUND DATA: Several surgical procedures have been used in the treatment of cervicothoracic kyphoscoliosis associated with myelopathy in patients with NF-1. However, to our knowledge, there has been no report that describes a surgical procedure for NF-1 patients with anomalous VA at the cervical spine. METHODS: A 45-year-old man with NF-1 developed cervical myelopathy. Preoperative examinations revealed severe cervicothoracic kyphoscoliosis, dystrophic changes of the cervical vertebrae, and the anomalous course of a VA and VA aneurysms. To assist in the preoperative planning and intraoperative navigation, we created 3-dimensional (3D) full-scale models of the patient's spine. Using a model, we performed a simulation of the planned surgery for spinal cord decompression with spinal fusion through both anterior and posterior approaches. RESULTS: Through the simulation, we could evaluate the risk of VA injury at the process of corpectomy, and altered the surgical procedure for the spinal cord decompression with spinal fusion from a posterior approach and a bone graft alone from an anterior approach. We accomplished the surgery successfully without any neurovascular complications. After surgery, the patient experienced relief from myelopathy. CONCLUSION: Preoperative surgical simulation using a 3D full-scale model was useful for improving the accuracy and safety of the surgery for cervicothoracic kyphoscoliosis with NF-1.

Efficacy and reliability of highly functional open source DICOM software (OsiriX) in spine surgery.

We evaluated the feasibility and reliability of open source Digital Imaging and COmmunication in Medicine (DICOM) imaging software, OsiriX (Antoine Rosset, 2003-2009), in spine surgery. CT data were used and processed with OsiriX and with commercial software for comparison. Images were reconstructed and compared in volume rendering (VR) and multi-planar reconstruction (MPR) mode. When all images were compared, the three-dimensional (3D) reconstructed images from both software packages showed considerable consistency in VR mode. Measurements in MPR mode also showed similar values with no statistically significant difference. These results demonstrate that OsiriX has approximately equivalent values to commercial software and provides reliable preoperative 3D information for the surgical field. In addition, any clinician, can obtain information using OsiriX at any time. Thus, OsiriX is a helpful tool in preoperative planning for spine surgery.