Noncontact Sensors for Vital Signs Measurement: A Narrative Review.

Vital signs are crucial for monitoring changes in patient health status. This review compared the performance of noncontact sensors with traditional methods for measuring vital signs and investigated the clinical feasibility of noncontact sensors for medical use. We searched the Medical Literature Analysis and Retrieval System Online (MEDLINE) database for articles published through September 30, 2023, and used the key search terms "vital sign," "monitoring," and "sensor" to identify relevant articles. We included studies that measured vital signs using traditional methods and noncontact sensors and excluded articles not written in English, case reports, reviews, and conference presentations. In total, 129 studies were identified, and eligible articles were selected based on their titles, abstracts, and full texts. Three articles were finally included in the review, and the types of noncontact sensors used in each selected study were an impulse radio ultrawideband radar, a microbend fiber-optic sensor, and a mat-type air pressure sensor. Participants included neonates in the neonatal intensive care unit, patients with sleep apnea, and patients with coronavirus disease. Their heart rate, respiratory rate, blood pressure, body temperature, and arterial oxygen saturation were measured. Studies have demonstrated that the performance of noncontact sensors is comparable to that of traditional methods of vital signs measurement. Noncontact sensors have the potential to alleviate concerns related to skin disorders associated with traditional skin-contact vital signs measurement methods, reduce the workload for healthcare providers, and enhance patient comfort. This article reviews the medical use of noncontact sensors for measuring vital signs and aimed to determine their potential clinical applicability.

Proteomic profiling of extracellular vesicles derived from human serum for the discovery of biomarkers in Avascular necrosis.

BACKGROUND: Avascular necrosis (AVN) is a medical condition characterized by the destruction of bone tissue due to a diminished blood supply. When the rate of tissue destruction surpasses the rate of regeneration, effective treatment becomes challenging, leading to escalating pain, arthritis, and bone fragility as the disease advances. A timely diagnosis is imperative to prevent and initiate proactive treatment for osteonecrosis. We explored the potential of differentially expressed proteins in serum-derived extracellular vesicles (EVs) as biomarkers for AVN of the femoral head in humans. We analyzed the genetic material contained in serum-derived exosomes from patients for early diagnosis, treatment, and prognosis of avascular necrosis. METHODS: EVs were isolated from the serum of both patients with AVN and a control group of healthy individuals. Proteomic analyses were conducted to compare the expression patterns of these proteins by proteomic analysis using LC-MS/MS. RESULTS: Our results show that the levels of IGHV3-23, FN1, VWF, FGB, PRG4, FCGBP, and ZSWIM9 were upregulated in the EVs of patients with AVN compared with those of healthy controls. ELISA results showed that VWF and PRG4 were significantly upregulated in the patients with AVN. CONCLUSIONS: These findings suggest that these EV proteins could serve as promising biomarkers for the early detection and diagnosis of AVN. Early diagnosis is paramount for effective treatment, and the identification of new osteonecrosis biomarkers is essential to facilitate swift diagnosis and proactive intervention. Our study provides novel insights into the identification of AVN-related biomarkers that can enhance clinical management and treatment outcomes.

Detection of Cervical Foraminal Stenosis from Oblique Radiograph Using Convolutional Neural Network Algorithm.

PURPOSE: This study was conducted to develop a convolutional neural network (CNN) algorithm that can diagnose cervical foraminal stenosis using oblique radiographs and evaluate its accuracy. MATERIALS AND METHODS: A total of 997 patients who underwent cervical MRI and cervical oblique radiographs within a 3-month interval were included. Oblique radiographs were labeled as "foraminal stenosis" or "no foraminal stenosis" according to whether foraminal stenosis was present in the C2-T1 levels based on MRI evaluation as ground truth. The CNN model involved data augmentation, image preprocessing, and transfer learning using DenseNet161. Visualization of the location of the CNN model was performed using gradient-weight class activation mapping (Grad-CAM). RESULTS: The area under the curve (AUC) of the receiver operating characteristic curve based on DenseNet161 was 0.889 (95% confidence interval, 0.851-0.927). The F1 score, accuracy, precision, and recall were 88.5%, 84.6%, 88.1%, and 88.5%, respectively. The accuracy of the proposed CNN model was significantly higher than that of two orthopedic surgeons (64.0%, p<0.001; 58.0%, p<0.001). Grad-CAM analysis demonstrated that the CNN model most frequently focused on the foramen location for the determination of foraminal stenosis, although disc space was also frequently taken into consideration. CONCLUSION: A CNN algorithm that can detect neural foraminal stenosis in cervical oblique radiographs was developed. The AUC, F1 score, and accuracy were 0.889, 88.5%, and 84.6%, respectively. With the current CNN model, cervical oblique radiography could be a more effective screening tool for neural foraminal stenosis.

Application of non-contact sensors for health monitoring in hospitals: a narrative review.

The continuous monitoring of the health status of patients is essential for the effective monitoring of disease progression and the management of symptoms. Recently, health monitoring using non-contact sensors has gained interest. Therefore, this study aimed to investigate the use of non-contact sensors for health monitoring in hospital settings and evaluate their potential clinical applications. A comprehensive literature search was conducted using PubMed to identify relevant studies published up to February 26, 2024. The search terms included "hospital," "monitoring," "sensor," and "non-contact." Studies that used non-contact sensors to monitor health status in hospital settings were included in this review. Of the 38 search results, five studies met the inclusion criteria. The non-contact sensors described in the studies were radar, infrared, and microwave sensors. These non-contact sensors were used to obtain vital signs, such as respiratory rate, heart rate, and body temperature, and were then compared with the results from conventional measurement methods (polysomnography, nursing records, and electrocardiography). In all the included studies, non-contact sensors demonstrated a performance similar to that of conventional health-related parameter measurement methods. Non-contact sensors are expected to be a promising solution for health monitoring in hospital settings.

Tibial Axis-to-Talus Distance: A Clinically Reliable Measurement for Sagittal Translation of the Talus in Total Ankle Arthroplasty.

Background: Sagittal talar translation is an important factor influencing the sagittal alignment of total ankle arthroplasty (TAA). Thus, accurate measurement of sagittal talar translation is crucial. This study proposes a simple method (tibiotalar distance [TTD]) that can quantify talar translation without being affected by the ankle and subtalar joint condition or the talar component position in patients with TAA. Methods: We enrolled 280 eligible patients (296 ankles) who underwent primary TAA between 2005 and 2019 and retrospectively reviewed them for sagittal talar translation. The TTD was measured for each patient on weight-bearing lateral ankle radiographs by 3 raters. In addition, we analyzed interrater and intrarater reliability for the TTD method. Results: We found that the TTD method could quantify the talar translation and was not affected by the preoperative condition of the ankle joint surface, subtalar joint pathologies, or the postoperative talar component position. The TTD method showed an excellent intraclass correlation coefficient (> 0.9) in all interrater and intrarater reliability analyses. In the analysis of 157 healthy, unoperated contralateral ankles, we identified that TTD showed a Gaussian distribution (p = 0.284) and a mean of 38.91 mm (normal range, 29.63-48.20 mm). Conclusions: The TTD method is a simple and reliable method that could be applied to patients with TAA to assess the sagittal talar translation regardless of the pre-and postoperative joint condition and implantation status.

Ultrasound-triggered three dimensional hyaluronic acid hydrogel promotes in vitro and in vivo reprogramming into induced pluripotent stem cells.

Cellular reprogramming technologies have been developed with different physicochemical factors to improve the reprogramming efficiencies of induced pluripotent stem cells (iPSCs). Ultrasound is a clinically applied noncontact biophysical factor known for regulating various cellular behaviors but remains uninvestigated for cellular reprogramming. Here, we present a new reprogramming strategy using low-intensity ultrasound (LIUS) to improve cellular reprogramming of iPSCs in vitro and in vivo. Under 3D microenvironment conditions, increased LIUS stimulation shows enhanced cellular reprogramming of the iPSCs. The cellular reprogramming process facilitated by LIUS is accompanied by increased mesenchymal to epithelial transition and histone modification. LIUS stimulation transiently modulates the cytoskeletal rearrangement, along with increased membrane fluidity and mobility to increase HA/CD44 interactions. Furthermore, LIUS stimulation with HA hydrogel can be utilized in application of both human cells and in vivo environment, for enhanced reprogrammed cells into iPSCs. Thus, LIUS stimulation with a combinatorial 3D microenvironment system can improve cellular reprogramming in vitro and in vivo environments, which can be applied in various biomedical fields.

Bone grafting for periprosthetic bone cysts following total ankle arthroplasty.

Aims: The purpose of this study was to assess the success rate and functional outcomes of bone grafting for periprosthetic bone cysts following total ankle arthroplasty (TAA). Additionally, we evaluated the rate of graft incorporation and identified associated predisposing factors using CT scan. Methods: We reviewed a total of 37 ankles (34 patients) that had undergone bone grafting for periprosthetic bone cysts. A CT scan was performed one year after bone grafting to check the status of graft incorporation. For accurate analysis of cyst volumes and their postoperative changes, 3D-reconstructed CT scan processed with 3D software was used. For functional outcomes, variables such as the Ankle Osteoarthritis Scale score and the visual analogue scale for pain were measured. Results: Out of 37 ankles, graft incorporation was successful in 30 cases. Among the remaining seven cases, four (10.8%) exhibited cyst re-progression, so secondary bone grafting was needed. After secondary bone grafting, no further progression has been noted, resulting in an overall 91.9% success rate (34 of 37) at a mean follow-up period of 47.5 months (24 to 120). The remaining three cases (8.1%) showed implant loosening, so tibiotalocalcaneal arthrodesis was performed. Functional outcomes were also improved after bone grafting in all variables at the latest follow-up (p < 0.05). The mean incorporation rate of the grafts according to the location of the cysts was 84.8% (55.2% to 96.1%) at the medial malleolus, 65.1% (27.6% to 97.1%) at the tibia, and 81.2% (42.8% to 98.7%) at the talus. Smoking was identified as a significant predisposing factor adversely affecting graft incorporation (p = 0.001). Conclusion: Bone grafting for periprosthetic bone cysts following primary TAA is a reliable procedure with a satisfactory success rate and functional outcomes. Regular follow-up, including CT scan, is important for the detection of cyst re-progression to prevent implant loosening after bone grafting.

Identification of L5 vertebra on lumbar spine radiographs using deep learning.

OBJECTIVE: Deep learning is an advanced machine-learning approach that is used in several medical fields. Here, we developed a deep learning model using an object detection algorithm to identify the L5 vertebra on anteroposterior lumbar spine radiographs, and assessed its detection accuracy. METHODS: We retrospectively recruited 150 participants for whom both anteroposterior whole-spine and lumbar spine radiographs were available. The anteroposterior lumbar spine radiographs of these patients were used as the input data. Of the 150 images, 105 (70%) were randomly selected as the training set, and the remaining 45 (30%) were assigned to the validation set. YOLOv5x, of the YOLOv5 family model, was used to detect the L5 vertebra area. RESULTS: The mean average precisions 0.5 and 0.75 of the trained L5 detection model were 99.2% and 96.9%, respectively. The model's precision was 95.7% and its recall was 97.8%. Furthermore, 93.3% of the validation data were correctly detected. CONCLUSION: Our deep learning model showed an outstanding ability to identify L5 vertebrae.

The applicability of noncontact sensors in the field of rehabilitation medicine.

A noncontact sensor field is an innovative device that can detect, measure, or monitor physical properties or conditions without direct physical contact with the subject or object under examination. These sensors use a variety of methods, including electromagnetic, optical, and acoustic technique, to collect information about the target without physical interaction. Noncontact sensors find wide-ranging applications in various fields such as manufacturing, robotics, automobiles, security, environmental monitoring, space industry, agriculture, and entertainment. In particular, they are used in the medical field, where they provide continuous monitoring of patient conditions and offer opportunities in rehabilitation medicine. This article introduces the potential of noncontact sensors in the field of rehabilitation medicine.

Extracellular Vesicles from Human Adipose Tissue-Derived Mesenchymal Stem Cells Suppress RANKL-Induced Osteoclast Differentiation via miR122-5p.

Researchers are increasingly interested in cell therapy using mesenchymal stem cells (MSCs) as an alternative remedy for osteoporosis, with fewer side effects. Thus, we isolated and characterized extracellular vesicles (EVs) from human adipose tissue-derived MSCs (hMSCs) and investigated their inhibitory effects on RANKL-induced osteoclast differentiation. Purified EVs were collected from the supernatant of hMSCs by tangential flow filtration. Characterization of EVs included typical evaluation of the size and concentration of EVs by nanoparticle tracking analysis and morphology analysis using transmission electron microscopy. hMSC-EVs inhibited RANKL-induced differentiation of bone marrow-derived macrophages (BMDMs) into osteoclasts in a dose-dependent manner. F-actin ring formation and bone resorption were also reduced by EV treatment of osteoclasts. In addition, EVs decreased RANKL-induced phosphorylation of p38 and JNK and expression of osteoclastogenesis-related genes in BMDMs treated with RANKL. To elucidate which part of the hMSC-EVs plays a role in the inhibition of osteoclast differentiation, we analyzed miRNA profiles in hMSC-EVs. The results showed that has-miR122-5p was present at significantly high read counts. Overexpression of miR122-5p in BMDMs significantly inhibited RANKL-induced osteoclast differentiation and induced defects in F-actin ring formation and bone resorption. Our results also revealed that RANKL-induced phosphorylation of p38 and JNK and osteoclast-specific gene expression was decreased by miR122-5p transfection, which was consistent with the results of hMSC-EVs. These findings suggest that hMSC-EVs containing miR122-5p inhibit RANKL-induced osteoclast differentiation via the downregulation of molecular mechanisms and could be a preventive candidate for destructive bone diseases.

Predictors of Vertebral Endplate Fractures after Oblique Lumbar Interbody Fusion.

Background: Cage subsidence after oblique lumbar interbody fusion (OLIF) induces restenosis and adversely affects patient outcomes. Many studies have investigated the causes of subsidence, one of which is endplate fracture (EF). This study aimed to identify predictors of EF after OLIF. Methods: This retrospective study reviewed consecutive patients who underwent OLIF at a single institute between August 2019 and February 2022. A total of 104 patients were enrolled. The patients' demographic data and surgical details were collected through chart reviews. Radiographic variables were measured. Related variables were also analyzed using binomial logistic regression, dividing each group into those with versus without EF. Results: EF occurred at 30 of 164 levels (18.3%), and the binary logistic analysis revealed that sex (odds ratio [OR], 11.07), inferior endplate concave depth (OR, 1.95), disc wedge angle (OR, 1.22), lumbar lordosis (OR, 1.09), pelvic incidence (OR, 1.07), sagittal vertical axis (OR, 1.02), sacral slope (OR, 0.9), L3-4 level (OR, 0.005), and L4-5 level (OR, 0.004) were significantly related to EF. Conclusions: OLIF in older Asian patients should be performed carefully after recognizing the high possibility of EF and confirming the factors that should be considered preoperatively.

Effect of diabetes mellitus on the outcomes of total ankle arthroplasty: is controlled diabetes mellitus a risk factor?

BACKGROUND: It is still uncertain whether diabetes mellitus (DM) is a risk factor for poor outcomes and increased complications after total ankle arthroplasty (TAA). The objective of this study was to compare clinical outcomes and complication rates of TAA in patients with and without DM. METHODS: This study enrolled patients with symptomatic end-stage ankle osteoarthritis with a minimum follow-up period of 24 months after TAA. A total of 252 patients (266 ankles) were classified into two groups according to the presence of DM: (1) DM group (59 patients, 67 ankles) and (2) non-DM group (193 patients, 199 ankles). We defined controlled diabetes as (1) HbA1c level < 7.0%, or (2) fasting glucose level < 130 mg/dL with HbA1c level ≥ 7.0% for hospitalization period. Clinical outcomes data (Ankle Osteoarthritis Scale, American Orthopedic Foot and Ankle Society ankle-hindfoot score, Short Form-36 Physical Component Summary score, and visual analog scale for pain) were compared preoperatively and at the final follow-up between the two groups. Complications following TAA were also compared between the two groups. RESULTS: All clinical variables had improved in both groups by the final follow-up (mean follow-up = 77.8 months). There was no significant difference in any clinical variable between the two groups at the final follow-up (P > 0.05). Of the 266 ankles, 73 ankles (19 in the DM group, 54 in the non-DM group) developed periprosthetic osteolysis. Although the DM group showed a higher prevalence of aseptic loosening or subsidence, the difference between the two groups was not statistically significant (P = 0.236). CONCLUSIONS: In the intermediate-term follow-up, TAA in patients with controlled DM showed clinical outcomes and complication rates comparable to patients without DM. Our results suggest that TAA can be done safely in diabetic patients if the DM is controlled in the perioperative period. LEVEL OF EVIDENCE: Therapeutic Level III.

Progressive unilateral leg weakness after lumbar decompression due to ischemic monomelic neuropathy misdiagnosed as epidural hematoma: A CARE-compliant case report.

RATIONALE: Ischemic monomelic neuropathy (IMN) is a disease that occurs after acute arterial occlusion or steal phenomenon in an extremity that results in single or multiple axonal mononeuropathies in the distal limb without the classical features of limb ischemia, including a skin color change, limb swelling, and ischemic claudication. IMN can easily be misdiagnosed as any other neuropathic disorder. Here we present a case of IMN that was misdiagnosed as spinal epidural hematoma. PATIENT CONCERNS: A 77-year-old man presented with sudden motor weakness and pain in his left foot and calf 5 days after a bilateral L4 to 5 posterior decompression for lumbar spinal stenosis. His symptoms progressed over the next 5 days. The strengths of the left ankle dorsiflexors, first toe extensors, and ankle plantar flexors were Medical Research Council 0. On brain and whole-spine magnetic resonance imaging, no specific abnormalities correlated with his symptoms were observed. Computed tomography angiography of the lower extremities revealed segmental occlusion of the left common femoral artery and multifocal severe stenoses in the bilateral anterior and posterior tibial arteries of the left leg. No skin color change or swelling was observed in the left lower extremity. DIAGNOSIS: Based on his clinical features and imaging findings, he was diagnosed with IMN. INTERVENTION: The patient underwent thrombectomy of the left femoral artery. OUTCOMES: After the treatment, his pain almost completely disappeared. LESSONS: When patients exhibit acute-onset pain in the unilateral limb with or without motor weakness but no correlated abnormality on spinal magnetic resonance imaging or computed tomography, clinicians should consider the possibility of IMN.

Therapeutic potential of small extracellular vesicles derived from mesenchymal stem cells for spinal cord and nerve injury.

Neural diseases such as compressive, congenital, and traumatic injuries have diverse consequences, from benign mild sequelae to severe life-threatening conditions with associated losses of motor, sensory, and autonomic functions. Several approaches have been adopted to control neuroinflammatory cascades. Traditionally, mesenchymal stem cells (MSCs) have been regarded as therapeutic agents, as they possess growth factors and cytokines with potential anti-inflammatory and regenerative effects. However, several animal model studies have reported conflicting outcomes, and therefore, the role of MSCs as a regenerative source for the treatment of neural pathologies remains debatable. In addition, issues such as heterogeneity and ethical issues limited their use as therapeutic agents. To overcome the obstacles associated with the use of traditional agents, we explored the therapeutic potentials of extracellular vesicles (EVs), which contain nucleic acids, functional proteins, and bioactive lipids, and play crucial roles in immune response regulation, inflammation reduction, and cell-to-cell communication. EVs may surpass MSCs in size issue, immunogenicity, and response to the host environment. However, a comprehensive review is required on the therapeutic potential of EVs for the treatment of neural pathologies. In this review, we discuss the action mechanism of EVs, their potential for treating neural pathologies, and future perspectives regarding their clinical applications.

High reliability and accuracy of dynamic magnetic resonance imaging in the diagnosis of cervical Spondylotic myelopathy: a multicenter study.

BACKGROUND: Cervical spondylotic myelopathy (CSM) is a critical condition that results in significant neurologic deterioration. An accurate diagnosis is essential for determining its outcome and prognosis. The pathology is strongly associated with dynamic factors; therefore, dynamic magnetic resonance (MR) image could be crucial to accurately detect CSM. However, very few studies have evaluated the reliability and accuracy of dynamic MR in CSM. In this study, we aimed to compare intra- and interobserver reliabilities and accuracy of dynamic MR in detecting CSM using sagittal MR scans of the neck in the flexed, neutral, and extended position. METHODS: Out of 131 patients who underwent surgical treatments for CSM, 107 were enrolled in this study. The patient underwent three-types of sagittal MR scans that were obtained separately in different neck positions (neutral, flexion, and extension postures). The MR scans of the cervical spine were evaluated independently by three spine professionals, on the basis of tabled questionnaires. For accuracy, we performed a receiver operator characteristic analysis, and the overall discriminating ability of each method was measured by calculating the area under the ROC curve. The Cohen's kappa coefficient and the Fleiss-generalized kappa coefficient was used to the inter- and intra-observer reliabilities. RESULTS: The intraobserver reliability (using the Cohen's kappa coefficient) and interobserver reliability (using the Fless kappa coefficient) were respectively 0.64 and 0.52 for the neutral sagittal MR. The accuracy of neutral sagittal MR in detecting CSM was 0.735 (95% CI, 0.720 to 0.741) while that of extension sagittal MRI was 0.932 (96% CI, 0.921 to 0.948). CONCLUSIONS: Dynamic MR significantly showed better diagnostic reliability and accuracy in detecting CSM compared to conventional MR. In particular, extension MR scans could provide a more accurate diagnosis than other images.

Diagnostic Technology for Spine Pathology.

Diagnostic techniques for spinal pathologies have been developed in accordance with advances in technology. Accurate diagnosis of spinal pathology is essential for appropriate management of spinal diseases. Since the development of X-rays in 1895 and computed tomography (CT) in 1967, several diagnostic imaging modalities have been utilized for detecting spinal pathologies, including radiography, CT, magnetic resonance imaging, and radionuclide imaging. In addition to diagnostic imaging technologies, electrodiagnostic tests, including electromyography and nerve conduction studies, play a significant role as diagnostic tools, as spinal diseases are mostly profoundly associated with pathologies of the neural structures, such as the spinal cord and nerve root, and extent of injury at the structure cannot be adequately detected by conventional imaging techniques. In patient-specific treatment strategies, usage of diagnostic modalities is of great importance; thus, we should be aware of the basic details and approaches of the different diagnostic modalities. In this review, the authors discuss the details of the technologies that aid in the diagnosis of spinal pathologies.

Periprosthetic Osteolysis as a Risk Factor for Revision After Total Ankle Arthroplasty: A Single-Center Experience of 250 Consecutive Cases.

BACKGROUND: Periprosthetic osteolysis after total ankle arthroplasty (TAA) is a challenging problem. This study aimed to evaluate the prevalence of and predisposing factors for osteolysis and its effects on clinical outcomes. METHODS: We enrolled 236 patients (250 ankles) who underwent primary TAA using a mobile-bearing HINTEGRA prosthesis, with a mean follow-up of 83.5 months (range, 36 to 182 months), and subsequently divided them into 2 groups: the osteolysis group (79 ankles) and non-osteolysis group (171 ankles). Clinical and radiographic outcomes were compared between the 2 groups, and a bivariable logistic regression analysis was performed to identify predisposing factors for the development of osteolysis. RESULTS: In the osteolysis group (31.6% of the 250 ankles), the mean time of detection was 28.8 months postoperatively. Forty of these ankles were closely monitored without surgical treatment. Another 29 ankles underwent bone grafting and exchange of the polyethylene inlay, and the remaining 10 ankles underwent revision TAA or arthrodesis. All clinical outcome variables were significantly lower in patients with osteolysis, compared with those without osteolysis, at the final follow-up (p < 0.05). In the investigation of predisposing factors, only rheumatoid arthritis was identified as having a significant association with an increased prevalence of osteolysis (p = 0.030). CONCLUSIONS: This study demonstrated that the prevalence of periprosthetic osteolysis after TAA was considerable and that the development of osteolysis negatively affected the clinical outcome. Therefore, the prevention and appropriate treatment of osteolysis are crucial for the satisfactory long-term survival of TAA. LEVEL OF EVIDENCE: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.

Older Korean men with inadequate vitamin D status have lower odds of radiologic osteoarthritis.

Most studies on osteoarthritis (OA) and vitamin D status were performed in Whites with relatively adequate vitamin D status. Associations may differ by baseline 25-hydroxyvitamin D (25(OH)D) and race. We assessed the odds of OA and joint pain according to vitamin D status in Korean adults ≥ 50 years of age in the nationally representative Korea National Health and Nutrition Examination Survey (n = 8575). Agreement between radiologic OA (ROA) and self-reported OA were also assessed. Multivariate logistic regression was performed and participants were stratified by sex. Adults with serum 25(OH)D < 12 ng/mL and 12 to < 20 ng/mL had 26% and 18% lower odds of knee ROA, respectively, compared to those with 25(OH)D ≥ 20 ng/mL. Similar results were observed in men, but not women. No associations were found between 25(OH)D and knee ROA severity, lumbar spine ROA, symptomatic OA, or knee pain. Sensitivity of self-reported OA was low (27%), indicating a weak possibility of reverse causation. Prospective studies are required to identify the possible causality of vitamin D on OA in Korean men.

Assessment of Fusion After Anterior Cervical Discectomy and Fusion Using Convolutional Neural Network Algorithm.

BACKGROUND: A convolutional neural network (CNN) is a deep learning (DL) model specialized for image processing, analysis, and classification. OBJECTIVE: In this study, we evaluated whether a CNN model using lateral cervical spine radiographs as input data can help assess fusion after anterior cervical discectomy and fusion (ACDF). STUDY DESIGN: Diagnostic imaging study using DL. PATIENT SAMPLE: We included 187 patients who underwent ACDF and fusion assessment with postoperative one-year computed tomography and neutral and dynamic lateral cervical spine radiographs. OUTCOME MEASURES: The performance of the CNN-based DL algorithm was evaluated in terms of accuracy and area under the curve. MATERIALS AND METHODS: Fusion or nonunion was confirmed by cervical spine computed tomography. Among the 187 patients, 69.5% (130 patients) were randomly selected as the training set, and the remaining 30.5% (57 patients) were assigned to the validation set to evaluate model performance. Radiographs of the cervical spine were used as input images to develop a CNN-based DL algorithm. The CNN algorithm used three radiographs (neutral, flexion, and extension) per patient and showed the diagnostic results as fusion (0) or nonunion (1) for each radiograph. By combining the results of the three radiographs, the final decision for a patient was determined to be fusion (fusion ≥2) or nonunion (fusion ≤1). By combining the results of the three radiographs, the final decision for a patient was determined as fusion (fusion ≥2) or nonunion (nonunion ≤1). RESULTS: The CNN-based DL model demonstrated an accuracy of 89.5% and an area under the curve of 0.889 (95% confidence interval, 0.793-0.984). CONCLUSION: The CNN algorithm for fusion assessment after ACDF trained using lateral cervical radiographs showed a relatively high diagnostic accuracy of 89.5% and is expected to be a useful aid in detecting pseudarthrosis.