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.

Lateral Deviation of the Hyoid Bone and Thyroid Cartilage Influences Prevertebral Soft-Tissue Swelling and Dysphagia After Anterior Cervical Diskectomy and Fusion.

BACKGROUND AND OBJECTIVES: Prevertebral soft-tissue swelling (PSTS) after anterior cervical diskectomy and fusion (ACDF) is known to be influenced by several factors. We considered the effect of lateral deviation on the traction force and attempted to find a relationship with the PSTS. This study was designed to evaluate the preoperative lateral deviation of the hyoid bone and thyroid cartilage and its effect on PSTS, airway collapse, and clinical outcomes after ACDF. METHODS: Preoperative lateral deviations of the hyoid bone and thyroid cartilage at the superior cornu and inferior cornu were measured. To assess the effect of lateral deviation, patients who underwent 1 or 2 level ACDF with the left-sided approach were divided into a deviation group (left-sided deviation >5 mm or >10 mm) and a nondeviation group (left-sided deviation <5 mm or <10 mm). Difference of preoperative and postoperative PSTS (dPSTS), airway collapse, dysphagia score, and Neck Disability Index were compared between the 2 groups. RESULTS: Lateral deviation was measured in 290 patients, and 145 were enrolled to assess the effect of lateral deviation. Left-sided deviation was more common than right-sided deviation in all 3 structures (the hyoid bone, superior cornu, and inferior cornu of the thyroid cartilage). The deviation group demonstrated a significantly larger dPSTS at the C3 and C4 levels, more airway collapse at the C4 level, and a higher dysphagia score. There was no significant difference in the Neck Disability Index between the 2 groups. Lateral deviation significantly correlated with dPSTS (C3, C4, C5, and C6 levels) and airway collapse (C3 and C4 levels). CONCLUSION: A left-sided deviation of more than 5 mm of the hyoid bone or thyroid cartilage discouraged the left-sided approach for ACDF because of the aggravation of dPSTS, airway collapse, and dysphagia postoperatively.

Preoperative Radiographic Simulation for Partial Uncinate Process Resection during Anterior Cervical Discectomy and Fusion to Achieve Adequate Foraminal Decompression and Prevention of Vertebral Artery Injury.

STUDY DESIGN: Retrospective radiographic study. PURPOSE: This study aims to demonstrate the proper resection trajectory of a partial posterior uncinate process resection combined with anterior cervical discectomy and fusion (ACDF) and evaluate whether foraminal stenosis or uncinate process degeneration increases the risk of vertebral artery (VA) injury. OVERVIEW OF LITERATURE: Appropriate resection trajectory that could result in sufficient decompression and avoid vertebral artery injury is yet unknown. METHODS: We retrospectively reviewed patients who underwent cervical magnetic resonance imaging and computed tomography angiography for preoperative ACDF evaluation. The segments were classified according to the presence of foraminal stenosis. The height, thickness, anteroposterior length, horizontal distance from the uncinate process to the VA, and vertical distance from the uncinate process baseline to the VA of the uncinate process were measured. The distance between the uncinate anterior margin and the resection trajectory (UAM-to-RT) was measured. RESULTS: There were no VA injuries or root injuries among the 101 patients who underwent ACDF (163 segments, mean age of 56.3±12.2). Uncinate anteroposterior length was considerably longer in foramens with foraminal stenosis, whereas uncinate process height, thickness, and distance between the uncinate process and VA were not significantly associated with foraminal stenosis. There were no significant differences in radiographic parameters based on uncinate degeneration. The UAM-to-RT distances for adequate decompression were 1.6±1.4 mm (range, 0-4.8 mm), 3.4±1.7 mm (range, 0-7.1 mm), 4.0±1.7 mm (range, 0-9.0 mm), and 4.5±1.2 mm (range, 2.5-7.5 mm) for C3-C4, C4-C5, C5-C6, and C6-C7, respectively. CONCLUSIONS: More than half of the uncinate process in the anteroposterior plane should be removed for adequate neural foramen decompression. Foraminal stenosis or uncinate degeneration did not alter the relative anatomy of the uncinate process and the VA and did not impact VA injury risk.