Conquering the Cobb Angle: A Deep Learning Algorithm for Automated, Hardware-Invariant Measurement of Cobb Angle on Radiographs in Patients with Scoliosis.

Scoliosis is a disease estimated to affect more than 8% of adults in the United States. It is diagnosed with use of radiography by means of manual measurement of the angle between maximally tilted vertebrae on a radiograph (ie, the Cobb angle). However, these measurements are time-consuming, limiting their use in scoliosis surgical planning and postoperative monitoring. In this retrospective study, a pipeline (using the SpineTK architecture) was developed that was trained, validated, and tested on 1310 anterior-posterior images obtained with a low-dose stereoradiographic scanning system and radiographs obtained in patients with suspected scoliosis to automatically measure Cobb angles. The images were obtained at six centers (2005-2020). The algorithm measured Cobb angles on hold-out internal (n = 460) and external (n = 161) test sets with less than 2° error (intraclass correlation coefficient, 0.96) compared with ground truth measurements by two experienced radiologists. Measurements, produced in less than 0.5 second, did not differ significantly (P = .05 cutoff) from ground truth measurements, regardless of the presence or absence of surgical hardware (P = .80), age (P = .58), sex (P = .83), body mass index (P = .63), scoliosis severity (P = .44), or image type (low-dose stereoradiographic image vs radiograph; P = .51) in the patient. These findings suggest that the algorithm is highly robust across different clinical characteristics. Given its automated, rapid, and accurate measurements, this network may be used for monitoring scoliosis progression in patients. Keywords: Cobb Angle, Convolutional Neural Network, Deep Learning Algorithms, Pediatrics, Machine Learning Algorithms, Scoliosis, Spine Supplemental material is available for this article. © RSNA, 2023.

Vertebral Deformity Measurements at MRI, CT, and Radiography Using Deep Learning.

PURPOSE: To construct and evaluate the efficacy of a deep learning system to rapidly and automatically locate six vertebral landmarks, which are used to measure vertebral body heights, and to output spine angle measurements (lumbar lordosis angles [LLAs]) across multiple modalities. MATERIALS AND METHODS: In this retrospective study, MR (n = 1123), CT (n = 137), and radiographic (n = 484) images were used from a wide variety of patient populations, ages, disease stages, bone densities, and interventions (n = 1744 total patients, 64 years ± 8, 76.8% women; images acquired 2005-2020). Trained annotators assessed images and generated data necessary for deformity analysis and for model development. A neural network model was then trained to output vertebral body landmarks for vertebral height measurement. The network was trained and validated on 898 MR, 110 CT, and 387 radiographic images and was then evaluated or tested on the remaining images for measuring deformities and LLAs. The Pearson correlation coefficient was used in reporting LLA measurements. RESULTS: On the holdout testing dataset (225 MR, 27 CT, and 97 radiographic images), the network was able to measure vertebral heights (mean height percentage of error ± 1 standard deviation: MR images, 1.5% ± 0.3; CT scans, 1.9% ± 0.2; radiographs, 1.7% ± 0.4) and produce other measures such as the LLA (mean absolute error: MR images, 2.90°; CT scans, 2.26°; radiographs, 3.60°) in less than 1.7 seconds across MR, CT, and radiographic imaging studies. CONCLUSION: The developed network was able to rapidly measure morphometric quantities in vertebral bodies and output LLAs across multiple modalities.Keywords: Computer Aided Diagnosis (CAD), MRI, CT, Spine, Demineralization-Bone, Feature Detection Supplemental material is available for this article. © RSNA, 2021.

Zone- and layer-specific differences in proteoglycan content in patellofemoral pain syndrome are detectable on T1ρ MRI.

OBJECTIVE: Determine if differences in T1ρ would be detected in specific regions or layers of patellofemoral cartilage between patients with symptomatic patellofemoral pain syndrome and asymptomatic control subjects. MATERIALS AND METHODS: Ten subjects diagnosed with patellofemoral pain syndrome were compared with ten age-, gender-, and BMI-matched control subjects with no knee pain or prior trauma. Conventional turbo (fast) spin echo sequences and T1ρ-weighted imaging were performed on the symptomatic knee in each of the ten subjects. At the patella and distal femur, cartilage regions of interest were divided into medial and lateral sub-regions, each then further sub-divided by layer (superficial, middle, or deep). Two-tailed t test and chi-squared tests were used to analyze demographic data. A mixed effect model was run for each sub-region of T1ρ imaging. Statistical significance was determined using the likelihood ratio test against reduced models without patellofemoral pain syndrome symptomatic status as a fixed effect. RESULTS: There was no difference in age, sex, or BMI between symptomatic and control patients. T1ρ values were significantly higher among patellofemoral pain syndrome patients when compared with controls in the superficial zone of the lateral patella (58.43 vs. 50.83, p = 0.03) and the middle zone of the lateral patella (52.67 vs. 43.60, p = 0.03). T1ρ was also higher in the superficial zone of the medial femur (50.94 vs. 46.70, p = 0.09) with a value approaching statistical significance. CONCLUSION: We report statistically significant differences in the T1ρ value in the superficial and middle zones of the lateral patella in patients with patellofemoral pain syndrome who had no abnormalities seen on conventional MRI sequences, suggesting an alteration the macromolecular structure of the cartilage in this population.

The Effect of Computed Tomography Scans Oriented Along the Longitudinal Scaphoid Axis on Measurements of Deformity and Displacement in Scaphoid Fractures.

PURPOSE: Reformatting computed tomography (CT) scans along the scaphoid longitudinal axis improves the ability to detect scaphoid fractures compared with reformats along the wrist axis. However, it remains unclear whether scaphoid axis reformats affect measurements of displacement or deformity, which are factors that drive the clinical decision to perform open reduction internal fixation. Our null hypothesis was that reformatting CT scans along the scaphoid axis does not affect measurements of fracture displacement and deformity. METHODS: Thirty patients with CT scans demonstrating scaphoid fractures (4 proximal pole, 17 midwaist fractures, and 9 distal) were identified and reformatted along 2 axes: the longitudinal axis of the scaphoid and the longitudinal axis of the wrist. The reformatted scans were sent to 2 musculoskeletal radiologists and 2 orthopedic hand surgeons who made the following measurements: (1) fracture gap, (2) displacement of the articular surface, (3) intrascaphoid angle, and (4) height to length (H:L) ratio. RESULTS: The reliability of each of the measurements cited above was compared for all raters between the 2 axes using intraclass correlation coefficients. Measurement of fracture gap and articular displacement trended toward more reliability in the wrist axis, whereas measurement of H:L ratio and intrascaphoid angle trended toward more reliability in the scaphoid axis. However, no differences in measurements between the 2 axes were statistically significant. CONCLUSIONS: This study demonstrates that reformatting CT scans in line with the axis of the scaphoid does not result in more reliable measurements of displacement or deformity. CLINICAL RELEVANCE: Measurements of displacement and deformity in scaphoid fractures can be made in the wrist axis with comparative reliability to those in the longitudinal scaphoid axis.

The Utility of Digital Tomosynthesis to the Practicing Orthopaedic Trauma Surgeon: A Case Series.

OBJECTIVE: To document the role of digital tomosynthesis (DTS) in the evaluation and treatment of orthopaedic trauma patients. DESIGN: Retrospective case series. SETTING: Level 1 trauma center with nonunion referral patient population. PARTICIPANTS: Four orthopaedic trauma patients with musculoskeletal injuries. INTERVENTION: Three revision surgical procedures and 1 conservative treatment for patients with periprosthetic fractures or nonunions. RESULTS: DTS successfully visualized 2 nonunions, 1 refracture, and 1 arthrodesis. MAIN OUTCOME MEASURE: Documented fracture or nonunion on imaging. CONCLUSIONS: DTS has the potential to be of significant value in the detection and follow-up of fractures. LEVEL OF EVIDENCE: Diagnostic Level IV. See Instructions for Authors for a complete description of levels of evidence.

T1rho MRI quantification of arthroscopically confirmed cartilage degeneration.

Nine asymptomatic subjects and six patients underwent T(1)rho MRI to determine whether Outerbridge grade 1 or 2 cartilage degeneration observed during arthroscopy could be detected noninvasively. MRI was performed 2-3 months postarthroscopy, using sagittal T(1)-weighted and axial and coronal T(1)rho MRI, from which spatial T(1)rho relaxation maps were calculated from segmented T(1)-weighted images. Median T(1)rho relaxation times of patients with arthroscopically documented cartilage degeneration and asymptomatic subjects were significantly different (P < 0.001), and median T(1)rho exceeded asymptomatic articular cartilage median T(1)rho by 2.5 to 9.2 ms. In eight observations of mild cartilage degeneration at arthroscopy (Outerbridge grades 1 and 2), mean compartment T(1)rho was elevated in five, but in all observations, large foci of increased T(1)rho were observed. It was determined that T(1)rho could detect some, but not all, Outerbridge grade 1 and 2 cartilage degeneration but that a larger patient population is needed to determine the sensitivity to these changes.