No abstract

A deep learning network was developed to determine Risser stage on adolescent pelvic radiographs. The network had similar accuracy to expert readers, and thus could be implemented to aid physicians to provide a second opinion on staging. Key PointsThe developed deep learning method to automate Risser stage assessment reached 78.0% accuracy, which was comparable to 74.5% agreement between expert readers. Risser stage assessment using deep learning models is promising for the evaluation of skeletal maturity in AIS and could reduce the propagation of error biases within clinical files.

Study Design Observational Cohort study. Objectives We aim to document the abandon and irregular compliance rate towards brace treatment during the COVID-19 pandemic and its impact on AIS progression. Methods We reviewed a database of AIS patients recruited between March and September 2020. We included AIS patients under brace treatment according to SRS criteria. The patients were divided in 2 cohorts: those with self-reported Good-Compliance (GC) to treatment and those who had a Bad-Compliance (BC). Data analysis included biometric and radiographic data at first visit and last follow-up and percentage of progression. Unpaired student-t tests and Chi2 were used for comparison. Results 152 patients met inclusion criteria. 89 patients (age:12.1y.o.±1.4) reported good adherence to treatment, while 63 patients (age:12.7y.o.±1.8) were not compliant. Within the BC group, 18 patients reported irregular brace wear, while 45 had completely abandoned treatment (abandon rate of 29%). The GC cohort started treatment with a mean main thoracic (MT) curve of 26° and finished with 27°. The mean difference between measurements was +.65°±7.5; mean progression rate was -4.6%. However, the BC cohort started with a mean MT curve of 27° and finished with 32°, with a mean increase of +5°±8 and a mean progression rate of -13%. The differences between the 2 cohorts were statistically significant ( P = .0002). Six patients from the BC group progressed and were offered surgery. Conclusion The abandon rate of brace treatment in AIS significantly increased during the first wave of COVID-19 pandemic. Patients who voluntarily discontinued treatment had significant increases in curve progression and surgical indication rates. Level of evidence III