An artificial intelligence powered platform for auto-analyses of spine alignment irrespective of image quality with prospective validation

• Published in: EClinicalMedicine 2022-01, Vol.43, p.101252-101252, Article 101252
• Main Authors: Meng, Nan, Cheung, Jason P.Y., Wong, Kwan-Yee K., Dokos, Socrates, Li, Sofia, Choy, Richard W., To, Samuel, Li, Ricardo J., Zhang, Teng
• Format: Article
• Language: English
• Subjects: Adolescent idiopathic scoliosis; Artificial intelligence; Deep learning; Open platform; Spine malalignment
• ISSN: 2589-5370; 2589-5370
• DOI: 10.1016/j.eclinm.2021.101252

Assessment of spine alignment is crucial in the management of scoliosis, but current auto-analysis of spine alignment suffers from low accuracy. We aim to develop and validate a hybrid model named SpineHRNet+, which integrates artificial intelligence (AI) and rule-based methods to improve auto-alignment reliability and interpretability. From December 2019 to November 2020, 1,542 consecutive patients with scoliosis attending two local scoliosis clinics (The Duchess of Kent Children's Hospital at Sandy Bay in Hong Kong; Queen Mary Hospital in Pok Fu Lam on Hong Kong Island) were recruited. The biplanar radiographs of each patient were collected with our medical machine EOS™. The collected radiographs were recaptured using smartphones or screenshots, with deidentified images securely stored. Manually labelled landmarks and alignment parameters by a spine surgeon were considered as ground truth (GT). The data were split 8:2 to train and internally test SpineHRNet+, respectively. This was followed by a prospective validation on another 337 patients. Quantitative analyses of landmark predictions were conducted, and reliabilities of auto-alignment were assessed using linear regression and Bland-Altman plots. Deformity severity and sagittal abnormality classifications were evaluated by confusion matrices. SpineHRNet+ achieved accurate landmark detection with mean Euclidean distance errors of 2·78 and 5·52 pixels on posteroanterior and lateral radiographs, respectively. The mean angle errors between predictions and GT were 3·18° and 6·32° coronally and sagittally. All predicted alignments were strongly correlated with GT (p  0·97), with minimal overall difference visualised via Bland-Altman plots. For curve detections, 95·7% sensitivity and 88·1% specificity was achieved, and for severity classification, 88·6–90·8% sensitivity was obtained. For sagittal abnormalities, greater than 85·2–88·9% specificity and sensitivity were achieved. The auto-analysis provided by SpineHRNet+ was reliable and continuous and it might offer the potential to assist clinical work and facilitate large-scale clinical studies. RGC Research Impact Fund (R5017–18F), Innovation and Technology Fund (ITS/404/18), and the AOSpine East Asia Fund (AOSEA(R)2019–06).