Smartphone-based surface topography app accurately detects clinically significant scoliosis
Matthew Rohde, Marleni Albarran, Anthony A. Catanzano, Elizabeth J. Sachs, Hiba Naz, Amishi Jobanputra, Jacob Ribet, Kali Tileston, John S. Vorhies
Abstract
PURPOSE: The purpose of this study was twofold: (1) to validate the predictive capabilities of the Scoliosis Assessment App using ST technology against X-ray "ground truth" in patients being evaluated for clinically significant scoliosis; and (2) to compare the diagnostic accuracy of the App versus the commonly used scoliometer tool. METHODS: A multicenter, prospective validation study was conducted among patients with known or suspected scoliosis. The App determined an Asymmetry Index to predict the likelihood of clinically significant disease (MCM ≥ 20°) as determined by X-ray. Outcomes included the sensitivity, specificity, and area under the receiver operating characteristic curve (ROC AUC) associated with the Apps prediction of clinically significant disease. RESULTS: Fifty-five patients were evaluated with a mean age of 13.6 ± 2.1 years. The App correctly classified 91% (50/55) of the patients compared to 69% (38/55) for the scoliometer. The sensitivity of the App was 96.4% (89.6-100% CI) versus 50% (28.1-71.9% CI) for the scoliometer (P < 0.05), while the specificity values were 85.2% (71.8-98.9% CI) and 88.9% (74.4-100% CI), respectively. ROC analysis indicated a statistically significant difference in accuracy (AUC) in favor of the App (95% versus 71%; P = 0.015). CONCLUSION: The Scoliosis Assessment App using ST technology offers an accurate, accessible, and non-ionizing method of detecting clinically significant scoliosis, suggesting that the App can be used for detection and monitoring as an alternative to radiography and as a replacement for scoliometer without diminishing the standard of care. Further studies are required to assess variations of sensitivity in a large cohort of patients and clinical utility as an alternative to radiographs.