Validating a postural evaluation method developed using a Digital Image-based Postural Assessment (DIPA) software

• Published in: Computer methods and programs in biomedicine Vol. 108; no. 1; pp. 203 - 212
• Main Authors: Furlanetto, Tássia Silveira, Candotti, Cláudia Tarragô, Comerlato, Tatiana, Loss, Jefferson Fagundes
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ireland Ltd 01.10.2012; Elsevier
• Subjects: Adult; Applied sciences; Artificial intelligence; Assessments; Biological and medical sciences; Classification; Computer programs; Computer science; control theory; systems; Digital; Digital photographs; Exact sciences and technology; Female; Fundamental and applied biological sciences. Psychology; Humans; Image Processing, Computer-Assisted; Instruments, apparatus, components and techniques common to several branches of physics and astronomy; Internal Medicine; Investigative techniques, diagnostic techniques (general aspects); Male; Medical sciences; Middle Aged; Osteoarticular system. Muscles; Other; Pattern recognition. Digital image processing. Computational geometry; Physics; Posture; Radionuclide investigations; Reproducibility; Scoliosis; Skeleton and joints; Software; Spinal postural evaluation; Validating; Vertebrates: osteoarticular system, musculoskeletal system; X- and γ-ray instruments and techniques; X-rays; Spinal postural evaluation; Validating; Scoliosis; Digital photographs; Correlation; Spine; X ray radiography; Digital image; Marker; Anatomy; Digital photography; Posture; Graphics; Statistical test; Reproducibility; Classification; Frontal; Localization
• ISSN: 0169-2607; 1872-7565; 1872-7565
• DOI: 10.1016/j.cmpb.2012.03.012

To investigate (1) the accuracy of the palpatory method to identify anatomical points by comparison with the X-ray exams, (2) the validity of classifying spinal posture in the frontal plane using Digital Image-Based Postural Assessment (DIPA) software by comparison with the X-ray exams and (3) the intra and inter-evaluator reproducibility of the DIPA software. The postural assessment and X-ray examination of the spine, both in the frontal plane and standing position, were performed consecutively in 24 subjects. The postural assessment protocol consisted of: (1) palpation and the use of reflective markers containing lead to mark the spinous processes (SP) of the C7, T2, T4, T6, T8, T10, T12, L2, L4 and S2 vertebrae and; (2) acquisition of photographic records. First, the X-ray examinations were used to check the correlation between the palpated and marked SP and the true location of the SP of the vertebra in question, by assessing the distance between them. The spinal posture was classified based on the calculation of the scoliosis arrows in the DIPA (DIPA-SA). The X-ray examinations provided the scoliosis arrows (X-SA), the Cobb angles and the classification of spinal posture based on the Cobb angle. The results from the DIPA protocol were compared to those from the X-ray examination-based protocol. The statistical tests used were: (1) Kruskal–Wallis – differences in terms of the numerical distance between the markers and the anatomical landmarks, (2) Pearson's Correlation – DIPA-SA and Cobb angles, (3) Pearson's Correlation – X-SA and DIPA-SA; (4) Bland and Altman's graphic representation – X-SA and DIPA-SA, (5) Spearman's Correlation – classification of spinal posture obtained using the X-ray and DIPA protocols, (6) the intraclass correlation test (ICC) for the relationship between the DIPA-SA made by each evaluator (inter-evaluator), and (7) independent t-test to compare the data from the two evaluation days (intra-evaluator), α=0.05. There were no significant differences between the location of the anatomical points located using palpation and identified with reflective markers and the respective location of the SP as identified using X-ray exams (χ2=9.366, p=0.404). Significant correlations were found between the DIPA-SA and the Cobb angles in the dorsal (r=0.75, p<0.001) and lumbar (r=0.76, p=0.007) regions; between the DIPA-SA and the X-SA in the dorsal (r=0.79, p<0.001) and lumbar (r=0.92, p<0.001) regions and; between the classifications of posture obtained with the DIPA and X-ray protocols (r=0.804, p<0.001). Bland and Altman's representation showed agreement between DIPA-SA and X-SA for both curvatures. Significant correlations were found for the intra-evaluator test in the thoracic (r=0.99, p<0.001) and lumbar (r=0.98, p<0.001) regions; for the inter-evaluator test in the thoracic (r=0.99, p<0.001) and lumbar (r=0.88, p<0.001) regions. The results suggest that the DIPA protocol constitutes a valid simple, practical and low-cost non-invasive tool for the evaluation of the spine in the frontal plane which can be used to obtain reproducible measurements (inter and intra-evaluators).