Computer-aided Measurement System Using Image Processing for Measuring Cobb Angle in Scoliosis

Abstract

Background: One of the spine deformities is scoliosis, and Cobb angle is generally used to assess it. Objectives: In this study, a computer-aided measurement system (CAMS) was presented as a new repeatable and reproducible approach to assess the Cobb angle in idiopathic scoliosis patients. Methods: Python libraries, including OpenCV and Numpy were used for image processing and design of the software. To assess the repeatability and reproducibility of the CAMS, a series of 98 anterior-posterior radiographs from patients with idiopathic scoliosis were used. Assessments were done by five independent observers. Each radiograph was assessed by each observer three times with a minimum break of two weeks among assessment. The single measure intraclass correlation coefficient (ICC), the mean absolute difference (MAD), and the standard error measurement (SEM) values were used for intra- and inter-observer reliability. Results: The inter-observer analysis indicated that the ICCs ranged from 0.94 - 0.99, and the MAD between manual and CAMS were less than 3°. For intra-observer measurements, the combined SEM between all observers for the manual and CAMS was 1.79° and 1.27°, respectively. An ICC value of 0.97 with 95% confidence interval (CI) was excellent in CAMS for inter-observer reliability. The MAD of CAMS was 2.18 ± 2.01 degrees. Conclusions: The CAMS is an effective and reliable approach for assessing scoliotic curvature in the standing radiographs of thoraco-lumbar. Moreover, CAMS can accelerate clinical visits, and its calculation results are reliable.

Reliability Analysis of Vertebral Landmark Labelling on Lumbar Spine X-ray Images

“…Although there is no direct comparison due to the lack of individual UVEL and LVEL analysis studies on each lumbar vertebra, we aim to compare our results with those of similar studies, such as Cobb angle. The variations in the angles of the UVEL and LVEL in the AP view of this study are comparable with Moftian et al [ 30 ]. In their study, mean variations were 2.96° ± 2.13° and 2.18° ± 2.01° for manual labelling and computer-aided measurement system, respectively.…”

“…Our measurements ranged from 0.577° ± 0.503° to 0.979° ± 0.871° ( Table 4 ). Since the target task is different, it is hard to say that our result is superior than [ 30 ]. Cobb angle studies are conducted with whole spine images.…”

“…A large set (1000 pairs) of AP and LAT view L-spine images were prepared, and 12 raters participated in the labelling process. Most previous studies used fewer than 100 images, and the number of raters varied from 2 to 10 [ 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 ]. To the best of our knowledge, our study uses the largest dataset among reliability studies of spinal landmarks to date.…”

“…One of the most popular quantitative spinal landmark analyses is vertebral angle assessment, including the Cobb angle [ 24 ]. Many previous studies [ 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 ] have reported the inter- and intrarater reliability of angle assessment by means of intraclass correlation coefficients [ 32 ]. Cobb angle assessment is the most popular application in this field; however, it is difficult to find studies utilizing landmark point locations.…”

Development of a CapsNet and Fuzzy Logic Decision Support System for Diagnosing the Scoliosis and Planning Treatments via Schroth Method

“…When the literature is examined, it is seen that there are different studies in terms of diagnosing the scoliosis. Some of these studies are based on using traditional Machine Learning techniques or some computational algorithms [24][25][26][27][28][29][53][54][55].…”