Advancing Use of DEXA Scans to Quantitatively and Qualitatively Evaluate Lateral Spinal Curves, for Preliminary Identification of Adolescent Idiopathic Scoliosis

This study used data from the Raine Study (www.rainestudy.org.au). This longitudinal cohort recruited 2868 mothers (generation one) and their children (generation two) born between 1989 and 1991, with prospective collection of a wide array of biological, psychological, and social variables. The Raine Study has been demonstrated to be an unbiased sample, representing the general population in Western Australia [10]. This study draws on DEXA scans collected at age 20 years (2009–2012), and International Classification of Diseases, Ninth Revision (ICD-9), code number 737.3 (idiopathic scoliosis/kyphoscoliosis) data collected at ages 1–20 years. This study was approved by the Raine Study Scientific Management Committee, ethics approval was obtained from the Human Research Ethics Committee of The University of Western Australia, and participants provided their own informed consent.

Participants who attended for a DEXA scan at age 20 years were included (n = 1238, females = 604, males = 634). The Norland XR-36 densitometer (Norland Medical Systems, Inc., Fort Atkinson, WA, USA) provided images of 96dpi with an image size of 280 × 730pixels, to measure lean and total fat mass. The research staff guided the participant’s upper and lower body to be in line with the pelvis while lying supine on the scanning table. Owing to skeletal maturity being likely at age 20, any scoliotic curves identified were expected to be at their maximum deformity (curve progression which is known to occur during skeletal growth has ceased) [12].

Quantitative Measurement of Spinal Curve from DEXA Images

The low resolution of DEXA scan images means that it is not considered a diagnostic tool for identifying AIS. Formal diagnosis of AIS with radiographs or other advanced imaging typically employs either the Cobb method, which takes reference from the most inclined vertebral end plates of the major curve [13], or the Ferguson method, which takes reference from axial rotation of segments [14]. For preliminary identification and measurement of scoliosis on DEXA scans, the modified Ferguson method uses a “normal spine line” through the T1 to L5 vertebrae as a reference to determine the center of the upper and lower end vertebrae of a scoliotic curve, with lines connecting to the apex of the curve creating the measured angle [11]. Employing the modified Ferguson method in the current study, DEXA images were viewed at 300–400% magnification and marked with an image processing application (Pixelmator, version 3.3, Lithuania), as detailed in Fig. 1. At the intersection of the lines connecting the upper vertebra to apex and the lower vertebra to the apex, the angle β was measured with the ImageJ (version 2.0.0, National Institute of Health, USA) angle tool. The modified Ferguson angle was calculated by 180°- β. One author (AC) marked and measured all scans.

Fig. 1figure 1

A DEXA scan with lines drawn for the modified Ferguson angle, with a flowchart describing the process (steps 1–7). This image shows symmetrical positioning with no obliquity of the shoulders or pelvis, and the lateral borders of the spine are clear to see without any cardiac shadowing

Inter-Rater Reliability of DEXA Ferguson Method

To test the inter-rater reliability of marking and measuring these angles, a sample of 41 unmarked original DEXA scans, with angles ranging between 6° and 40° (as measured by examiner AC), were independently marked and measured by a second examiner (PN), who was blind to the previous measurements.

Qualitative Examination of Scoliosis Curve

An apparent curve on a DEXA scan may represent poor participant positioning for the scan, a shadow effect from internal organs, a spinal pathology other than AIS, or spinal curves consistent with AIS. Due to these potential confounders, a clinical research expert with 25 years of experience and publications in relation to AIS image analysis and clinical deformity measurement (author-MI) [15,16,17] conducted a qualitative evaluation of DEXA scans that had modified Ferguson angles ≥ 8° (n = 142). The criteria that the expert applied to qualitatively evaluate scans are provided in Table 1, and exemplar images are provided in Online Appendix.

Table 1 Qualitative criteria used by the expert reviewer for evaluation of spinal curves on DEXA images, with the aim to identify likely adolescent idiopathic scoliosisReported Diagnosis of Idiopathic Scoliosis/Kyphoscoliosis

At ages 1, 5, 8, 10, 14, 17, and 20 years, participants/parents of the Raine Study reported on their health conditions that had been diagnosed by a health professional, based on the question: “Does your child have now, or has your child had in the past, any of the following health professional diagnosed medical conditions or health problems?”. These reports were subsequently classified by Raine Study research staff using the ICD-9 [18]. Those who had DEXA scans at age 20 and reported a diagnosis of “idiopathic scoliosis/kyphoscoliosis” at ages 1, 5, and 8 years (n = 2) were excluded from the dataset for likely AIS owing to juvenile-onset rather than adolescent-onset of their condition.

Participant Grouping

To group DEXA scans as (i) “likely AIS,” (ii) “without scoliosis,” and (iii) “uncertain to have AIS,” both the quantitative modified Ferguson scoliosis curve measurements and qualitative expert evaluation of the DEXA scans were used (Fig. 2). Of note, a reported diagnosis of idiopathic scoliosis/kyphoscoliosis at any time from ages 10–20 years was also noted (Fig. 2), but not considered to be reliable for grouping participants as some participants reported diagnosis where scans showed no scoliosis curve. Rather, those identified as “likely AIS” had both a modified Ferguson angle ≥ 10°, and the expert reviewer determined that the scoliosis curves were consistent with AIS. The group “without scoliosis” had a modified Ferguson angle of < 10°, and the expert reviewer determined that any scoliosis curves between 8 and 10° were not consistent with AIS. The group that was “uncertain to have AIS” presented with either a discord between criteria for a modified Ferguson angle ≥ 10° and expert evaluation or had a reported diagnosis of AIS with neither sufficient angle or expert evaluation agreement.

Fig. 2figure 2

Consolidated standards of reporting trials (CONSORT) diagram to detail the recruitment and allocation of participants for this study. Participants were allocated into: (i) likely adolescent idiopathic scoliosis, (ii) without scoliosis, and (iii) uncertain to have adolescent idiopathic scoliosis based on their curve angle (measured modified Ferguson angle ≥ 10°), and expert evaluation. Reported diagnosis of idiopathic scoliosis/kyphoscoliosis was not part of the allocation criteria but is indicated in the diagram to determine its concordance with DEXA-identified adolescent idiopathic scoliosis. ✅: did meet criteria; ❌ did not meet criteria

Statistical Analysis

Descriptive statistics were used to report prevalence and curve characteristics. To assess the inter-rater reliability for modified Ferguson measures from DEXA scans, statistical analysis was performed using RStudio (2020) software (version 1.3.959, PBC, Boston, MA, USA). Bland–Altman plots were used to visualize the data, and intra-class correlation (ICC) with a two-way random effects model with absolute agreement were calculated. The ICC values were interpreted as poor (< 0.40), fair (0.40–0.59), good (0.60–0.74), and excellent (0.75–1.00) [19] inter-rater agreement. To assess the consistency between the two raters, the standard error of measurement (SEM) was calculated as the standard deviation of the difference in modified Ferguson angle values between the two raters divided by the √2. The SEM was then used to determine a minimum detectable change, MDC95 = 1.96 × SEM × √2 (also referred to as 95% limit of agreement). The distribution of the difference scores was assessed using the Shapiro–Wilk test. The level of significance was set at p < 0.05.

留言 (0)

沒有登入
gif