Investigating the ability to use the CT scan projection radiograph to monitor adolescent idiopathic scoliosis

Introduction: Adolescent idiopathic scoliosis (AIS) is a spinal deformity that causes the spine to bend laterally. Patients with AIS undergo frequent X-ray examinations to monitor the progression of the deformity through the measurement of the Cobb angle, increasing the risk of developing radiation-induced cancer. The aim of this study was to investigate the use of scan projection radiograph (SPR) in computed tomography (CT) to assess AIS by quantifying radiation dose from the SPR acquisitions and comparing it to those of digital radiography (DR) and a dedicated scoliosis imaging system (EOS) and by evaluating the accuracy of Cobb angle measurements on SPR images using a bespoke validated phantom. Methods: A dosimetry phantom representing a 10-year-old child and thermoluminescent dosimeters were used for measuring organ dose to calculate effective dose (ED) and effective risk (ER). Twenty-seven CT SPR protocols were used. A comparison was made to doses from imaging protocols using DR and the EOS system. The effectiveness of a scoliosis shawl for selected projections was also tested. To test the accuracy of Cobb angle measurements on SPR images, a scoliotic phantom was constructed and validated. Poly-methyl methacrylate (PMMA) and plaster of Paris (PoP) were used to represent human soft tissue and bone tissue, respectively, to construct a phantom exhibiting a 15° lateral curve of the spine. The phantom was validated by comparing the Hounsfield unit (HU) of its vertebrae with those of a human and an animal. Additionally, comparisons of signal-to-noise ratio (SNR) to those from a commercially available phantom were made. The angle of the curve in the phantom was measured directly to confirm that it was 15°. The constructed phantom was scanned in CT SPR mode, and the resulting images were visually evaluated against set criteria to determine their suitability for Cobb angle measurements. Those deemed of insufficient quality were excluded. Cobb angle measurements were then performed on the remaining images (n = 10) by 13 observers. Results: EOS had the lowest ED and ER when it was used to irradiate the phantom in AP positions. Five SPR AP imaging protocols and seven PA imaging protocols delivered significantly lower radiation dose and risk than their corresponding imaging positions in DR (p < 0.05). The scoliosis shawl significantly lowered the ED and ER of SPR and DR AP imaging protocols (p < 0.05). The validation of the PoP phantom revealed that the HU of the PoP vertebrae was 628 (SD= 56), human vertebrae was 598 (SD= 79) and sheep vertebra was 605 (SD= 83). The SNR values of the two phantoms correlated strongly (r = 0.93 [(p < 0.05]). The measured scoliosis angle was 14 degrees. When the phantom was imaged using SPR, the difference between the measured Cobb angle and the known angle was, on average, –2.75° (SD = 1.46°). The agreement among the observers was good (p = 0.861, 95% CI [0.70–0.95]) and comparable to similar studies on other imaging modalities which are used for Cobb angle estimation. Conclusion: EOS had the lowest dose. Where this technology is not available, there is a potential for organ dose (OD) reduction in AIS imaging using CT SPR compared with DR. The PoP phantom has physical characteristics (in terms of spinal deformity) and radiological characteristics (in terms of HU and SNR values) of the spine of a 10-year-old child with AIS. CT SPR images can be used for AIS assessment with the 5° margin of error that is clinically acceptable. A few SPR imaging protocols (CT4, 8 and 11) had the lower radiation risk compared with the DR and provided the most accurate Cobb angle measurements. Implications for practice: The bespoke phantom can be used to investigate new X-ray imaging techniques and technology in the assessment of scoliosis and has utility for the optimisation of X-ray imaging techniques in 10-year-old children. Overall, the outcome is promising for patients and health providers because it provides an opportunity to reduce patient dose and achieve clinically acceptable Cobb angle measurements whilst using existing CT technology