- Email: [email protected]
MDCT in suspected lumbar spine fracture: comparison of standard and reduced dose settings using iterative reconstruction
Clinical Radiology 73 (2018) 675.e9e675.e15
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
MDCT in suspected lumbar spine fracture: comparison of standard and reduced dose settings using iterative reconstruction J.M. Weinrich a, *, L. Well a, M. Regier a, C. Behzadi a, S. Sehner b, G. Adam a, A. Laqmani a a
Department for Diagnostic and Interventional Radiology and Nuclear Medicine, University Medical Center Hamburg-Eppendorf, Hamburg, Germany b Department of Medical Biometry and Epidemiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
art icl e i nformat ion Article history: Received 13 November 2017 Accepted 16 February 2018
AIM: To compare standard (SD-) and reduced-dose computed tomography (RD-CT) in combination with iterative reconstruction (IR) in emergency patients with suspected lumbar spine fracture. MATERIAL AND METHODS: Forty emergency patients with suspected lumbar spinal disorder who underwent RD-CT and 40 body mass index-matched patients undergoing SD-CT were enrolled in this retrospective study. Raw data for RD-CT were reconstructed using two increasing IR levels (IRL) 4 and 6, while SD-CT was reconstructed with IRL3. Two radiologists assessed image quality, image noise, and reader confidence in interpreting findings of spinal fractures in a blinded manner. RESULTS: Effective radiation dose was reduced by 50% using RD-CT. Overall subjective image quality (SIQ) was high for both protocols and slightly superior in the RD-CT protocol for both IRL compared to SD-CT. The detection rate of spinal disorders was high for both protocols with a high interobserver agreement. CONCLUSION: RD-CT with higher levels of IR results in substantial dose reduction of 50% in lumbar spine CT while maintaining an excellent subjective image quality resulting in a high diagnostic confidence. Ó 2018 Published by Elsevier Ltd on behalf of The Royal College of Radiologists.
Introduction Traumatic spinal injuries are the leading cause of death and disability in young adults,1 whereas elderly patients
* Guarantor and correspondent: J. Weinrich, Department of Radiology, University of Hamburg, Martinistraße 52, Hamburg, 20251, Germany. Tel.: þ49 (0) 152 2281 7461. E-mail address: [email protected] (J.M. Weinrich).
mainly suffer from osteoporotic fractures.2 Lumbar spine radiography remains the first radiographic examination for suspected fracture, even though there is evidence of its limited value.3 Additionally, radiographs of the lumbar spine result in considerably high radiation doses of approximately 1 mSv.4,5 Even though effective dose of lumbar spine computed tomography (CT) can be as high as 19 mSv,6 it is regarded as the imaging technique of choice for the evaluation of vertebral fractures in emergency patients.7,8
https://doi.org/10.1016/j.crad.2018.02.015 0009-9260/Ó 2018 Published by Elsevier Ltd on behalf of The Royal College of Radiologists.
675.e10
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
Due to the constant advances in CT technology, the amount of examinations has constantly risen over the last decade. Even though reduced-dose CT protocols are increasingly established in clinical practice for certain diagnostic questions as for the detection of urolithiasis or for the characterising patterns of pulmonary disease, the cumulative radiation exposure from medical examination has increased.9 In order to achieve low radiation doses, optimisation of CT protocols without sacrificing image quality is needed.10 Dose reduction is associated with increased image noise, which can impair image quality.11 Iterative reconstruction (IR) algorithms in CT enable the reduction of patient exposure by decreasing image noise.11,12 There is only scarce literature regarding dose reduction in lumbar spine CT,4,13e15 especially in the evaluation of lumbar spine fractures. To the authors’ knowledge, there have been no studies on dose reduction in evaluating emergency patients with suspected lumbar spine trauma. Therefore, the aim of the present study was to compare the effect of reduced-dose (RD) CT with IR to standard-dose (SD) CT with a lower level of IR in emergency patients with suspected lumbar spine fracture.
Materials and methods
Between March and July 2016, 40 patients who were referred to the emergency department for a clinically and radiographically indicated RD-CT of the lumbar spine for suspected fracture were included consecutively. As a reference group for radiation dose comparison, datasets of 40 body mass index (BMI; range: 18.8e40.4 kg/m2), gender, and age-matched patients submitted for suspected lumbar spine fracture who had undergone a SD-CT for suspected trauma were also included. Demographic patient data including sex, age, height, and weight, and imaging data were noted (Table 1). BMI was calculated using the formula (patient weight [kg]/patient height [m]2). Formal ethics Table 1 Demographic and CT specific baseline data.
Sex (percent female) Numbers Age (years) BMI (kg2/m) Tube currentetime product (mAs) CTDIvol (mGy) DLP (mGy$cm) Scan length (cm) Effective dose (mSv)
Imaging protocols All examinations were performed on a 256-section CT system (Brilliance iCT, Philips, Best, the Netherlands). Patients were placed in supine position with the head first on the CT table and arms above the head for both protocols. Details are given in Table 2. An established IR algorithm (iDose,4 Philips Healthcare) was used for IR.16 Within iDose,4 increasing levels represent increasing strength of noise reduction. The technical details of IR have been described elsewhere.17 Based on preliminary unpublished results of cadaver studies, which were performed at University Medical Center Hamburg-Eppendorf and the assumption of higher image noise reduction through the application of higher levels of IR,18 higher levels of IR were applied in RD-CT compared to the SD-CT protocol. Images were reconstructed using the iDose4 levels 4 (IRL4) and 6 (IRL6) for RD-CT and level 3 for SD-CT (IRL3). Images were reconstructed with a 3 mm section-thickness for qualitative image quality evaluation. All datasets were reconstructed in a bone and soft-tissue window setting in the axial, coronal, and sagittal planes.
Radiation dose
Study population
SD-CT
approval was not required for retrospective analysis of anonymised imaging data.
RD-CT
p-Value
Mean (SD)
Mean (SD)
50% 40 66.316.1 25.95.1 170.270
50% 40 66.316.1 26.14.6 67.132.2
0.826 <0.001
11.44.6 403.7188.3 25.28.3 6.22.9
6.93.3 209.2115.2 23.56.5 3.21.8
<0.001 <0.001 0.32 <0.001
The CT dose index (CTDIvol) and doseelength product (DLP) were derived from the automatically generated dose protocol of each examination. The effective dose (ED) was estimated by multiplying the DLP by a conversion factor of 0.0153 mSv/mGy∙cm for 120 kV and 0.0155 mSv/mGy∙cm for 140 kV.19 Estimated dose saving was expressed as the estimated percentage of radiation reduced between the two protocols.
Quantitative image analysis Quantitative image analysis was performed using a dedicated PACS workstation (PACS IW, GE Healthcare, Milwaukee, MI, USA) by one experienced radiologist (J.W.). Attenuation was measured in the axial plane by placing circular regions of interest (ROIs) of 100 mm2 at the centre Table 2 Imaging protocols. SD-CT
Tube currentetime product (mAs) values are given as the mean modulated values. SD-CT, standard-dose computed tomography; RD-CT, reduced-dose computed tomography (RD-CT).
Collimation (mm) Rotation time (s) Pitch Tube voltage (kV) Reference tube current (mAs) Automatic tube current modulation Level of iteration (iDose level) Kernel Section thickness (mm) Reformations Field of view SD-CT, standard-dose computed computed tomography (RD-CT).
RD-CT
21280.0625 0.5 0.993 0.579 120 140 158 70 Z-DOM 3 4 and 6 B and D 3 Axial, coronal, sagittal 250250 mm tomography;
RD-CT,
reduced-dose
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
of the abdominal aorta in each patient. To prevent bias resulting from a single ROI measurement, each region was measured at three adjacent 1-mm sections. Results were then averaged for further analyses. The standard deviation of the attenuation values served as objective image noise (OIN).
Qualitative image analysis For qualitative image analysis, data sets of 80 RD-CT (40 RD-CT examinations reconstructed with the two increasing iDose4 levels IRL4 and IRL6) and 40 SD-CT examinations (reconstructed with IRL3) were anonymised and reviewed in a blinded manner. All 120 reconstructions were randomised and three separate review sessions after a period of 14 days were performed to ensure a maximum level of blinding. Despite the attempts to blind the evaluation process regarding different reconstruction algorithms and levels, unambiguous blinding might not have been achieved, due to slightly altered image texture in different reconstructions. The images were reviewed independently by two radiologists with 4 and 5 years of experience in musculoskeletal imaging. Bone (window width, 2500 HU; window centre, 500 HU) and soft-tissue (window width, 360 HU; window centre, 60 HU) windows were used. For overall image quality, an established five-point scale was used based on subjective image quality, image noise and presence of artefacts16: 1, poor image quality, major artefacts, major image noise; 2, reduced image quality, substantial artefacts, extensive image noise; 3, acceptable image quality, moderate artefacts, moderate image noise; 4, good image quality, minor artefacts, minor image noise; and 5, excellent image quality, no artefacts, no perceived image noise. In order to evaluate visibility of anatomical structures a modified grading method based on the European Guidelines on Quality Criteria Computed Tomography was used.20 Three anatomical regions were included for analysis: (1) cortical and trabecular bone, intervertebral foramina, (2) pedicle and intervertebral joints, and (3) spinous and transverse processes. The grading system was based on a three-point scale as follows: 1, structures are well-defined with sharp contours, free of artefacts, and adequate for diagnosis; 2, structures are visible, but contours are not completely demarcated; mild artefacts can be seen, but they do not prevent diagnosis; and 3, structures are unclear with blurred anatomical boundaries, have considerable artefacts, and prevent proper diagnosis.
Diagnostic reliability Spinal disorders were independently evaluated by both readers in concordance with international standards reported elsewhere.21 Both readers evaluated the following pathologies: fracture yes/no, acute versus old fracture, spondylolisthesis, osteolytic or osteoblastic metastases. Reviewer’s confidence in detection of findings was rated on a five-point Likert scale: 1, definitely absent; 2, probably
675.e11
absent; 3, equivocal; 4, probably present; 5, definitely present.
Diagnostic validity One radiologist retrospectively retrieved the pertinent medical records of all patients and compared final clinical diagnoses with detected CT findings. The final clinical diagnosis was determined by review of the patients’ medical records in the hospital information system. The diagnosis was confirmed by review of the patient’s hospital discharge records, reports of surgical procedures, hospital discharge records, or clinical follow-up information.
Statistical analysis Sample characteristics are given as absolute and relative frequencies or mean standard deviation, whichever is appropriate. Baseline characteristics of the both groups (RD-CT versus SD-CT) were compared using the Welch’s approximation for t-test. To account for the repeated measurement structure of the data, defined by three 1 mm adjacent image sections within the aorta per patient and reconstruction level, a linear mixed effect model repeat measurement (MMRM) was calculated for the quantitative parameters. In this model, as a predictor for the quantitative parameters the reconstruction level was included. For the qualitative Likert scale an analogous model was used while the repeated measurements were defined by the ratings of the two readers. To evaluate the subjective image quality a linear mixed effect model repeat measurement was used with the interaction between reconstruction level and each subjective image quality characteristic as a predictor. To evaluate the visibility (well defined versus visible) a logistic mixed effect model repeat measurement was used with the interaction between IRL and the three different anatomical structures as a predictor. If interaction terms were insignificant, only the main effects were included. Moreover, all models were controlled for the potential confounder of BMI, age, gender, tube current, and effective dose CTDIvol. The results were reported as estimated means, which are represented in graphs with their corresponding 95% confidence intervals (95% CIs). Post hoc tests for comparison of the estimated means were calculated with contrast tests, using Wald tests. To test for interrater reliability regarding presence of a pathology Cohen’s kappa test was performed. Two sided p-values of <0.05 were considered significant. All analyses were computed using Stata 14.1 (STATA Corporation, College Station, TX, USA).
Results Radiation dose The mean exposure, CTDIvol, DLP, and effective dose in RD-CT were significantly reduced compared to SD-CT (Table 1). CTDIvol was reduced by 39.5% and effective dose
675.e12
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
was reduced by approximately 50% in the RD-CT protocol: RD-CT: 3.21.8 mGy versus SD-CT: 6.22.9 mGy.
Quantitative results Quantitative results (95% confidence interval [CI]) are graphically displayed in Fig 1. None of the confounder showed a significant effect on the quantitative parameters. The mean CT-N in SD-CT studies reconstructed with IRL3 were slightly but not significantly higher than in RD-CT reconstructions with IRL4 and IRL6 (SD-CT: IRL3: 55.9 HU; RD-CT: IRL4: 51.4 HU and IRL6: 51.7 HU). Compared to SDCT reconstructed with IRL3, objective image noise was reduced up to 20% and 37% in RD-CT reconstructed with IRL4 and IRL6 (Fig 1).
Qualitative results Mean scores (95% CIs) of qualitative image analysis are graphically displayed in Fig 2. None of the confounder showed a significant interaction with level of IR or image quality. Overall subjective image quality was high and was only slightly improved in the RD-CT IRL4 (p¼0.049) when compared to SD-CT IRL3 (Fig 2). Subjective image quality further improved with the use of RD-CT IRL6 compared to SD-CT IRL3 (p¼0.006).
Visibility of anatomical structures Overall visibility of anatomical structures was high in both protocols and a score of 3 was not given. Therefore, only scores of 1 and 2 were considered in the statistical analyses. Visibility of ‘‘cortical and the trabecular bone’’ was rated better for RD-CT IRL4 when compared to SD-CT IRL3 (p¼0.0016), while there was no difference compared to RDCT IRL6 (p¼0.238). There was no significant difference between SD-CT IRL3 and RD-CT IRL4 and IRL6 for the remaining two criteria “intervertebral foramina, pedicles and intervertebral joints” and “spinous and transverse processes” (all p>0.05; Fig 3).
Figure 2 Qualitative analysis of SD-CT IRL3 and RD-CT reconstructed with IRL4 and IRL6. Plots show mean scores for subjective image quality. Error bars represent the 95% CI. Subjective five-point grading scale (1 indicating worst through 5 indicating best).
Diagnostic confidence Overall confidence was high and there were no equivocal findings. For SD-CT, both readers could definitely assure their diagnosis in every patient (80/80). RD-CT IRL4 and IRL6 achieved 77/80 and 79/80 definite diagnoses, respectively (Table 3).
Reliability Both readers diagnosed the same amount of pathologies. Interobserver agreement for detecting pathologies was high for both protocols (SD-CT: 97.5%, kappa: 0.950.09; RD-CT IRL4 and IRL6: 100%, kappa: 10.09). There was only one disagreement (probably versus definitely present) in the SD-CT protocol regarding an old vertebrae fracture.
Figure 1 Quantitative analysis of SD-CT IRL3 and RD-CT reconstructed with IRL4 and IRL6. Error bars represent the 95% CI and points between intervals are given as mean.
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
675.e13
Figure 3 Visibility analysis of anatomical structures of SD-CT reconstructed with IRL3 and RD-CT reconstructed with IRL4 and IRL 6. The plots show the mean probability of visibility (error bars represent the 95% CI).
CT findings and validity Twenty-four out of 40 patients in the SD-CT group and 26/40 patients in the LDCT group had an acute vertebral fracture (Fig 4), whereas 21/80 did not have any spinal disorders. Four of 80 patients revealed old fractures and 3/ 80 had osteolytic or osteoblastic lesions. There were only 2/ 80 cases of spondylolisthesis. Seventeen of 50 patients with acute fractures also had at least one old fracture.
Discussion In the present study, the effective radiation dose to lumbar spine was reduced by approximately 50% in patients with suspected fracture who underwent MDCT, while at the same time maintaining high diagnostic confidence and subjective image quality. The present results show that a RD-CT protocol combined with higher IR levels does not only provide an image quality that is at least on par with SDCT with lower levels of IR, but also enables an equivalent and even better visualisation of anatomical structures. Especially, the visibility of cortical and trabecular bone was rated better for RD-CT IRL4 when compared to SD-CT IRL3. Higher levels of IR improved subjective image quality of almost all diagnostic criteria analysed within the RD-CT protocol when compared to the SD-CT. In accordance with discharge records and further treatment both readers detected all pathologies and showed a very high confidence in evaluating both, RD-CT IRL4 and 6 and SD-CT IRL3. The present results show that higher levels of IR reliably decreased image noise in the RD-CT when compared to the SD-CT protocol, while the attenuation values remain
Table 3 Confidence for diagnosing and excluding spinal disorders using standard (SD-) and reduced-dose computed tomography (RDeCT).
Confidence Probably present, n (%) Definitely present, n (%)
SD-CT
RD-CT
IRL3 0 (0%) 80 (100%)
IRL4 3 (3.75%) 77 (96.25%)
IRL6 1 (1.25%) 79 (98.75%)
unchanged. This is in line with previous studies and the concept of iterative reconstruction.16 Even though there are a few studies regarding dose reduction in MDCT for imaging the lumbar spine,4,13,14,22 there are currently no studies evaluating its implementation in emergency patients with suspected fracture. Previous studies addressed the issue of reducing whole-spine MDCT for patients suffering from multiple myeloma and concluded that it is superior in detecting osseous lesions compared to conventional radiography.23,24 In comparison to the SD setting, the present reduced dose protocol was based on a higher tube voltage setting and a lower reference tube current, which is in accordance with Gleeson et al. proclaiming that these are the optimal parameters regarding the “as low as reasonably achievable” (ALARA) principles.25,26 The authors state because of the inherent high contrast between bone and soft tissues, a higher tube voltage would prove less susceptible to variations in body mass.26 A recent study evaluated reduced dose CT in comparison to conventional radiographs in patients with lower back pain.4 It was shown that RD-CT provided superior quality and anatomical information except for sharp reproduction of cortical and trabecular bone. As these authors did not examine patients with vertebral fractures or osseous lesions, it remains unclear if the mentioned pathologies could have been missed. In the present study, dose reduction did not influence diagnostic confidence and further improved visualisation of anatomical structures compared to SD-CT with lower level of IR. Studies focusing on evaluating the cervical spine using RD-CT with IR demonstrated that a significant reduction of radiation exposure did not change subjective image quality.27,28 These findings are in accordance with the present results emphasising the use of RD-CT in patients suspected for lumbar spine fractures. As MDCT provides more information than plain radiographs for spinal trauma7 and disorders29 it has a high clinical value. MDCT is often inevitable to provide an accurate diagnosis in suspected spine fracture. Unfortunately, it leads to a higher radiation exposure than plain radiography.6,30 Tube current and voltage modulations along with
675.e14
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
Figure 4 Sagittal and axial CT images of acute spine fractures in two BMI matched (26 kg2/m) patients (56 and 57 years) who underwent RD-CT reconstructed with IRL4 (a) and IRL6 (b) and SD-CT with IRL3 (c). The fractures of L3 (aeb) and L4 (c) were detected with high confidence, while image quality of IRL4 and IRL6 (b and c) were rated best.
IR could significantly reduce radiation in emergency patients with suspected spine fractures. As the high diagnostic value of spinal MDCT has been outlined in numerous studies, the focus should lie on optimisation of RD-CT protocols in order to comply to the ALARA principles. RD-CT in combination with IR allows accurate diagnosis in suspected lumbar spine trauma, but further studies on a larger collective are warranted to confirm this hypothesis. Due to the high confidence in diagnostic accuracy in patients with suspected lumbar spine fracture using RD-CT combined with IR, further dose reduction could be achieved and should be investigated in future studies. The present study has several limitations. First, due to ethical reasons and its retrospective design, it was not possible to maintain intra-individual comparisons of RD-CT and SD-CT. Second, discharge reports as the standard of reference might be biased as MDCT usually serves as a reference in diagnosing vertebral fractures and not every patient underwent surgery or additional magnetic resonance imaging. Lastly, the settings for the RD-CT protocol were based on not yet published cadaver studies with the
aim to reduce the radiation dose. Pitch, tube voltage, and tube current as well as level of iteration were chosen differently compared to SD-CT. These changes and the lack of intra-individual examination compromise the comparison of RD-CT and SD-CT. In conclusion, the present study demonstrates that RDCT with a tube voltage of 140 kV at 70 quality reference mAs combined with higher levels of IR allows for a significant reduction of radiation dose while preserving high image quality and consequently high confidence in the detection of lumbar spine fractures.
References 1. Hu R, Mustard CA, Burns C. Epidemiology of incident spinal fracture in a complete population. Spine 1996;21(4):492. 2. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet 2002;359(9319):1761e7. 3. Venkatesan M, Fong A, Sell PJ. CT scanning reduces the risk of missing a fracture of the thoracolumbar spine. J Bone Jt Surg Br 2012;94B(8):1097e100. 4. Alshamari M, Geijer M, Norrman E, et al. Low dose CT of the lumbar spine compared with radiography: a study on image quality with implications for clinical practice. Acta Radiol 2016;57(5):602e11.
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15 5. Hart D, Hillier M, Shrimpton P. Doses to patients from radiographic and fluoroscopic X-ray imaging procedures in the UK. The Health Protection Agency (HPA), centre for radiation, chemical and environmental hazards (CRCE). 2012. 6. Biswas D, Bible JE, Bohan M, et al. Radiation exposure from musculoskeletal computerized tomographic scans. J Bone Jt Surg Am 2009;91(8):1882e9. 7. Antevil JL, Sise MJ, Sack DI, et al. Spiral computed tomography for the initial evaluation of spine trauma: a new standard of care? J Trauma Acute Care Surg 2006;61(2):382e7. 8. Wintermark M, Mouhsine E, Theumann N, et al. Thoracolumbar spine fractures in patients who have sustained severe trauma: depiction with multiedetector row CT. Radiology 2003;227(3):681e9. 9. Brenner DJ, Hall EJ. Computed tomographydan increasing source of radiation exposure. N Engl J Med 2007;357(22):2277e84. 10. Prasad KN, Cole WC, Haase GM. Radiation protection in humans: extending the concept of as low as reasonably achievable (ALARA) from dose to biological damage. Br J Radiol 2004;77(914):97e9. 11. Singh S, Kalra MK, Hsieh J, et al. Abdominal CT: comparison of adaptive statistical iterative and filtered back projection reconstruction techniques. Radiology 2010;257(2):373e83. 12. O’Neill SB, Laughlin PDM, Crush L, et al. A prospective feasibility study of sub-millisievert abdominopelvic CT using iterative reconstruction in Crohn’s disease. Eur Radiol 2013;23(9):2503e12. 13. Yang C-H, Wu T-H, Lin C-J, et al. Knowledge-based iterative model reconstruction technique in computed tomography of lumbar spine lowers radiation dose and improves tissue differentiation for patients with lower back pain. Eur J Radiol 2016. 14. Bohy P, de Maertelaer V, Roquigny A, et al. Multidetector CT in patients suspected of having lumbar disk herniation: comparison of standard-dose and simulated low-dose techniques. Radiology 2007;244(2):524e31. 15. Gervaise A, Osemont B, Lecocq S, et al. CT image quality improvement using adaptive iterative dose reduction with wide-volume acquisition on 320-detector CT. Eur Radiol 2012;22(2):295e301. 16. Veldhoen S, Laqmani A, Derlin T, et al. 256-MDCT for evaluation of urolithiasis: iterative reconstruction allows for a significant reduction of the applied radiation dose while maintaining high subjective and objective image quality. J Med Imaging Radiat Oncol 2014;58(3):283e90. 17. Willemink MJ, de Jong PA, Leiner T, et al. Iterative reconstruction techniques for computed tomography. Part 1: technical principles. Eur Radiol 2013;23(6):1623e31.
675.e15
€l PB, Fingerle AA, Renger B, et al. Initial performance characterization 18. Noe of a clinical noise-suppressing reconstruction algorithm for MDCT. AJR Am J Roentgenol 2011;197(6):1404e9. 19. Deak PD, Smal Y, Kalender WA. Multisection CT protocols: sex- and agespecific conversion factors used to determine effective dose from doseelength product. Radiology 2010;257(1):158e66. 20. Bongartz G, Golding SJ, Jurik AJ. European guidelines for multislice computed tomography: report EUR 16262 EN 2004. Luxembourg: European Commission; 2004. Retrieved October 15, 2017, from: http://www. drs.dk/guidelines/ct/quality/index.htm. 21. Shah LM, Ross JS. Imaging of spine trauma. Neurosurgery 2016;79(5): 626e42. 22. Abul-Kasim K. Low-dose spine CT: optimisation and clinical implementation. Radiat Prot Dosim 2010;139(1e3):169e72. 23. Horger M, Claussen CD, Bross-Bach U, et al. Whole-body low-dose multidetector row-CT in the diagnosis of multiple myeloma: an alternative to conventional radiography. Eur J Radiol 2005;54(2):289e97. 24. Chassang M, Grimaud A, Cucchi JM, et al. Can low-dose computed tomographic scan of the spine replace conventional radiography? An evaluation based on imaging myelomas, bone metastases, and fractures from osteoporosis. Clin Imaging 2007;31(4):225e7. 25. Gleeson TG, Moriarty J, Shortt CP, et al. Accuracy of whole-body lowdose multidetector CT (WBLDCT) versus skeletal survey in the detection of myelomatous lesions, and correlation of disease distribution with whole-body MRI (WBMRI). Skeletal Radiol 2009;38(3):225e36. 26. Gleeson TG, Byrne B, Kenny P, et al. Image quality in low-dose multidetector computed tomography: a pilot study to assess feasibility and dose optimization in whole-body bone imaging. Can Assoc Radiol J 2010;61(5):258e64. € rner M, Hempel R, et al. Evaluation of a dedicated MDCT 27. Geyer LL, Ko protocol using iterative image reconstruction after cervical spine trauma. Clin Radiol 2013;68(7):e391e6. 28. Patro SN, Chakraborty S, Sheikh A. The use of adaptive statistical iterative reconstruction (ASiR) technique in evaluation of patients with cervical spine trauma: impact on radiation dose reduction and image quality. Br J Radiol 2016;89(1060):20150082. 29. Sakai T, Sairyo K, Takao S, et al. Incidence of lumbar spondylolysis in the general population in Japan based on multidetector computed tomography scans from two thousand subjects. Spine 2009;34(21):2346e50. 30. Richards PJ, George J, Metelko M, et al. Spine computed tomography doses and cancer induction. Spine 2010;35(4):430e3.
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
MDCT in suspected lumbar spine fracture: comparison of standard and reduced dose settings using iterative reconstruction J.M. Weinrich a, *, L. Well a, M. Regier a, C. Behzadi a, S. Sehner b, G. Adam a, A. Laqmani a a
Department for Diagnostic and Interventional Radiology and Nuclear Medicine, University Medical Center Hamburg-Eppendorf, Hamburg, Germany b Department of Medical Biometry and Epidemiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
art icl e i nformat ion Article history: Received 13 November 2017 Accepted 16 February 2018
AIM: To compare standard (SD-) and reduced-dose computed tomography (RD-CT) in combination with iterative reconstruction (IR) in emergency patients with suspected lumbar spine fracture. MATERIAL AND METHODS: Forty emergency patients with suspected lumbar spinal disorder who underwent RD-CT and 40 body mass index-matched patients undergoing SD-CT were enrolled in this retrospective study. Raw data for RD-CT were reconstructed using two increasing IR levels (IRL) 4 and 6, while SD-CT was reconstructed with IRL3. Two radiologists assessed image quality, image noise, and reader confidence in interpreting findings of spinal fractures in a blinded manner. RESULTS: Effective radiation dose was reduced by 50% using RD-CT. Overall subjective image quality (SIQ) was high for both protocols and slightly superior in the RD-CT protocol for both IRL compared to SD-CT. The detection rate of spinal disorders was high for both protocols with a high interobserver agreement. CONCLUSION: RD-CT with higher levels of IR results in substantial dose reduction of 50% in lumbar spine CT while maintaining an excellent subjective image quality resulting in a high diagnostic confidence. Ó 2018 Published by Elsevier Ltd on behalf of The Royal College of Radiologists.
Introduction Traumatic spinal injuries are the leading cause of death and disability in young adults,1 whereas elderly patients
* Guarantor and correspondent: J. Weinrich, Department of Radiology, University of Hamburg, Martinistraße 52, Hamburg, 20251, Germany. Tel.: þ49 (0) 152 2281 7461. E-mail address: [email protected] (J.M. Weinrich).
mainly suffer from osteoporotic fractures.2 Lumbar spine radiography remains the first radiographic examination for suspected fracture, even though there is evidence of its limited value.3 Additionally, radiographs of the lumbar spine result in considerably high radiation doses of approximately 1 mSv.4,5 Even though effective dose of lumbar spine computed tomography (CT) can be as high as 19 mSv,6 it is regarded as the imaging technique of choice for the evaluation of vertebral fractures in emergency patients.7,8
https://doi.org/10.1016/j.crad.2018.02.015 0009-9260/Ó 2018 Published by Elsevier Ltd on behalf of The Royal College of Radiologists.
675.e10
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
Due to the constant advances in CT technology, the amount of examinations has constantly risen over the last decade. Even though reduced-dose CT protocols are increasingly established in clinical practice for certain diagnostic questions as for the detection of urolithiasis or for the characterising patterns of pulmonary disease, the cumulative radiation exposure from medical examination has increased.9 In order to achieve low radiation doses, optimisation of CT protocols without sacrificing image quality is needed.10 Dose reduction is associated with increased image noise, which can impair image quality.11 Iterative reconstruction (IR) algorithms in CT enable the reduction of patient exposure by decreasing image noise.11,12 There is only scarce literature regarding dose reduction in lumbar spine CT,4,13e15 especially in the evaluation of lumbar spine fractures. To the authors’ knowledge, there have been no studies on dose reduction in evaluating emergency patients with suspected lumbar spine trauma. Therefore, the aim of the present study was to compare the effect of reduced-dose (RD) CT with IR to standard-dose (SD) CT with a lower level of IR in emergency patients with suspected lumbar spine fracture.
Materials and methods
Between March and July 2016, 40 patients who were referred to the emergency department for a clinically and radiographically indicated RD-CT of the lumbar spine for suspected fracture were included consecutively. As a reference group for radiation dose comparison, datasets of 40 body mass index (BMI; range: 18.8e40.4 kg/m2), gender, and age-matched patients submitted for suspected lumbar spine fracture who had undergone a SD-CT for suspected trauma were also included. Demographic patient data including sex, age, height, and weight, and imaging data were noted (Table 1). BMI was calculated using the formula (patient weight [kg]/patient height [m]2). Formal ethics Table 1 Demographic and CT specific baseline data.
Sex (percent female) Numbers Age (years) BMI (kg2/m) Tube currentetime product (mAs) CTDIvol (mGy) DLP (mGy$cm) Scan length (cm) Effective dose (mSv)
Imaging protocols All examinations were performed on a 256-section CT system (Brilliance iCT, Philips, Best, the Netherlands). Patients were placed in supine position with the head first on the CT table and arms above the head for both protocols. Details are given in Table 2. An established IR algorithm (iDose,4 Philips Healthcare) was used for IR.16 Within iDose,4 increasing levels represent increasing strength of noise reduction. The technical details of IR have been described elsewhere.17 Based on preliminary unpublished results of cadaver studies, which were performed at University Medical Center Hamburg-Eppendorf and the assumption of higher image noise reduction through the application of higher levels of IR,18 higher levels of IR were applied in RD-CT compared to the SD-CT protocol. Images were reconstructed using the iDose4 levels 4 (IRL4) and 6 (IRL6) for RD-CT and level 3 for SD-CT (IRL3). Images were reconstructed with a 3 mm section-thickness for qualitative image quality evaluation. All datasets were reconstructed in a bone and soft-tissue window setting in the axial, coronal, and sagittal planes.
Radiation dose
Study population
SD-CT
approval was not required for retrospective analysis of anonymised imaging data.
RD-CT
p-Value
Mean (SD)
Mean (SD)
50% 40 66.316.1 25.95.1 170.270
50% 40 66.316.1 26.14.6 67.132.2
0.826 <0.001
11.44.6 403.7188.3 25.28.3 6.22.9
6.93.3 209.2115.2 23.56.5 3.21.8
<0.001 <0.001 0.32 <0.001
The CT dose index (CTDIvol) and doseelength product (DLP) were derived from the automatically generated dose protocol of each examination. The effective dose (ED) was estimated by multiplying the DLP by a conversion factor of 0.0153 mSv/mGy∙cm for 120 kV and 0.0155 mSv/mGy∙cm for 140 kV.19 Estimated dose saving was expressed as the estimated percentage of radiation reduced between the two protocols.
Quantitative image analysis Quantitative image analysis was performed using a dedicated PACS workstation (PACS IW, GE Healthcare, Milwaukee, MI, USA) by one experienced radiologist (J.W.). Attenuation was measured in the axial plane by placing circular regions of interest (ROIs) of 100 mm2 at the centre Table 2 Imaging protocols. SD-CT
Tube currentetime product (mAs) values are given as the mean modulated values. SD-CT, standard-dose computed tomography; RD-CT, reduced-dose computed tomography (RD-CT).
Collimation (mm) Rotation time (s) Pitch Tube voltage (kV) Reference tube current (mAs) Automatic tube current modulation Level of iteration (iDose level) Kernel Section thickness (mm) Reformations Field of view SD-CT, standard-dose computed computed tomography (RD-CT).
RD-CT
21280.0625 0.5 0.993 0.579 120 140 158 70 Z-DOM 3 4 and 6 B and D 3 Axial, coronal, sagittal 250250 mm tomography;
RD-CT,
reduced-dose
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
of the abdominal aorta in each patient. To prevent bias resulting from a single ROI measurement, each region was measured at three adjacent 1-mm sections. Results were then averaged for further analyses. The standard deviation of the attenuation values served as objective image noise (OIN).
Qualitative image analysis For qualitative image analysis, data sets of 80 RD-CT (40 RD-CT examinations reconstructed with the two increasing iDose4 levels IRL4 and IRL6) and 40 SD-CT examinations (reconstructed with IRL3) were anonymised and reviewed in a blinded manner. All 120 reconstructions were randomised and three separate review sessions after a period of 14 days were performed to ensure a maximum level of blinding. Despite the attempts to blind the evaluation process regarding different reconstruction algorithms and levels, unambiguous blinding might not have been achieved, due to slightly altered image texture in different reconstructions. The images were reviewed independently by two radiologists with 4 and 5 years of experience in musculoskeletal imaging. Bone (window width, 2500 HU; window centre, 500 HU) and soft-tissue (window width, 360 HU; window centre, 60 HU) windows were used. For overall image quality, an established five-point scale was used based on subjective image quality, image noise and presence of artefacts16: 1, poor image quality, major artefacts, major image noise; 2, reduced image quality, substantial artefacts, extensive image noise; 3, acceptable image quality, moderate artefacts, moderate image noise; 4, good image quality, minor artefacts, minor image noise; and 5, excellent image quality, no artefacts, no perceived image noise. In order to evaluate visibility of anatomical structures a modified grading method based on the European Guidelines on Quality Criteria Computed Tomography was used.20 Three anatomical regions were included for analysis: (1) cortical and trabecular bone, intervertebral foramina, (2) pedicle and intervertebral joints, and (3) spinous and transverse processes. The grading system was based on a three-point scale as follows: 1, structures are well-defined with sharp contours, free of artefacts, and adequate for diagnosis; 2, structures are visible, but contours are not completely demarcated; mild artefacts can be seen, but they do not prevent diagnosis; and 3, structures are unclear with blurred anatomical boundaries, have considerable artefacts, and prevent proper diagnosis.
Diagnostic reliability Spinal disorders were independently evaluated by both readers in concordance with international standards reported elsewhere.21 Both readers evaluated the following pathologies: fracture yes/no, acute versus old fracture, spondylolisthesis, osteolytic or osteoblastic metastases. Reviewer’s confidence in detection of findings was rated on a five-point Likert scale: 1, definitely absent; 2, probably
675.e11
absent; 3, equivocal; 4, probably present; 5, definitely present.
Diagnostic validity One radiologist retrospectively retrieved the pertinent medical records of all patients and compared final clinical diagnoses with detected CT findings. The final clinical diagnosis was determined by review of the patients’ medical records in the hospital information system. The diagnosis was confirmed by review of the patient’s hospital discharge records, reports of surgical procedures, hospital discharge records, or clinical follow-up information.
Statistical analysis Sample characteristics are given as absolute and relative frequencies or mean standard deviation, whichever is appropriate. Baseline characteristics of the both groups (RD-CT versus SD-CT) were compared using the Welch’s approximation for t-test. To account for the repeated measurement structure of the data, defined by three 1 mm adjacent image sections within the aorta per patient and reconstruction level, a linear mixed effect model repeat measurement (MMRM) was calculated for the quantitative parameters. In this model, as a predictor for the quantitative parameters the reconstruction level was included. For the qualitative Likert scale an analogous model was used while the repeated measurements were defined by the ratings of the two readers. To evaluate the subjective image quality a linear mixed effect model repeat measurement was used with the interaction between reconstruction level and each subjective image quality characteristic as a predictor. To evaluate the visibility (well defined versus visible) a logistic mixed effect model repeat measurement was used with the interaction between IRL and the three different anatomical structures as a predictor. If interaction terms were insignificant, only the main effects were included. Moreover, all models were controlled for the potential confounder of BMI, age, gender, tube current, and effective dose CTDIvol. The results were reported as estimated means, which are represented in graphs with their corresponding 95% confidence intervals (95% CIs). Post hoc tests for comparison of the estimated means were calculated with contrast tests, using Wald tests. To test for interrater reliability regarding presence of a pathology Cohen’s kappa test was performed. Two sided p-values of <0.05 were considered significant. All analyses were computed using Stata 14.1 (STATA Corporation, College Station, TX, USA).
Results Radiation dose The mean exposure, CTDIvol, DLP, and effective dose in RD-CT were significantly reduced compared to SD-CT (Table 1). CTDIvol was reduced by 39.5% and effective dose
675.e12
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
was reduced by approximately 50% in the RD-CT protocol: RD-CT: 3.21.8 mGy versus SD-CT: 6.22.9 mGy.
Quantitative results Quantitative results (95% confidence interval [CI]) are graphically displayed in Fig 1. None of the confounder showed a significant effect on the quantitative parameters. The mean CT-N in SD-CT studies reconstructed with IRL3 were slightly but not significantly higher than in RD-CT reconstructions with IRL4 and IRL6 (SD-CT: IRL3: 55.9 HU; RD-CT: IRL4: 51.4 HU and IRL6: 51.7 HU). Compared to SDCT reconstructed with IRL3, objective image noise was reduced up to 20% and 37% in RD-CT reconstructed with IRL4 and IRL6 (Fig 1).
Qualitative results Mean scores (95% CIs) of qualitative image analysis are graphically displayed in Fig 2. None of the confounder showed a significant interaction with level of IR or image quality. Overall subjective image quality was high and was only slightly improved in the RD-CT IRL4 (p¼0.049) when compared to SD-CT IRL3 (Fig 2). Subjective image quality further improved with the use of RD-CT IRL6 compared to SD-CT IRL3 (p¼0.006).
Visibility of anatomical structures Overall visibility of anatomical structures was high in both protocols and a score of 3 was not given. Therefore, only scores of 1 and 2 were considered in the statistical analyses. Visibility of ‘‘cortical and the trabecular bone’’ was rated better for RD-CT IRL4 when compared to SD-CT IRL3 (p¼0.0016), while there was no difference compared to RDCT IRL6 (p¼0.238). There was no significant difference between SD-CT IRL3 and RD-CT IRL4 and IRL6 for the remaining two criteria “intervertebral foramina, pedicles and intervertebral joints” and “spinous and transverse processes” (all p>0.05; Fig 3).
Figure 2 Qualitative analysis of SD-CT IRL3 and RD-CT reconstructed with IRL4 and IRL6. Plots show mean scores for subjective image quality. Error bars represent the 95% CI. Subjective five-point grading scale (1 indicating worst through 5 indicating best).
Diagnostic confidence Overall confidence was high and there were no equivocal findings. For SD-CT, both readers could definitely assure their diagnosis in every patient (80/80). RD-CT IRL4 and IRL6 achieved 77/80 and 79/80 definite diagnoses, respectively (Table 3).
Reliability Both readers diagnosed the same amount of pathologies. Interobserver agreement for detecting pathologies was high for both protocols (SD-CT: 97.5%, kappa: 0.950.09; RD-CT IRL4 and IRL6: 100%, kappa: 10.09). There was only one disagreement (probably versus definitely present) in the SD-CT protocol regarding an old vertebrae fracture.
Figure 1 Quantitative analysis of SD-CT IRL3 and RD-CT reconstructed with IRL4 and IRL6. Error bars represent the 95% CI and points between intervals are given as mean.
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
675.e13
Figure 3 Visibility analysis of anatomical structures of SD-CT reconstructed with IRL3 and RD-CT reconstructed with IRL4 and IRL 6. The plots show the mean probability of visibility (error bars represent the 95% CI).
CT findings and validity Twenty-four out of 40 patients in the SD-CT group and 26/40 patients in the LDCT group had an acute vertebral fracture (Fig 4), whereas 21/80 did not have any spinal disorders. Four of 80 patients revealed old fractures and 3/ 80 had osteolytic or osteoblastic lesions. There were only 2/ 80 cases of spondylolisthesis. Seventeen of 50 patients with acute fractures also had at least one old fracture.
Discussion In the present study, the effective radiation dose to lumbar spine was reduced by approximately 50% in patients with suspected fracture who underwent MDCT, while at the same time maintaining high diagnostic confidence and subjective image quality. The present results show that a RD-CT protocol combined with higher IR levels does not only provide an image quality that is at least on par with SDCT with lower levels of IR, but also enables an equivalent and even better visualisation of anatomical structures. Especially, the visibility of cortical and trabecular bone was rated better for RD-CT IRL4 when compared to SD-CT IRL3. Higher levels of IR improved subjective image quality of almost all diagnostic criteria analysed within the RD-CT protocol when compared to the SD-CT. In accordance with discharge records and further treatment both readers detected all pathologies and showed a very high confidence in evaluating both, RD-CT IRL4 and 6 and SD-CT IRL3. The present results show that higher levels of IR reliably decreased image noise in the RD-CT when compared to the SD-CT protocol, while the attenuation values remain
Table 3 Confidence for diagnosing and excluding spinal disorders using standard (SD-) and reduced-dose computed tomography (RDeCT).
Confidence Probably present, n (%) Definitely present, n (%)
SD-CT
RD-CT
IRL3 0 (0%) 80 (100%)
IRL4 3 (3.75%) 77 (96.25%)
IRL6 1 (1.25%) 79 (98.75%)
unchanged. This is in line with previous studies and the concept of iterative reconstruction.16 Even though there are a few studies regarding dose reduction in MDCT for imaging the lumbar spine,4,13,14,22 there are currently no studies evaluating its implementation in emergency patients with suspected fracture. Previous studies addressed the issue of reducing whole-spine MDCT for patients suffering from multiple myeloma and concluded that it is superior in detecting osseous lesions compared to conventional radiography.23,24 In comparison to the SD setting, the present reduced dose protocol was based on a higher tube voltage setting and a lower reference tube current, which is in accordance with Gleeson et al. proclaiming that these are the optimal parameters regarding the “as low as reasonably achievable” (ALARA) principles.25,26 The authors state because of the inherent high contrast between bone and soft tissues, a higher tube voltage would prove less susceptible to variations in body mass.26 A recent study evaluated reduced dose CT in comparison to conventional radiographs in patients with lower back pain.4 It was shown that RD-CT provided superior quality and anatomical information except for sharp reproduction of cortical and trabecular bone. As these authors did not examine patients with vertebral fractures or osseous lesions, it remains unclear if the mentioned pathologies could have been missed. In the present study, dose reduction did not influence diagnostic confidence and further improved visualisation of anatomical structures compared to SD-CT with lower level of IR. Studies focusing on evaluating the cervical spine using RD-CT with IR demonstrated that a significant reduction of radiation exposure did not change subjective image quality.27,28 These findings are in accordance with the present results emphasising the use of RD-CT in patients suspected for lumbar spine fractures. As MDCT provides more information than plain radiographs for spinal trauma7 and disorders29 it has a high clinical value. MDCT is often inevitable to provide an accurate diagnosis in suspected spine fracture. Unfortunately, it leads to a higher radiation exposure than plain radiography.6,30 Tube current and voltage modulations along with
675.e14
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15
Figure 4 Sagittal and axial CT images of acute spine fractures in two BMI matched (26 kg2/m) patients (56 and 57 years) who underwent RD-CT reconstructed with IRL4 (a) and IRL6 (b) and SD-CT with IRL3 (c). The fractures of L3 (aeb) and L4 (c) were detected with high confidence, while image quality of IRL4 and IRL6 (b and c) were rated best.
IR could significantly reduce radiation in emergency patients with suspected spine fractures. As the high diagnostic value of spinal MDCT has been outlined in numerous studies, the focus should lie on optimisation of RD-CT protocols in order to comply to the ALARA principles. RD-CT in combination with IR allows accurate diagnosis in suspected lumbar spine trauma, but further studies on a larger collective are warranted to confirm this hypothesis. Due to the high confidence in diagnostic accuracy in patients with suspected lumbar spine fracture using RD-CT combined with IR, further dose reduction could be achieved and should be investigated in future studies. The present study has several limitations. First, due to ethical reasons and its retrospective design, it was not possible to maintain intra-individual comparisons of RD-CT and SD-CT. Second, discharge reports as the standard of reference might be biased as MDCT usually serves as a reference in diagnosing vertebral fractures and not every patient underwent surgery or additional magnetic resonance imaging. Lastly, the settings for the RD-CT protocol were based on not yet published cadaver studies with the
aim to reduce the radiation dose. Pitch, tube voltage, and tube current as well as level of iteration were chosen differently compared to SD-CT. These changes and the lack of intra-individual examination compromise the comparison of RD-CT and SD-CT. In conclusion, the present study demonstrates that RDCT with a tube voltage of 140 kV at 70 quality reference mAs combined with higher levels of IR allows for a significant reduction of radiation dose while preserving high image quality and consequently high confidence in the detection of lumbar spine fractures.
References 1. Hu R, Mustard CA, Burns C. Epidemiology of incident spinal fracture in a complete population. Spine 1996;21(4):492. 2. Cummings SR, Melton LJ. Epidemiology and outcomes of osteoporotic fractures. Lancet 2002;359(9319):1761e7. 3. Venkatesan M, Fong A, Sell PJ. CT scanning reduces the risk of missing a fracture of the thoracolumbar spine. J Bone Jt Surg Br 2012;94B(8):1097e100. 4. Alshamari M, Geijer M, Norrman E, et al. Low dose CT of the lumbar spine compared with radiography: a study on image quality with implications for clinical practice. Acta Radiol 2016;57(5):602e11.
J.M. Weinrich et al. / Clinical Radiology 73 (2018) 675.e9e675.e15 5. Hart D, Hillier M, Shrimpton P. Doses to patients from radiographic and fluoroscopic X-ray imaging procedures in the UK. The Health Protection Agency (HPA), centre for radiation, chemical and environmental hazards (CRCE). 2012. 6. Biswas D, Bible JE, Bohan M, et al. Radiation exposure from musculoskeletal computerized tomographic scans. J Bone Jt Surg Am 2009;91(8):1882e9. 7. Antevil JL, Sise MJ, Sack DI, et al. Spiral computed tomography for the initial evaluation of spine trauma: a new standard of care? J Trauma Acute Care Surg 2006;61(2):382e7. 8. Wintermark M, Mouhsine E, Theumann N, et al. Thoracolumbar spine fractures in patients who have sustained severe trauma: depiction with multiedetector row CT. Radiology 2003;227(3):681e9. 9. Brenner DJ, Hall EJ. Computed tomographydan increasing source of radiation exposure. N Engl J Med 2007;357(22):2277e84. 10. Prasad KN, Cole WC, Haase GM. Radiation protection in humans: extending the concept of as low as reasonably achievable (ALARA) from dose to biological damage. Br J Radiol 2004;77(914):97e9. 11. Singh S, Kalra MK, Hsieh J, et al. Abdominal CT: comparison of adaptive statistical iterative and filtered back projection reconstruction techniques. Radiology 2010;257(2):373e83. 12. O’Neill SB, Laughlin PDM, Crush L, et al. A prospective feasibility study of sub-millisievert abdominopelvic CT using iterative reconstruction in Crohn’s disease. Eur Radiol 2013;23(9):2503e12. 13. Yang C-H, Wu T-H, Lin C-J, et al. Knowledge-based iterative model reconstruction technique in computed tomography of lumbar spine lowers radiation dose and improves tissue differentiation for patients with lower back pain. Eur J Radiol 2016. 14. Bohy P, de Maertelaer V, Roquigny A, et al. Multidetector CT in patients suspected of having lumbar disk herniation: comparison of standard-dose and simulated low-dose techniques. Radiology 2007;244(2):524e31. 15. Gervaise A, Osemont B, Lecocq S, et al. CT image quality improvement using adaptive iterative dose reduction with wide-volume acquisition on 320-detector CT. Eur Radiol 2012;22(2):295e301. 16. Veldhoen S, Laqmani A, Derlin T, et al. 256-MDCT for evaluation of urolithiasis: iterative reconstruction allows for a significant reduction of the applied radiation dose while maintaining high subjective and objective image quality. J Med Imaging Radiat Oncol 2014;58(3):283e90. 17. Willemink MJ, de Jong PA, Leiner T, et al. Iterative reconstruction techniques for computed tomography. Part 1: technical principles. Eur Radiol 2013;23(6):1623e31.
675.e15
€l PB, Fingerle AA, Renger B, et al. Initial performance characterization 18. Noe of a clinical noise-suppressing reconstruction algorithm for MDCT. AJR Am J Roentgenol 2011;197(6):1404e9. 19. Deak PD, Smal Y, Kalender WA. Multisection CT protocols: sex- and agespecific conversion factors used to determine effective dose from doseelength product. Radiology 2010;257(1):158e66. 20. Bongartz G, Golding SJ, Jurik AJ. European guidelines for multislice computed tomography: report EUR 16262 EN 2004. Luxembourg: European Commission; 2004. Retrieved October 15, 2017, from: http://www. drs.dk/guidelines/ct/quality/index.htm. 21. Shah LM, Ross JS. Imaging of spine trauma. Neurosurgery 2016;79(5): 626e42. 22. Abul-Kasim K. Low-dose spine CT: optimisation and clinical implementation. Radiat Prot Dosim 2010;139(1e3):169e72. 23. Horger M, Claussen CD, Bross-Bach U, et al. Whole-body low-dose multidetector row-CT in the diagnosis of multiple myeloma: an alternative to conventional radiography. Eur J Radiol 2005;54(2):289e97. 24. Chassang M, Grimaud A, Cucchi JM, et al. Can low-dose computed tomographic scan of the spine replace conventional radiography? An evaluation based on imaging myelomas, bone metastases, and fractures from osteoporosis. Clin Imaging 2007;31(4):225e7. 25. Gleeson TG, Moriarty J, Shortt CP, et al. Accuracy of whole-body lowdose multidetector CT (WBLDCT) versus skeletal survey in the detection of myelomatous lesions, and correlation of disease distribution with whole-body MRI (WBMRI). Skeletal Radiol 2009;38(3):225e36. 26. Gleeson TG, Byrne B, Kenny P, et al. Image quality in low-dose multidetector computed tomography: a pilot study to assess feasibility and dose optimization in whole-body bone imaging. Can Assoc Radiol J 2010;61(5):258e64. € rner M, Hempel R, et al. Evaluation of a dedicated MDCT 27. Geyer LL, Ko protocol using iterative image reconstruction after cervical spine trauma. Clin Radiol 2013;68(7):e391e6. 28. Patro SN, Chakraborty S, Sheikh A. The use of adaptive statistical iterative reconstruction (ASiR) technique in evaluation of patients with cervical spine trauma: impact on radiation dose reduction and image quality. Br J Radiol 2016;89(1060):20150082. 29. Sakai T, Sairyo K, Takao S, et al. Incidence of lumbar spondylolysis in the general population in Japan based on multidetector computed tomography scans from two thousand subjects. Spine 2009;34(21):2346e50. 30. Richards PJ, George J, Metelko M, et al. Spine computed tomography doses and cancer induction. Spine 2010;35(4):430e3.