Low back pain is an increasingly common complaint in pediatric patients.1 Among the common structural causes of back pain is spondylolysis,1–3 a condition that implies a continuum of disease, including pars stress reactions, true spondylolysis, and spondylolisthesis.2 Herman and Pizzutillo4 describe spondylolysis as a “defect or abnormality of the pars interarticularis and surrounding lamina and pedicle.” The authors identify three distinct types of spondylolysis. Stress reactions are described as “intraosseous edema with surrounding sclerosis of the pars, lamina, or pedicle without cortical or trabecular disruption.”4 Stress fractures are described as a “disruption of trabecular or cortical bone of the pars without a bony gap or lysis.”4 Pars nonunions or spondylolytic defects are “complete disruptions of the pars interarticularis with a gap and surrounding sclerosis at the edges of the defect.”4 Spondylolisthesis is “the translation of one vertebral segment relative to the next caudal segment.”4 The radiographic diagnosis of spondylolysis has long been a topic of controversy.5 When the clinical diagnosis of spondylolysis is in question, the physical examination and radiographic confirmation is difficult and imaging strategies vary considerably among practitioners and published studies.6 Existing published imaging guidelines are vague and recommend initial investigation of pediatric patients with low back pain with radiographs first, then magnetic resonance, or nuclear medicine modalities.7 There is not a specific recommendation for spondylolysis.
The imaging modalities commonly used to evaluate spondylolysis include plain film radiographs, bone scintigraphy (bone scan), computed tomography (CT), single-photon emission CT (SPECT) or SPECT-CT, and magnetic resonance imaging (MRI).3,6 Each of these modalities has advantages and disadvantages when investigating low back pain presumed to be caused by spondylolysis. Plain films are arguably advantageous in that they are inexpensive and provide a low dose of radiation exposure in a susceptible population of pediatric patients.6 However, they do not provide the same degree of sensitivity as the advanced imaging modalities unless there is an obvious pars defect. Where the diagnosis is in question, MRI has been shown to be superior in detecting early pars stress reactions that cannot be seen on plain films or CT, and avoids radiation exposure entirely.3 Bone scan involves a larger radiation dose6 but has the advantage of being quite sensitive in the detection of lesions.8 CT is best at assessing the bony spinal elements to determine if a lytic defect has occurred.9
Although there is a considerable amount of literature that describes the use of these modalities for the radiographic diagnosis of spondylolysis, there are few consistent recommendations for how to effectively utilize imaging in making a diagnosis. There is little consensus regarding which modality to employ first, at what point in the evolution of the disease each is most effective, and if there is any particular sequence of imaging studies that should be used preferentially. This is in part due to the inherent difficulty of comparing multiple imaging modalities used to make a diagnosis without a reliable gold standard. To our knowledge, the use of imaging modalities in pediatric spondylolysis has not been systematically reviewed to date. The aim of this study is to provide an evidence-based recommendation for how and when to employ imaging studies when diagnosing pediatric back pain thought to be caused by spondylolysis.
MATERIALS AND METHODS
A literature search was performed in PubMed and Cochrane databases. The search terms comprised “spondylolysis,” “pediatric,” “adolescent,” “juvenile,” “young,” “lumbar,” “MRI,” “bone scan,” “CT,” and “SPECT.” A set of formalized inclusion criteria was developed. The screening criteria deemed studies appropriate for consideration that were English language, published in the past 15 years, had titles and content related to the diagnostic imaging of pediatric spondylolysis, included a comparison of imaging modalities or made a recommendation concerning a diagnostic imaging algorithm, and could be classified as level III evidence or better using the Journal of Bone and Joint Surgery criteria for levels of evidence. After an initial search using broad search terms, 4732 articles were identified. After the removal of duplicates and screening for relevant content, 17 articles remained. Based upon the exclusion criteria and a full-text review of these articles, level of evidence was assigned, and an additional four were excluded, leaving 13 articles for review.3,6,8–18 This process was replicated in a blinded fashion by two authors and an independent referee reconciled any disagreements in level of evidence assignment or exclusion criteria. All of these articles were included in the qualitative synthesis, while eight had adequate data for quantitative review.
Physical Examination Findings
Two papers included performance figures for physical examination findings (Table 1). Kobayashi et al.12 investigated lumbar spine extension, flexion, Kemp test, and percussion, but did not find that any specific maneuver correlated with imaging findings. Masci et al.17 investigated the one-legged hyperextension test and found that the test was roughly 50% sensitive when compared with several advanced imaging modalities. No study made a concrete recommendation in favor of the use of physical examination findings for definitive diagnosis.12,17
Ten papers included sensitivity calculations or kappa values for comparing imaging performance3,6,8,10,12–15,17,18 (Table 2). There were five data points comparing the sensitivity of MRI and CT with CT as the gold standard.3,13,14,17 The average sensitivity of MRI versus CT was 81.4%. One paper calculated a kappa value of 0.83 for MRI with CT as the gold standard.3 The sensitivity of CT was also calculated using SPECT as the gold standard. Using two data points, the average sensitivity of CT versus SPECT was 85%.8,18 One study calculated the sensitivity of MRI using SPECT as the gold standard 80%.17 The kappa value of this same comparison was 0.79.3 In addition, the comparison of MRI and SPECT revealed that MRI correlated more consistently with subsequent lysis, while positive lesions on SPECT with negative MRI did not consistently correlate with subsequent lysis.3 Two of the studies reviewed found MRI to be a superior to CT in performance for early lesions and CT superior to MRI for longitudinal follow-up.9,13 CT and MRI were both found to be comparable to SPECT in terms of sensitivity, making the ability of all three modalities to detect spondylolysis on imaging relatively similar.3,8,17,18
In comparison of plain film views, one study found that two-view plain films are equivalent to four view studies.6 Only one reported study made a useful comparison of the sensitivity of plain films, putting their performance at an inferior but still clinically useful threshold of 75% sensitivity when compared with CT and bone scan.6
Twelve studies made a recommendation as to how best to perform diagnostic imaging of patients with clinically suspected spondylolysis (Table 3).3,6,8–17 Three studies recommended plain films followed by MRI, with some recommending CT as a differentiator in case of equivocal findings.10,12,15 Two studies recommended MRI as the initial investigation of choice, while three others conditionally recommended MRI as an early investigative study.3,10,12–14 Two other studies recommended against MRI as an initial investigation, in both cases due to concerns regarding false negative imaging findings.14,17 Two studies advocated for SPECT followed by confirmatory CT if SPECT was positive.8,17 Another study advocated for plain films followed by SPECT.11
Of the studies that made recommendations regarding a preferred diagnostic algorithm, a majority conditionally preferred MRI as an initial or initial advanced imaging study, with some advocating follow-up CT in equivocal situations.3,9,10,12,13,15,16 Studies that advocated against MRI as an initial investigation found that the risk of false negatives was the primary concern.14,17 In addition, authors cited the risk of visualizing additional imaging findings such as disc pathology that might not be primarily responsible for symptoms. MRI performed best during the early course of spondylolysis and also at lower lumbar levels.13 It detected changes at additional levels not seen on CT and avoided the significant radiation exposure of SPECT or CT. CT, on the contrary, was especially useful for assessing healing in the longitudinal setting.12 SPECT was found to have a high rate of false positives, and more importantly, positive imaging findings on SPECT did not reliably correlate with locations of subsequent lysis.8,17 For this reason, studies advocated for a follow-up CT as a confirmatory study if SPECT imaging was positive.8,17
A number of studies advocated for initial workup with plain films, despite their reduced sensitivity, as they were inexpensive and require only a minimal dose of radiation.6,11,12,15 Finally, one study questioned the role of advanced imaging altogether, as conservative management with activity modification was often the common final endpoint in the treatment of pediatric low back pain including spondylolysis, regardless of imaging findings.6
One 2013 study made an attempt to quantify radiation exposure with the use of different imaging modalities.6 The authors calculated that four-view plain films and CT had approximately double the dose of two-view plain film studies, while bone scans had seven to nine times the effective dose of two-view plain films studies. In addition, the pelvic organs received an unusually significant proportion of the overall dose due to the proximity to the lumbar spine, the focus of the imaging investigation.
One study quantified the cost of several imaging modalities based upon charges at their institution.6 The reported costs of imaging studies were as follows: two-view plain films $350; four-view plain films $500; CT $2000; bone scan $1650; MRI $2900.
While low back pain in pediatric patients cannot be explained by diagnosable causes in up to 78% of initial presenting complaints,6 it is theorized that the early radiographic confirmation of acute lytic defects coupled with appropriate conservative management could lead to resolution of the pars lesion.12 Appropriate management strategies could thus in theory prevent the progression to spondylolisthesis and other clinical sequelae including ongoing symptoms of back pain.12 When there is clinical suspicion suggesting a diagnosis of spondylolysis in a pediatric patient, there is clinical utility in confirming the diagnosis using appropriate diagnostic imaging.3,9,16 Although this study is limited in that it does not address in detail when it is appropriate to investigate a complaint of pediatric back pain with diagnostic imaging, we have attempted to profile the advantages and disadvantages of the various imaging modalities and provide recommendations regarding their best use practices.
The uniform lack of high-quality studies among the included investigations makes it difficult to formulate concrete recommendations for clinical practice. In addition, there is a great deal of inconsistency in the various ways in which imaging study performance is reported and measured. The patient cohorts are heterogeneous among the included studies and study protocols are variable. It is clear that further rigorous comparison of the performance of the included imaging modalities would be beneficial in determining how best to effectively use them in the clinical setting.
Given these caveats, we conditionally propose a diagnostic sequence based upon the literature reviewed. In patients with low back pain suggestive of spondylolysis, in the absence of neurological examination abnormalities, due to their moderate effectiveness, low cost, and low radiation exposure, we find two-view plain films to be the best initial study. Physical examination findings diagnostic for spondylolysis remain unreliable among the patient cohorts reviewed. The clinical examination of patients with suspected spondylolysis is an unreliable indicator of bony spinal pathology and imaging therefore must be the primary means of diagnostic confirmation. If plain films do not yield a diagnosis or the patient does not improve clinically, advanced imaging can be undertaken if further investigation is clinically warranted. CT is comparable to MRI overall in terms of sensitivity, but practitioners should consider the time course of the symptoms and use MRI in early diagnosis and CT in more persistent courses. We discourage serial follow-up with CT, as the radiation burden sustained as a result of this approach would be substantial. We find the radiation burden of SPECT to be prohibitive in light of the availability of alternatives that perform similarly. However, if the clinical recommendation following negative plain films would be conservative management and activity modification regardless of the results of advanced imaging, we advise that providers utilize advanced imaging judiciously and only in circumstances in which it may change their clinical management, or in situations wherein symptoms are clinically severe, persistent, or neurological examination abnormalities are present.
The results reported in this systemic review propose a diagnostic imaging sequence for evaluating pediatric spondylolysis based upon the available literature to date. Although the studies reviewed are largely retrospective reviews and contain heterogeneous patient cohorts and variable imaging protocols, the recommendations suggested here provide a useful starting point for guiding a diagnostic workup. However, rigorous comparative studies would provide the ultimate determination of imaging performance.
1. Altaf F, Heran MKS, Wilson LF. Back pain in children and adolescents. Bone Joint J
2. Cavalier R, Herman MJ, Cheung EV, et al. Spondylolysis
and spondylolisthesis in children and adolescents: I. Diagnosis, natural history, and nonsurgical management. J Am Acad Orthop Surg
3. Campbell RSD, Grainger AJ, Hide IG, et al. Juvenile spondylolysis
: a comparative analysis of CT
. Skeletal Radiol
4. Herman MJ, Pizzutillo PD. Spondylolysis
and spondylolisthesis in the child and adolescent
: a new classification. Clin Orthop Relat Res
5. Gehweiler JA, Daffner RH. Low back pain: the controversy of radiologic evaluation. AJR Am J Roentgenol
6. Miller R, Beck NA, Sampson NR, et al. Imaging modalities for low back pain in children: a review of spondyloysis and undiagnosed back pain. J Pediatr Orthop
7. Armstrong P, Ringertz H, Bischof Delaloye A. Referral guidelines for imaging. European Commission Directorate-General for the Environment, Luxembourg. 2001:44–58.
8. Yang J, Servaes S, Edwards K, et al. Prevalence of stress reaction in the pars interarticularis in pediatric
patients with new-onset lower back pain. Clin Nucl Med
9. Dunn AJ, Campbell RSD, Mayor PE, et al. Radiological findings and healing patterns of incomplete stress fractures of the pars interarticularis. Skeletal Radiol
10. Rush JK, Astur N, Scott S, et al. Use of magnetic resonance imaging in the evaluation of spondylolysis
. J Pediatr Orthop
11. Bhatia NN, Chow G, Timon SJ, et al. Diagnostic modalities for the evaluation of pediatric
back pain: a prospective study. J Pediatr Orthop
12. Kobayashi A, Kobayashi T, Kato K, et al. Diagnosis of radiographically occult lumbar spondylolysis
athletes by magnetic resonance imaging. Am J Sports Med
13. Ganiyusufoglu AK, Onat L, Karatoprak O, et al. Diagnostic accuracy of magnetic resonance imaging versus computed tomography in stress fractures of the lumbar
spine. Clin Radiol
14. Yamaguchi KT, Skaggs DL, Acevedo DC, et al. Spondylolysis
is frequently missed by MRI
in adolescents with back pain. J Child Orthop
15. Goda Y, Sakai T, Sakamaki T, et al. Analysis of MRI
signal changes in the adjacent pedicle of adolescent
patients with fresh lumbar spondylolysis
. Eur spine J
16. Sairyo K, Katoh S, Takata Y, et al. MRI
signal changes of the pedicle as an indicator for early diagnosis of spondylolysis
in children and adolescents: a clinical and biomechanical study. Spine (Phila Pa 1976)
17. Masci L, Pike J, Malara F, et al. Use of the one-legged hyperextension test and magnetic resonance imaging in the diagnosis of active spondylolysis
. Br J Sports Med
2006; 40:940–946. discussion 946.
18. Gregory PL, Batt ME, Kerslake RW, et al. Single photon emission computerized tomography and reverse gantry computerized tomography findings in patients with back pain investigated for spondylolysis
. Clin J Sport Med