Medical Policy: 02.04.69
Original Effective Date: December 2017
Reviewed: December 2020
This policy contains information which is clinical in nature. The policy is not medical advice. The information in this policy is used by Wellmark to make determinations whether medical treatment is covered under the terms of a Wellmark member's health benefit plan. Physicians and other health care providers are responsible for medical advice and treatment. If you have specific health care needs, you should consult an appropriate health care professional. If you would like to request an accessible version of this document, please contact customer service at 800-524-9242.
Benefit determinations are based on the applicable contract language in effect at the time the services were rendered. Exclusions, limitations or exceptions may apply. Benefits may vary based on contract, and individual member benefits must be verified. Wellmark determines medical necessity only if the benefit exists and no contract exclusions are applicable. This medical policy may not apply to FEP. Benefits are determined by the Federal Employee Program.
This Medical Policy document describes the status of medical technology at the time the document was developed. Since that time, new technology may have emerged or new medical literature may have been published. This Medical Policy will be reviewed regularly and be updated as scientific and medical literature becomes available.
Adolescent idiopathic scoliosis (AIS) is a disease of unknown etiology that causes mild-to-severe spinal deformity in approximately 1% to 3% of adolescents. Males and females are affected equally, however, the risk of curve progression and therefore the need for treatment is ten times higher in females than in males. While there is controversy about the value of both screening and treatment, once diagnosed, patients are closely followed. In cases with significant progression of curvature, both medical (bracing) and surgical (spinal fusion) interventions are considered. It has been proposed that the ScoliScore AIS prognostic DNA-based test using an algorithm incorporating results of testing for 53 single-nucleotide variants (SNVs), along with the patient’s presenting spinal curve (Cobb angle), to generate a risk score (range 1-200), can be used qualitatively or quantitatively to predict the likelihood of spinal curve progression.
Diagnosis is established by radiologic observation in adolescents (age 10 years until the age of skeletal maturity) of a lateral spine curvature of 10 degrees or more, as measured using the Cobb angle. The Cobb angle is defined by the angulation measures between the maximally tilted proximal and distal vertebrae of the curve. Curvature is considered mild (<25 degrees), moderate (25-40 degrees), or severe (> 40 degrees) in a patient still growing. Once diagnosed, patients must be monitored over several years, usually with serial radiographs for curve progression. If the curve progresses, spinal bracing is the generally accepted first-line treatment. If the curve progresses in spite of bracing, spinal fusion may be recommended.
Adolescent idiopathic scolisis (AIS) is a complex disorder that appears to result from the interaction of multiple genetic loci and the environment, but the details of these interactions are not fully understood. A familial nature of this disease is noted in the literature, about one-quarter of patients report a positive family history, and studies in twins have consistently supported shared genetic factors. The ScoliScore AIS prognostic DNA-based test which uses an algorithm incorporating results of testing for 53 SNVs (single nucleotide variants), along with the patient’s presenting spinal curve (Cobb angle) to generate a risk score (range 1-200) to predict the likelihood of spinal curve progression. The genetic markers were identified from unpublished genome-wide association studies (GWAS). The test is intended for white (Caucasian) patients, ages 9 to 13 years with a primary diagnosis of AIS with mild scoliotic curve defined as < 25 degrees.
The development and validation of the ScoliScore SNV based prognostic algorithm were described in 2010 by Ward et. al. in an industry-sponsored study. The prognostic algorithm was developed in a cohort of 2192 female patients from prior studies. Candidate genes selected were selected based on previous GWAS data from same investigators. The independent effect of each SNV and of clinical factors (initial Cobb angle) and all gene-gene interaction terms were tested in a stepwise logistic regression using a backward-selection procedure, and then using a forward-selection procedure. The final predictive model including 53 SNV markers, multiple gene-gene interaction terms, and the patient’s initial Cobb angle. Prediction probabilities were converted to a numeric score ranging from 1 to 200. A priori, low risk progression was determined to be less than 1%; from the generation cohort, a score of less than 41 was selected as an initial cutoff.
The purpose of the ScoliScore AIS prognostic DNA based test and other individual single-nucleotide variant (SNV) based tests for scoliosis prognosis is primarily to determine whether patients with scoliosis are at a higher likelihood for curve progression. Such patients could undergo more frequent surveillance than they would without testing. The current standard for management of patients with scoliosis that is not severe enough to undergo bracing or surgery is observation with routine radiographic or clinical follow-up.
The relevant population of interest is individuals with a diagnosis of AIS that is not yet severe enough to require bracing or surgery.
The intervention of interest is testing for SNVs, including testing with the specific ScoliScore AIS prognostic tests, which uses multiple SNVs along with the Cobb angle in an algorithm.
Patients would be seen in the outpatient setting.
The following practices are currently being used to make decisions about follow-up for patients with AIS that is not severe enough to require bracing or surgery: routine radiographic or clinical follow-up, at an interval that is generally determined by the individual patient in shared decision making. The test is an adjunct to existing clinical information and tests results.
The general outcomes of interest are change in disease severity (i.e. progression in scoliosis curve), morbid events (i.e. development of sever scoliosis, which is generally considered to be a Cobb angle > 40 degrees), or back pain.
Beneficial outcomes resulting from a true test result, if a true test result is followed by earlier detection of scoliosis by either clinical or radiologic testing, would be earlier detection and treatment of scoliosis. Potential harms from the test include those from a false-positive or a false-negative: false-positive results could lead to increased clinical or radiologic surveillance, while false-negative test could lead to premature stopping of surveillance.
The relevant follow-up periods depends on the timing of presentation relative to the cessation of growth; however, it is generally over the course of 2 to 3 years.
A test must detect the presence or absence of a condition, the risk of developing a condition in the future, or treatment response (beneficial or adverse).
For the evaluation of clinical validity of the ScoliScore AIS test for scoliosis progression, studies that meet the following eligibility were considered:
In 2010, Ward et. al. described the validation of the ScoliScore algorithm in a group of patients who had a diagnosis of AIS but who had not been previously involved in any AIS/genotype-related studies. These subjects were preselected by curvature severity (mild, moderate, severe) and assigned into 3 cohorts identified as: (1) a screening cohort of white females; (2) a spinal surgery practice cohort of white females; and (3) a male cohort. Inclusion/exclusion criteria were cited as being used, but not explicitly provided, although a component of cohort development was matching of disease prevalence by severity according to that expected from review of the literature or survey of clinical practices. Ward provided minimal information about the demographics of patients assigned to each cohort. Assignment of curvature severity was performed using expert opinion of a single orthopedic spine surgeon and was supplemented by external blinded review of the spinal surgery practice patients using an outside panel of 3 independent scoliosis experts.
The screening cohort was composed of 277 patients recruited to ensure 85% exhibited mild or improved curves, 12% moderate curve progression, and 3% severe curve progression. Using a risk score cutoff of 41 or less, the predictive value of a negative test (defined as identification of patients without severe curve progression) was 100% (95% confidence interval [CI], 98.6% to 100%). No analysis was performed to demonstrate whether this was a statistically significant improvement in prediction of negatives, given the low initial prevalence of patients expected to exhibit severe progression.
The spine surgery practice cohort was composed of 257 patients recruited to ensure 68% exhibited mild or improved curves, 21% moderate curve progression, and 11% severe curve progression. Using the risk score cutoff of 41 or less, the predictive value of a negative test (defined as identification of patients without severe curve progression) was 99% (95% CI, 95.4% to 99.6%). No analysis was performed to demonstrate whether this was a statistically significant improvement in prediction of negatives. In the male cohort (n=163), the prevalence of patients with progression to sever curvature was 11% before testing. The negative predictive value after tests was 97% (95% CI, 93.3% to 99%).
Although there are a description of positive predictive value calculations using a risk score cutoff of 190 or more, recruitment of patients into this category appears to have been derived from patients pooled from different and undescribed sources, making interpretation difficult.
In 2012, Roye et al reported retrospective results for 91 patients evaluated using ScoliScore. Although they noted a positive correlation between Cobb angle and ScoliScore results (r=0.581, p<0.001), ScoliScore appeared to be providing information very different from that observed using a standard risk score, with a marked increase in low-risk patients and a decrease in high-risk patients. However, no clinical end points were examined in association with classification results, and so interpretation of results observed remains unclear.
In 2015, Roye et. al. reported on an independent validation of the ScoliScore algorithm in a sample of 126 patients with AIS who were enrolled at 2 centers using a retrospective cohort design. Eligible patients had AIS with an initial Cobb angle of 10° to 25° and were white with skeletal immaturity. ScoliScore results were provided as continuous and categoric variables; categories were low (1-50 points), intermediate (51-179 points), or high (180-200 points) risk for progression. Outcomes were defined as progression (curve progression to >40° or requirement for spinal fusion) or nonprogression (reached skeletal maturity without curve progression >40°). The mean ScoliScore overall was 103 (SD=60). In unadjusted analysis, the continuous ScoliScore value was not significantly associated with curve progression (odds ratio [OR], 0.999; 95% CI, 0.991 to 1.006; p=0.664). The proportion of patients with curve progression did not differ significantly by ScoliScore risk group. The ScoliScore test PPV and NPV were 0.27 (95% CI, 0.09 to 0.55) and 0.87 (95% CI, 0.69 to 0.96), respectively.
In 2016, Bohl et al reported results from a small retrospective cohort study comparing ScoliScore results among patients with AIS undergoing bracing whose scoliosis progressed to those undergoing bracing who did not have progression. Authors contacted 25 patients with AIS treated at a single institution who underwent nighttime bracing; 16 subjects provided saliva samples to allow ScoliScore testing. Authors reported that the 8 patients whose curves progressed to greater than 45 had a higher mean ScoliScore than those whose curves did not progress (176 vs 112, respectively; p=0.03). No patient with a ScoliScore below 135 progressed to greater than 45 degrees. The interpretation of these results is unclear due to the study’s small size and potential for selective response bias.
Some studies have evaluated subsets of the SNVs (single nucleotide variants) used in the ScoliScore algorithm. Tang et al (2015) evaluated the association between 25 of the 53 SNVs used in the Ward et al study (previously described), along with 27 additional SNVs in high linkage disequilibrium with the other SNVs, and severe scoliosis in a case-control study involving 476 AIS patients of French-Canadian background. None of the SNVs was significantly associated with scoliosis severity.
The ScoliScore algorithm was developed and validated in a sample of white patients. Other studies have evaluated the association of specific SNVs from the algorithm in nonwhite populations.
In 2015, Xu et. al. reported on the association between the 53 SNVs in the ScoliScore panel with scoliosis in a retrospective case-control study of 990 female Han Chinese patients with AIS and 1188 age-matched healthy controls. At 4 loci, patients with AIS differed from controls: they had had higher frequency of alleles G at rs12618119 (46.5% vs 40.2%, OR=1.29; 95% CI, 1.15 to 1.46; p<0.001) and A at rs9945359 (22.6% vs 18.4%; OR=1.29; 95% CI, 1.12 to 1.50; p<0.001), and lower frequency of alleles T at rs4661748 (15.6% vs 19.4%; OR=0.77, 95% CI, 0.66 to 0.90; p<0.001) and C at rs4782809 (42.4% vs 47.4%; OR=0.82, 95% CI, 0.72 to 0.92; p<0.001).
In 2016, Xu et. al. reported on the association between the 53 SNVs in the ScoliScore panel with scoliosis progression in a retrospective case-control study of 670 female Han Chinese patients with AIS. Patients were identified from a set of patients who visited trialists’ scoliosis center for a time period that overlapped with that for the patients in the 2015 Xu study, but it is not specified whether the data overlap. Of the 670 patients, 313 were assigned to the nonprogression group (defined as a Cobb angle <25° at final follow-up) and 357 were assigned to the progression group (defined as a Cobb angle of >40° at final follow-up). The overall follow-up duration was not specified. At 2 loci, allele frequencies differed between groups: the progression group had a significantly higher frequency of allele A at rs9945359 (25.7% vs 19.5%; OR=1.42; 95% CI, 1.09 to 1.88; p=0.01) and a significantly lower frequency of allele A at rs17044552 (11.5% vs 16.4%; OR=0.65; 95% CI, 0.47 to 0.91; p=0.01).
There was no association between the 53 SNVs in the ScoliScore panel and curve progression in an earlier study of 2117 Japanese patients with AIS.
In addition to studies evaluating the clinical validity of the ScoliScore algorithm specifically, other studies have reported results for associations between SNVs and scoliosis progression.
In 2015, Noshchenko et. al. reported on a systematic review and meta-analysis of predictors of progression in AIS, which included studies evaluating the association between ScoliScore and SNVs and curve progression. In total, reviewers included 25 studies, across a range of physiologic measures. Reviewers selected 2 studies that evaluated ScoliScore - Ward et al (2010) and Bohl et al (2016). Pooled results were presented; however, given the differences in intervention in the studies (Bohl et. al. evaluated response to bracing), the results are more appropriately considered as individual studies, which are described above in the Clinical Validity of ScoliScore SNV-Based Testing section. Studies evaluating 6 additional SNVs in multiple genes, including CALM1, ER1, TPH1, IGF1, NTF3, IL17RC, and MTNR1B (N=7 studies) were included. The level of evidence based on GRADE for the studies was considered very low or low. Estimates for the pooled odds ratios for the association of the variant with the outcome ranged from 1.5 to 3.3. Reviewers concluded that “the levels of association were relatively low with small predictive capacity. All these findings have very low level of evidence due to the limitations of the studies’ design and that fact that only one study reported each finding.”
Sharma et al (2011) reported genome-wide association study results evaluating 327,000 SNVs in 419 families with AIS that found 3 loci significantly associated with scoliosis progression, which did not include any of the 53 SNVs included in the Ward et al study previously described.
In 2013, Fendri et al reported results from a case-control study 6 AIS patients and 6 non-AIS controls evaluating differential gene expression profiling in AIS. Gene expression profiles from primary osteoblasts derived from spinal vertebrae of AIS patients (n=6) were compared with profiles from the same cells collected from age- and sex-matched previously healthy patients who underwent spinal surgery for trauma (n=6). One hundred forty-five genes displayed significant expression changes in AIS osteoblasts compared with non-AIS osteoblasts. After hierarchical clustering gene ontology analysis, the authors identified 5 groups based on molecular function and biologic process that fell into 4 pathways: developmental/growth differentiation of skeletal elements (ie, HOXB8, HOXB2, MEOX2, PITX1), cellular signaling (ie, HOXA11, BARX1), connecting structural integrity of the extracellular matrix to the structural integrity of a bone or a muscle fiber (ie, COMP, HOXA2, HOXA11), and cellular signaling and cartilage damage (GDF15).
Four retrospective case-control studies have reported on the clinical validity of the marketed ScoliScore test; 2 of them permitted a determination of the association of the test with curve progression, and they have conflicting results and are limited by their retrospective designs. A number of additional studies have reported on the association between scoliosis progression or presence and various other SNVs, with inconsistent results. The evidence is insufficient to draw conclusions on clinical validity.
A test is clinically useful if use of the results informs management decisions that improve the net health outcome of care. The net health outcome can be improved if patients receive correct therapy, or more effective therapy, or avoid unnecessary therapy or avoid unnecessary testing.
No studies examining the impact of DNA-based predictive testing for scoliosis on net health outcomes were identified. The value of early identification and intervention(s) for people at risk for progression of disease and whether laboratory testing improves disease identification beyond clinical evaluation are unknown. It is not possible to construct a chain of evidence for clinical utility due to the lack of clinical validity.
For individuals with adolescent idiopathic scoliosis (AIS) who receive clinical management with prognostic testing with an algorithm incorporating single-nucleotide variant (SNV) based testing, the evidence includes cross-sectional studies reporting on the clinical validity of the ScoliScore test, along with cross-sectional studies reporting on the association between SNVs in various genes and scoliosis progression. A single study on the clinical validity for the ScoliScore AIS prognostic DNA-based test has reported a high negative predictive value for ruling out the possibility of progression to severe curvature in a population with a low baseline likelihood of progression. It is not clear if the increase in predictive accuracy provided by testing is statistically or clinically meaningful. Other genetic studies have not demonstrated significant associations between the SNVs used in the ScoliScore and scoliosis progression. Studies have identified additional SNVs that may be associated with AIS severity, but these associations have not been reliably replicated. The clinical validity of DNA-based testing (either through testing of individual SNVs or through an algorithm incorporating SNV results) for predicting scoliosis progression in patients with AIS has not been established. There is no direct evidence demonstrating that use of this test results in changes in management that improve outcomes. The value of early identification and intervention(s) for people at risk for progression of disease and whether laboratory testing improves disease identification beyond clinical evaluation is unknown. The evidence is insufficient to determine the effects of the technology on health outcomes.
In 2018, the U.S. Preventative Services Task Force (USPSTF) updated their recommendation regarding adolescent idiopathic scoliosis screening:
The USPSTF recommendation did not discuss the use of DNA-based prognostic testing for adolescent idiopathic scoliosis.
In 2016, the Scientific Society on Scoliosis Orthopaedic and Rehabilitation Treatment (SOSORT) updated their guidelines on the treatment of idiopathic scoliosis. They state the following:
In 2015, the Scoliosis Research Society (SRS), American Academy of Orthopedic Surgeons (AAOS), Pediatric Orthopaedic Society of North America (POSNA) and American Academy of Pediatrics (AAP) issued a position statement on screening for early detection for idiopathic scoliosis in adolescents. This position statement did not address the role of DNA-based prognostic testing.
Clinical laboratories may develop and validate tests in-house and market them as a laboratory service; laboratory developed tests (LDTs) must meet the general regulatory standards of the Clinical Laboratory Improvement Amendments (CLIA). Laboratories that offer LDTs must be licensed by CLIA for high-complexity testing. To date, the U.S. Food and Drug Administration has chosen not to require any regulatory review of this test.
DNA-based prognostic testing for adolescent idiopathic scoliosis (AIS) is considered investigational.
At the current time, there is a lack of evidence in the peer-reviewed medical literature to confirm the clinical utility and validity of DNA-based testing for adolescent idiopathic scoliosis (AIS). There is no direct evidence demonstrating that the use of this test results in changes in management of adolescent idiopathic scoliosis (AIS) that would improve outcomes. In addition, the value of early identification and intervention(s) for individuals at risk for progression of AIS is unclear. Further research of both clinical utility and validity is needed. The evidence is insufficient to determine the effects of the technology on health outcomes.
To report provider services, use appropriate CPT* codes, Alpha Numeric (HCPCS level 2) codes, Revenue codes and / or diagnosis codes.
Wellmark medical policies address the complex issue of technology assessment of new and emerging treatments, devices, drugs, etc. They are developed to assist in administering plan benefits and constitute neither offers of coverage nor medical advice. Wellmark medical policies contain only a partial, general description of plan or program benefits and do not constitute a contract. Wellmark does not provide health care services and, therefore, cannot guarantee any results or outcomes. Participating providers are independent contractors in private practice and are neither employees nor agents of Wellmark or its affiliates. Treating providers are solely responsible for medical advice and treatment of members. Our medical policies may be updated and therefore are subject to change without notice.
*CPT® is a registered trademark of the American Medical Association.