Reliability and reproducibility analysis of the AOSpine thoracolumbar spine injury classification system by Chinese spinal surgeons

From Radiographic Evaluation to Treatment Decisions in Neurologically Intact Patients With Thoraco-lumbar Burst Fractures

We propose that the key to improving care for these patients is to truly understand the processes that take place from the interpretation of radiographic findings, through the assessment of the severity of various injuries, to inclusion within a classification category and finally to selecting a specific treatment.

The AO Spine Thoracolumbar Injury Classification System and Treatment Algorithm in Decision Making for Thoracolumbar Burst Fractures Without Neurologic Deficit

STUDY DESIGN

Prospective Observational Study.

OBJECTIVE

To determine the alignment of the AO Spine Thoracolumbar Injury Classification system and treatment algorithm with contemporary surgical decision making.

METHODS

183 cases of thoracolumbar burst fractures were reviewed by 22 AO Spine Knowledge Forum Trauma experts. These experienced clinicians classified the fracture morphology, integrity of the posterior ligamentous complex and degree of comminution. Management recommendations were collected.

RESULTS

There was a statistically significant stepwise increase in rates of operative management with escalating category of injury (P < .001). An excellent correlation existed between recommended expert management and the actual treatment of each injury category: A0/A1/A2 (OR 1.09, 95% CI 0.70-1.69, P = .71), A3/4 (OR 1.62, 95% CI 0.98-2.66, P = .58) and B1/B2/C (1.00, 95% CI 0.87-1.14, P = .99). Thoracolumbar A4 fractures were more likely to be surgically stabilized than A3 fractures (68.2% vs 30.9%, P < .001). A modifier indicating indeterminate ligamentous injury increased the rate of operative management when comparing type B and C injuries to type A3/A4 injuries (OR 39.19, 95% CI 20.84-73.69, P < .01 vs OR 27.72, 95% CI 14.68-52.33, P < .01).

CONCLUSIONS

The AO Spine Thoracolumbar Injury Classification system introduces fracture morphology in a rational and hierarchical manner of escalating severity. Thoracolumbar A4 complete burst fractures were more likely to be operatively managed than A3 fractures. Flexion-distraction type B injuries and translational type C injuries were much more likely to have surgery recommended than type A fractures regardless of the M1 modifier. A suspected posterior ligamentous injury increased the likelihood of surgeons favoring surgical stabilization.

Automated detection and classification of acute vertebral body fractures using a convolutional neural network on computed tomography

BACKGROUND

Acute vertebral fracture is usually caused by low-energy injury with osteoporosis and high-energy trauma. The AOSpine thoracolumbar spine injury classification system (AO classification) plays an important role in the diagnosis and treatment of the disease. The diagnosis and description of vertebral fractures according to the classification scheme requires a great deal of time and energy for radiologists.

PURPOSE

To design and validate a multistage deep learning system (multistage AO system) for the automatic detection, localization and classification of acute thoracolumbar vertebral body fractures according to AO classification on computed tomography.

MATERIALS AND METHODS

The CT images of 1,217 patients who came to our hospital from January 2015 to December 2019 were collected retrospectively. The fractures were marked and classified by 2 junior radiology residents according to the type A standard in the AO classification. Marked fracture sites included the upper endplate, lower endplate and posterior wall. When there were inconsistent opinions on classification labels, the final result was determined by a director radiologist. We integrated different networks into different stages of the overall framework. U-net and a graph convolutional neural network (U-GCN) are used to realize the location and classification of the thoracolumbar spine. Next, a classification network is used to detect whether the thoracolumbar spine has a fracture. In the third stage, we detect fractures in different parts of the thoracolumbar spine by using a multibranch output network and finally obtain the AO types.

RESULTS

The mean age of the patients was 61.87 years with a standard deviation of 17.04 years, consisting of 760 female patients and 457 male patients. On vertebrae level, sensitivity for fracture detection was 95.23% in test dataset, with an accuracy of 97.93% and a specificity of 98.35%. For the classification of vertebral body fractures, the balanced accuracy was 79.56%, with an AUC of 0.904 for type A1, 0.945 for type A2, 0.878 for type A3 and 0.942 for type A4.

CONCLUSION

The multistage AO system can automatically detect and classify acute vertebral body fractures in the thoracolumbar spine on CT images according to AO classification with high accuracy.

Reliability and reproducibility analysis of the AOSpine Sacral Fractures Classification System by spinal and pelvic surgeons

STUDY DESIGN

Retrospective Cohort Study OBJECTIVES: The AOSpine Sacral Classification System was proposed as a comprehensive and universally accepted new classification for Sacral Fractures, and was recently internally validated. However, an external, independent and multidisciplinary reports on validation of this classification is lacking. Aim of the present study is to analyze the interobserver reliability and intraobserver reproducibility of the AOSpine Sacral Classification System for Sacral Fractures between orthopedic spinal and pelvic surgeons with different levels of experience.

METHODS

Our institutional database was searched to retrieve patients with acute, traumatic sacral injury admitted from June 2017 to June 2020. For each patients, X-Rays and CT scans were collected. Three Orthopedic Pelvic Surgeons (Group A) and three Spine Surgeons (Group B), with different level of experience (Junior, 〈 5 years; Middle, 5-10 years; Expert 〉 10 years) independently classified all the sacral fractures included in the dataset, with two separate evaluation three weeks apart. Both intra and interobserver reliability were calculated with k-coefficient.

RESULTS

Overall, 150 patients were included in the final dataset, for a total of 1800 different assessments, with all the subtypes reported. The intraobserver reproducibility for the whole group was substantial (κ=0.72). Overall, the interobserver reliability was moderate, with a κ=0.57. When only fracture type was taken in account, the κ value became substantial (κ=0.62). No significant differences were found comparing group A and group B (0.55 vs κ 0.55, p>0.05). No significant differences according to surgeon's experience were found; however, the κ value was slightly lower among the junior surgeons.

CONCLUSIONS

Our findings confirmed the reliability and reproducibility of this classification in clinical practice. In the current study the surgeon's expertise (pelvic and spinal trauma) and the level of experience does not influence the reliability of the classification system.

Is the AO spine thoracolumbar injury classification system reliable and practical? a systematic review

Controversy surrounding the classification of thoracolumbar injuries has given rise to various classification systems over the years, including the most recent AOSpine Thoracolumbar Injury Classification System (ATLICS). This systematic review aims to provide an up-to-date evaluation of the literature, including assessment of a further three studies not analysed in previous reviews. In doing so, this is the first systematic review to include the reliability among non-spine subspecialty professionals and to document the wide variety between reliability across studies, particularly with regard to sub-type classification. Relevant studies were found via a systematic search of PubMed, EBESCO, Cochrane and Web of Science. Data extraction and quality assessment were conducted in line with Cochrane Collaboration guidelines. Twelve articles assessing the reliability of ATLICS were included in this review. The overall inter-observer reliability varied from fair to substantial, but the three additional studies in this review, compared to previous reviews, presented on average only fair reliability. The greatest variation of results was seen in A1 and B3 subtypes. Least reliably classified on average was A4 subtype. This systematic review concludes that ATLICS is reliable for the majority of injuries, but the variability within subtypes suggests the need for further research in assessing the needs of users in order to increase familiarity with ATLICS or perhaps the necessity to include more subtype-specific criteria into the system. Further research is also recommended on the reliability of modifiers, neurological classification and the application of ATLICS in a paediatric context.

Myelomeningocele: a new functional classification

PURPOSE

In myelomeningocele, several classifications have been used. The present manuscript proposes a new functional classification to better assess the prognosis and management of these patients.

METHODS

The manual muscle test is what defines the actual group in which the patient should be included. Furthermore, this new classification brings information about the bracing and external supports recommended to each functional level. We also recommend that the patient's Functional Mobility Scale should always be mentioned together with their functional level.

RESULTS

The four levels in this classification are MMFC1, MMFC2, MMFC3 and MMFC4. The MMFC1 group includes patients with significant muscle weakness. They need to use high braces crossing the hip joint with a walker to achieve some ambulation. The MMFC2 group includes patients who have functional hip flexors, knee extensors and knee flexors. However, the hip abductors are quite weak. These patients usually need to use a walker - or crutches - and Ankle-Foot Orthosis (AFOs). The MMFC3 group includes patients with functional hip flexors, knee extensors, knee flexors and hip abductors. However, the ankle plantar flexion function is absent. Most of them are able to walk independently, only using AFOs without any external support. The MMFC4 group includes patients who have preserved function in the entire lower limb musculature. These patients don't need any assistive devices to achieve an adequate ambulation pattern.

CONCLUSIONS

We hope that this new classification is a system that is simple to understand, serves as a gait prognosis guide and facilitates communication among healthcare professionals.

LEVEL OF EVIDENCE

V.

Reliability and Clinical Usefulness of Current Classifications in Traumatic Thoracolumbar Fractures: A Systematic Review of the Literature

BACKGROUND

A validated classification remains the key to an appropriate treatment algorithm of traumatic thoracolumbar fractures. Considering the development of many classifications, it is remarkable that consensus about treatment is still lacking. We conducted a systematic review to investigate which classification can be used best for treatment decision making in thoracolumbar fractures.

METHODS

A comprehensive search was conducted using PubMed, Embase, CINAHL, and Cochrane using the following search terms: classification (mesh), spinal fractures (mesh), and corresponding synonyms. All hits were viewed by 2 independent researchers. Papers were included if analyzing the reliability (kappa values) and clinical usefulness (specificity or sensitivity of an algorithm) of currently most used classifications (Magerl/AO, thoracolumbar injury classification and severity score [TLICS] or thoracolumbar injury severity score, and the new AO spine).

RESULTS

Twenty articles are included. The presented kappa values indicate moderate to substantial agreement for all 3 classifications. Regarding the clinical usefulness, > 90% agreement between actual treatment and classification recommendation is reported for most fractures. However, it appears that over 50% of the patients with a stable burst fracture (TLICS 2, AO-A3/A4) in daily practice are operated, so in these cases treatment decision is not primarily based on classification.

CONCLUSION

AO, TLICS, and new AO spine classifications have acceptable accuracy (kappa > 0.4), but are limited in clinical usefulness since the treatment recommendation is not always implemented in clinical practice. Differences in treatment decision making arise from several causes, such as surgeon and patient preferences and prognostic factors that are not included in classifications yet. The recently validated thoracolumbar AO spine injury score seems promising for use in clinical practice, because of inclusion of patient-specific modifiers. Future research should prove its definite value in treatment decision making.

LEVEL OF EVIDENCE

2.

CLINICAL RELEVANCE

Without the appropriate treatment, the impact of traumatic thoracolumbar fractures can be devastating. Therefore it is important to achieve consensus in the treatment of thoracolumbar fractures.

Inter- and intra-rater reliability of vertebral fracture classifications in the Swedish fracture register

AIM

To investigate the inter- and intra-rater reliability of the vertebral fracture classifications used in the Swedish fracture register.

METHODS

Radiological images of consecutive patients with cervical spine fractures (n = 50) were classified by 5 raters with different experience levels at two occasions. An identical process was performed with thoracolumbar fractures (n = 50). Cohen’s kappa was used to calculate the inter- and intra-rater reliability.

RESULTS

The mean kappa coefficient for inter-rater reliability ranged between 0.54 and 0.79 for the cervical fracture classifications, between 0.51 and 0.72 for the thoracolumbar classifications (overall and for different sub classifications), and between 0.65 and 0.77 for the presence or absence of signs of ankylosing disorder in the fracture area. The mean kappa coefficient for intra-rater reliability ranged between 0.58 and 0.80 for the cervical fracture classifications, between 0.46 and 0.68 for the thoracolumbar fracture classifications (overall and for different sub classifications) and between 0.79 and 0.81 for the presence or absence of signs of ankylosing disorder in the fracture area.

CONCLUSION

The classifications used in the Swedish fracture register for vertebral fractures have an acceptable inter- and intra-rater reliability with a moderate strength of agreement.

Reliability and Agreement of Different Spine Fracture Classification Systems: An Independent Intraobserver and Interobserver Study

OBJECTIVE

Currently, no spinal classification system has achieved universal acceptance. Therefore, it is important to choose a reliable classification within clinical practice. The objective of this study was to determine and compare the intraobserver and interobserver agreement of the Load Sharing Classification (LSC), the Thoracolumbar Injury Classification System (TLICS), and the AOSpine Thoracolumbar Spine Injury Classification System.

METHODS

In this web-based intraobserver and interobserver study (www.spine.hostei.com), plain radiographs and computed tomographic scans of traumatic thoracolumbar fractures (T12-L2) were evaluated. By use of a questionnaire, fractures were classified according to the LSC, the TLICS, and the AOSpine classification. Data were analyzed with SPSS (Version 21, 76 Chicago, Illinois, USA). Intraobserver and interobserver agreement was determined by the Cohen κ. Statistical significance was defined as P < 0.05.

RESULTS

Data from 91 patients were classified twice by 7 board-certified spine surgeons. The intraobserver and interobserver reliability considering the LSC total score was noted as fair (intraobserver/interobserver reliability: κ = 0.26/0.22). Considering the resulting TLICS total score, a moderate intraobserver agreement (κ = 0.41) was noted, whereas the interobserver results presented only fair reliability (κ = 0.23). In contrast to the LSC and the TLICS, the AOSpine classification showed substantial agreement considering the fracture type (A;B;C) (intraobserver/interobserver reliability: κ = 0.71/0.61) and moderate agreement considering the fracture subtype (e.g., A0;A1;…;B1;…) (intraobserver/interobserver reliability: κ = 0.57/0.48).

CONCLUSION

In conclusion, the reliability of the AOSpine fracture classification is superior to the TLICS and the LSC. Therefore, this classification system could best be applied within clinical practice.

COMPARATIVE ANALYSIS OF TWO CLASSIFICATION SYSTEMS OF THORACOLUMBAR SPINE FRACTURES

ABSTRACT Objective: To evaluate the reproducibility and the interobserver coefficient of concordance between the AO/Magerl and AOSpine classifications for thoracolumbar spine fractures. Methods: Retrospective study of radiographic data analysis. Data were collected from 31 radiographic studies of patients with thoracolumbar spine fracture and distributed to a team involving spinal surgeons and residents. The fractures were classified according to the AO/Magerl and AOSpine classifications. Statistical analysis was performed using the Cohen Kappa test to assess the coefficient of concordance. Results: The Kappa value for interobserver concordance of AO/Magerl classification was κ = 0.70 and standard deviation was 0.16. For the AOSpine classification, we observed κ = 0.76, both with significance level α = 0.05 and P<0.001. Conclusions: We conclude that the interobserver concordance of the new AOSpine classification is similar to the AO/Magerl classification. This conclusion reinforces the reproducibility of the new AOSpine classification. Level of evidence: IV,Type of Study: Case series.

AOSpine Classification Systems (Subaxial, Thoracolumbar)

Numerous classification systems for subaxial and thoracolumbar spine injuries were proposed in the past with the attempt to facilitate communication between physicians. The AO-Magerl, thoracolumbar system, and Subaxial Cervical Spine Injury Classification systems are all well known, but did not achieve universal international adoption. A group of international experienced spine trauma surgeons were brought together by AOSpine with the goal to develop a comprehensive yet simple classification system for spinal trauma. This article is a synopsis of the proposed subaxial and thoracolumbar classification systems. In several studies, this classification system was developed using an iterative consensus process among the clinical experts in sufficient number and quality of DICOM images of real cases searching for meaningful and reproducible patterns. Both systems are based on 3 injury morphology types: compression injuries (A), tension band injuries (B), and translational injuries (C) with a total of 9 subgroups. In the subaxial cervical spine 4 additional subtypes for facet injuries exist. Patient-specific modifiers and neurologic status were also included to aid surgeons in therapeutic decision making. The proposed classification systems for subaxial and thoracolumbar injuries showed substantial intraobserver and interobserver reliability (κ = 0.64-0.85) for grading fracture type. Grading for the subtypes varied considerably due to the low frequency of certain injury subtypes among other reasons. In summary, the AOSpine thoracolumbar and subaxial cervical spine injury systems show substantial reliability, thus being valuable tools for clinical and research purposes.