Adolescent idiopathic scoliosis: Lenke type I-VI case studies

Automatic Lenke classification of adolescent idiopathic scoliosis with deep learning

Purpose

The Lenke classification system is widely utilized as the preoperative evaluation protocol for adolescent idiopathic scoliosis (AIS). However, manual measurement is susceptible to observer-induced variability, which consequently impacts the evaluation of progression. The goal of this investigation was to develop an automated Lenke classification system utilizing innovative deep learning algorithms.

Methods

Using the database from the First Affiliated Hospital of Sun Yat-sen University, the whole spinal x-rays images were retrospectively collected. Specifically, images collection was divided into AIS and control group. The control group consisted of individuals who underwent routine health checks and did not have scoliosis. Afterwards, relative features of all images were annotated. Deep learning was implemented through the utilization of the key-point based detection method to realize the vertebral detection, and Cobb angle measurement and scoliosis classification were performed based on relevant standards. Besides, the segmentation method was employed to achieve the recognition of lumbar vertebral pedicle to determine the type of lumbar spine modifier. Finally, the model performance was further quantitatively analyzed.

Results

In the study, a total of 2082 spinal x-ray images were collected from 407 AIS patients and 227 individuals in the control group. The model for vertebral detection achieved an F1-score of 0.809 for curve type evaluation and an F1-score of 0.901 for thoracic sagittal profile. The intraclass correlation efficient (ICC) of the Cobb angle measurement was 0.925. In the analysis of performance for vertebra pedicle segmentation model, the F1-score of lumbar modification profile was 0.942, the intersection over union (IOU) of the target pixels was 0.827, and the Hausdorff distance (HD) was 6.565 ± 2.583 mm. Specifically, the F1-score for ultimate Lenke type classifier was 0.885.

Conclusions

This study has constructed an automated Lenke classification system by employing the deep learning networks to achieve the recognition pattern and feature extraction. Our models require further validation in additional cases in the future.

Does anterior shoulder balance in adolescent idiopathic scoliosis correlate with posterior shoulder balance clinically and radiographically?

PURPOSE

Patients with adolescent idiopathic scoliosis (AIS) often present with a disfiguring shoulder imbalance. Shoulder balance (Sh.B) is of significant importance to the patient's self-perception. Previous studies have correlated Sh.B with respect to only the clinical posterior view correlated with radiographs. It is important, however, to address Sh.B with respect to anterior view of the patients' shoulders as if patients were viewing in a mirror. In this study, we evaluated the anterior Sh.B and correlated it with posterior Sh.B clinically and radiographically in Lenke type 1 and 2 curves.

METHOD

An online scoliosis database was queried to identify 74 AIS patients with Lenke 1 (n = 55, age 15.28 ± 3.35) and 2 (n = 19, age 15.66 ± 3.72) curves with a complete set of PA radiographs and anterior and posterior photos. Radiographic measures for Sh.B included Cobb angles, T1 tilt, first rib angle, and clavicle-rib intersection angle. Clinical measures for Sh.B included inner shoulder angle, outer shoulder angle, and axillary fold angle. Regression analysis with Pearson's correlation and ANOVA for statistical significance was used for analysis.

RESULTS

For Lenke 1 curves, there was moderate statistically significant correlation between anterior and posterior clinical Sh.B (R = 0.35-0.41). There was only weak to moderate correlation between radiographic and clinical measures. For Lenke 2 curves, there was a weak to moderate correlation between anterior and posterior clinical Sh.B (R = 0.25-0.45), though not statistically significant. There was no statistically significant correlation between any radiographic measures and posterior Sh.B. There was, however, moderate and significant correlation between radiographic measures and anterior Sh.B.

CONCLUSION

There is no strong correlation between anterior and posterior clinical Sh.B, and surgeons should evaluate both sides in planning deformity correction, especially in Lenke 2 curves. None of the radiographic measures showed strong correlation (R > 0.8) with anterior or posterior clinical Sh.B. A stronger correlation existed between radiographic measures and anterior Sh.B measurements compared with posterior clinical Sh.B measurements in Lenke 2 curves further necessitating anterior evaluation in this group.

Subjective evaluation of treatment outcomes of instrumentation with pedicle screws or hybrid constructs in Lenke Type 1 and 2 adolescent idiopathic scoliosis: what happens when judges are blinded to the instrumentation?

Superiority of pedicle screws over hybrid/hook instrumentation or vice versa in the treatment of Lenke Type 1 and 2 adolescent idiopathic scoliosis (AIS) remains unresolved for moderate curves. Our objective was therefore to compare the assessment of pedicle screw and hybrid/hooks instrumentation with special attention to cosmesis and uninstrumented spine using novel assessment methods. We carried out a retrospective study of radiographs and clinical photos of 40 cases of thoracic AIS between 40 degrees and 70 degrees of Cobb angle Lenke Type 1 and 2, treated with either pedicle screws or hybrid/hooks. The cases were subjectively assessed by four spine surgeons (SRS Travelling Fellows) for radiographic and operative cosmetic result, shoulder balance, trunk shift, rib hump, and waist asymmetry. Instrumentation in the radiographs was obscured with only the non-instrumented part visible, and the surgeons were asked to guess the instrumentation being used. Eighty photographs of patients before and after surgery were assessed for cosmesis by ten non-medical judges for overall cosmetic score, shoulder balance, waist asymmetry, and shoulder blade prominence. Objective assessment of radiographs and clinical photos was performed for Cobb angle of instrumented and non-instrumented spine, global coronal and sagittal balance, number of unfused vertebrae, disc angulation, tilt of last instrumented vertebra, shoulder balance, waist asymmetry, rib prominence, and percent correction. SRS-24 questionnaire was used to measure health-related quality of life in patients. Subjective assessments by surgeons and non-medical judges showed no significant difference by instrumentation (P > or = 0.05) for all variables. Out of the 160 guesses by surgeons of the cases with instrumentation blocked in the radiographs, they were unable to guess the instrumentation in 92% of the cases. Objective assessment of all variables and SRS-24 scores of all five domains showed no significant difference by instrumentation (P > or = 0.05). In this first-ever conducted study in a blinded-fashion, we conclude that there is no significant difference between the pedicle screw and hybrid/hooks instrumentations used to treat AIS for Lenke Type 1 and 2 curves for moderate curves between 40 degrees and 70 degrees .