511 - Congenital Scoliosis: An Update to Its Natural History and Clinical Approach

Publication Number: 4501.511

Priya Singh, Montefiore Medical Center, Bronx, NY, United States; Mihir Sharma, Albert Einstein College of Medicine, Bronx, NY, United States; Alexander M. Bhatt, Albert Einstein College of Medicine, Bronx, NY, United States; Jason Brenner, Albert Einstein College of Medicine, 15 Elmwood Court, NY, United States; Ashlyn A. Callan, Albert Einstein College of Medicine, Palo Alto, CA, United States; Edina Gjonbalaj, Montefiore Einstein, Department of Orthopaedic Surgery, Bronx, NY, United States; Leila Mehraban Alvandi, Montefiore-Einstein, Bronx, NY, United States; Jacob Schulz, The Children's Hospital at Montefiore, bronx, NY, United States; Jaime A. Gomez, Nicklaus Children’s Hospital, Miami, FL, United States

Background: Congenital scoliosis exhibits a highly variable natural history. Predictive modeling based on radiographic features could improve individualized care.

Objective: The natural history by McMaster and Ohtsuka remains the framework for scoliosis treatment, and we sought to confirm and expand upon it.

Design/Methods: This retrospective multicenter study included 464 patients with congenital scoliosis with baseline and follow-up X-rays. Patients with neurological conditions were excluded. Deformities were classified using McMaster's system. The primary outcome was a Cobb angle ≥50° assessed at final follow-up. A Kaplan-Meier survival analysis was used to analyze curve progression by deformity type, deformity location, and gender. A Chi-squared Automatic Interaction Detection (CHAID) decision tree was used to find predictors of progression based on baseline Cobb angle and deformity type. A Cox regression model was used to assess independent associations between progression and baseline age, deformity type, and gender.

Results: The mean baseline and follow-up ages were 3.5 ± 3.5 years and 8.8 ± 4.7 years, respectively. Most patients underwent surgical correction (55.3%), followed by observation (40.1%), or casting/bracing (4.5%). Baseline and follow-up Cobb angles averaged 46.5° ± 18.9° and 51.3° ± 22.9°, respectively. Kaplan-Meier analysis demonstrated that patients with unilateral unsegmented bars, with or without hemivertebra (HV), were at the highest risk of curve progression. Thoracic and thoracolumbar curves also showed higher progression risk, and female patients were more likely to progress compared with males.

Cox regression identified higher baseline Cobb angle (OR 1.05 per degree increase, p < 0.001), older age at baseline (OR 1.12 per year increase, p < 0.001), and female sex (OR 1.38, p = 0.024) as independent predictors of progression.

CHAID analysis through recursion identified baseline Cobb angle as the strongest predictor.
Cobb ≤38°: Low progression risk (20.8%), especially in block vertebra, incarcerated HV, and complex anomalies (6.9%).
Cobb 38-54°: Moderate progression risk (53.2%), with higher risk in patients with semisegmented HV, wedge, and unilateral unsegmented bar deformities (70.6%).
Cobb >54°: High risk of maintenance at >50° (92.2%) regardless of deformity type.

Conclusion(s): Baseline Cobb angle is a robust predictor of curve progression in congenital scoliosis. Deformity classification adds prognostic value in mid-range curves. These findings offer a clinically applicable framework for early risk stratification and individualized treatment planning.

Figure 1: Deformity Types Used in Classification.
Image was taken from Tachdjian's Pediatric Orthopedics textbook.

Figure 2: Kaplan-Meier Curve for Cobb Angles Greater than 50 Degrees at Last Follow-Up.
The unilateral unsegmented bar and unsegmented bar with contralateral hemivertebra were associated with the greatest risk of progression overtime, whereas incarcerated and block vertebra were associated with the least.

Figure 3: CHAID Analysis Results.
The results show that deformity type is significantly associated with different rates of progression depending on baseline curve size.