SRS-Lenke-Aubin 3D AIS Classification System

Adolescent Idiopathic Scoliosis (AIS) is a true three-dimensional spinal deformity, involving not only lateral curvature (coronal plane), but also rotational deformity (transverse plane) and changes in kyphosis/lordosis and balance (sagittal plane).

Yet, most clinical evaluations remain 2D-based, primarily relying on Cobb angles and the Lenke classification. This 2D framework does not account for vertebral rotation or regional 3D shape, which are increasingly relevant in the era of modern surgical strategies (e.g., direct vertebral derotation, growth modulation). Also, these ignore critical transverse plane features, which can influence:

• The mechanical function of the spine and trunk
• Cosmetic appearance
• Surgical outcomes

The SRS–Lenke–Aubin 3D classification was developed to:

• Reflect the true spatial complexity of AIS,
• Provide a structured and clinically intuitive 3D language,
• Enhance surgical planning and clinical as well as scientific communication,
• Bridge the gap between 2D practice and 3D-informed interventions.

The SRS–Lenke–Aubin 3D classification is a modular extension of the widely used Lenke classification. It maintains all core components of the Lenke system:

• Curve type (1–6)
• Lumbar spine modifier (A, B, C)
• Thoracic sagittal profile modifier (−, N, +)

And introduces two new transverse plane descriptors for each spinal region (PT, MT, TL/L):

1. ORPD – Orientation of the Regional Plane of Deformation, which captures the regional orientation of each scoliotic curve in the transverse plane,
2. AVR – Apical Vertebral Rotation, which quantifies local vertebral torsion at the apex of each curve

Each is categorized into 3 classes, yielding a 4-modifier system that captures coronal, sagittal, and now axial deformity at both regional and local levels

As a result, the classification expands into a four-modifier system:

Curve Type (1–6) + Lumbar Modifier (A/B/C) + Sagittal Modifier (−/N/+) + Transverse Plane Modifiers, with the transverse component comprising:

• Regional modifier (ORPD: 1, 2, 3) and
• Local modifier (AVR: s, m, l)

applied independently to the proximal thoracic (P), main thoracic (T), and thoracolumbar/lumbar (L) segments.
 

• ORPD (Regional descriptor) measures how much the apical region of the curve (between end vertebrae for PT, MT and TL/L regions) deviates from the sagittal plane. It reflects regional curvature and convexity of the spine.
• AVR (Local descriptor) quantifies the degree of vertebral rotation at the apex of each curve.

Both indices are computed from 3D models reconstructed from biplanar or standard radiographs (i.e. from any radiographic system), and are categorized as follows:

• ORPD
  • 1 = <70° (mostly preserved kyphosis/lordosis)
  • 2 = 70°–90° (flattened sagittal profile)
  • 3 = >90° (inverted sagittal curvature)
• AVR
  • s = slight (<10°)
  • m = moderate (10°–20°)
  • l = large (>20°)

These provide precise information on the rotational and torsional transverse plane characteristics of each spine region, beyond what 2D imaging can show

No. The system is designed to work with any standard radiographic setup. It does not require a proprietary low-dose biplanar system; the 3D measurements can be generated using validated self-calibrated reconstruction methods applied to conventional radiographs.

Enhanced surgical planning

• Identifies patients with high vertebral rotation, torsion or axial deviation who may benefit from specific derotation maneuvers or tailored rod contouring.
• Supports arc-based correction strategies that go beyond uniplanar translation.

Greater diagnostic precision

• Differentiates patients with similar coronal patterns but different 3D morphologies (e.g., Lenke 1 cases with very different rotation profiles).
• Supports personalized treatment decisions, avoiding one-size-fits-all approaches.

Built for clinical use

• Compatible with standard radiographs (JPEG/DICOM), no need for expensive equipment.
• Designed for eventual automation and seamless integration into clinical workflows.

Future-proof

• Scalable for inclusion of additional parameters (e.g., pelvic obliquity, shoulder imbalance).
• Already paired with a companion study analyzing the complementary predictive value of the new 3D descriptors.

Because scoliosis is not a 2D problem. Several studies (and real surgical cases) have shown:

• Patients with similar Cobb angles can have very different axial rotations
• 2D images underestimate the severity and complexity of spinal torsion
• Certain correction failures or suboptimal outcomes are likely due to unrecognized 3D features

As surgery becomes increasingly guided by 3D biomechanical principles and as surgical techniques and technologies continue to evolve, incorporating 3D descriptors is no longer optional—it's essential.

Modern AIS surgery increasingly targets:

• True 3D correction, especially derotation,
• Better restoration of trunk symmetry, not just Cobb angle reduction,
• More personalized alignment goals (based on curve shape, torsion, and growth potential).

However, many patients with similar Lenke types show vastly different 3D morphologies—with variation in axial rotation, apical position, or curve orientation. 2D fails to capture this.
A 3D classification provides:

• A common language to describe this hidden variation,
• A tool for more refined preoperative planning,
• A framework to stratify outcomes and predict behavior of curves,
• A pathway to automated, AI-supported tools for the future.
   

• Better surgical planning: identify cases with high torsion or anterior apices requiring specific derotation strategies.
• More informed fusion level decisions: understand transverse plane involvement per segment.
• Improved patient education: visualize curve complexity in a relatable, intuitive way.
• Outcome research: stratify and correlate results with 3D morphology.
• Training: teach trainees how to reason in 3D with a structured, reproducible framework.
   

The system is conceived primarily for clinical use, and can also be used for research:

• Uses standard imaging (radiographs from any radiographic system, or low-dose biplanar),
• Is designed to use automated 3D reconstructions,
• Was developed to be intuitive, modular, and scalable,
• Has been tested on a large cohort (n=285) and validated for real-world applicability,
• Is part of an ongoing SRS implementation effort, including software support and educational resources.
   

Yes, but on a restricted basis for now. It has been retrospectively applied to 285 surgically treated AIS cases, confirming its robustness, reproducibility, and ability to stratify patients across a wide spectrum of 3D morphologies. All nine combinations of the ORPD × AVR matrix were represented.

In parallel, the SRS 3D Classification Task Force is developing:

• Automated classification tools 
• Interactive educational modules to support clinical learning
• Integration into clinical decision-support systems
• Dedicated training initiatives to educate the spine surgery community
• The system will soon be made accessible to all SRS members

Stay tuned for updates and training opportunities!