Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-0263
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-07-19T00:56:47.840601Z
status: published
lens: g_space
sector: health
headline: Predicting postoperative coronal imbalance in Lenke 1/2 adolescent idiopathic scoliosis: A machine learning model with clinical interpretability
dek: Selective posterior thoracic fusion (sPTF) for Lenke 1/2 adolescent idiopathic scoliosis (AIS) aims to reconcile multi-planar correction with motion preservation. Nevertheless, postoperative coronal imbalance (CIB) frequently compromises these objectives. This study developed an interpretable machine learning architecture to stratify CIB risk and identify key predictors. Data from 282 patients were analyzed. Following dual-stage dimensionality reduction (Boruta and LASSO) on 24 candidate predictors, ten machine…
gain_title: LightGBM model trained on 282 Lenke 1/2 AIS patients predicted postoperative coronal imbalance with AUC 0.885 training and 0.824 internal validation, identifying LIV-LSTV, Lumbar Modifier, and Risser grade as key predictors.
problem_title: (none)
trace_subject: (none)
gain_reading: LightGBM model trained on 282 Lenke 1/2 AIS patients predicted postoperative coronal imbalance with AUC 0.885 training and 0.824 internal validation, identifying LIV-LSTV, Lumbar Modifier, and Risser grade as key predictors.
gain_evidence: The LightGBM model demonstrated favorable performance, achieving peak AUC of 0.885 (training) and 0.824 (internal validation) | SHAP identified three key predictors: the spatial relationship between the lowest instrumented vertebra and the last substantially touching vertebra (LIV-LSTV), regional lumbar adaptability (Lumbar Modifier), and skeletal maturity (Risser grade)
problem_reading: (none)
problem_evidence: (none)
quick_read: In 282 patients with Lenke 1/2 adolescent idiopathic scoliosis treated with selective posterior thoracic fusion, investigators built an interpretable machine learning pipeline to stratify risk of postoperative coronal imbalance. After reducing 24 candidates to key features, a LightGBM model achieved the highest discrimination with AUC 0.885 in training and 0.824 in internal validation, with SHAP highlighting LIV-LSTV relationship, Lumbar Modifier, and Risser grade.

The work matters because coronal imbalance frequently compromises correction and motion preservation goals in this population, and a transparent risk estimate could inform distal fusion level choice. As of the July 2026 publication date, performance is limited to internal validation, so generalizability, calibration in other centers, and impact on surgical decisions remain unproven.
limitation: Model has only split-sample internal validation and lacks independent external validation needed for clinical use.
tag: Evidence-backed gain
key_points: Study analyzed 282 patients undergoing selective posterior thoracic fusion for Lenke 1/2 adolescent idiopathic scoliosis. | Dual-stage dimensionality reduction using Boruta and LASSO reduced 24 candidate predictors before training ten machine learning architectures. | SHAP analysis found lower LIV-LSTV values, Lumbar Modifiers C, and lower Risser grades were associated with higher predicted risk of CIB.
rundown: Researchers applied Boruta and LASSO to 24 candidate predictors, then trained and evaluated ten machine learning architectures using split-sample internal validation with AUC, Brier score, and decision curve analysis.

The framework is presented as incorporating distal instrumentation selection, lumbar curve morphology, and skeletal maturity to support assessment of the trade-off between deformity correction and preservation of lumbar motion segments.
sources:
- peer_reviewed | International Orthopaedics | https://doi.org/10.1007/s00264-026-06953-6 | 2026-07-18
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

7e183f1ddae670fcbb7395a3e7c96b939eedb15a6d3f3b3506a5554eca8fa102

previous

0000000000000000000000000000000000000000000000000000000000000000