Saturday, September 29, 2018 10:30 am–12:00 pm Applying Deformity Concepts in Your Practice

Abstract

BACKGROUND CONTEXT Despite being one of the most common spinal deformities, no classification for Scheuermann's kyphosis (SK) exists to highlight key components of the pathology, produce a common language and guide surgical treatment. Historically, SK patients have been broadly grouped as having a thoracic or thoracolumbar apex. However, recent studies have reported that the SK population is actually markedly more complex, with considerable variability in clinical and radiographic outcomes. PURPOSE To develop a valid, modular, and reproducible classification to group SK patients into clinically-meaningful categories and guide treatment of SK. STUDY DESIGN/SETTING Expert consensus or reliability study, using multicenter radiographic data. PATIENT SAMPLE Multicenter database of 223 SK patients age 21 and younger undergoing deformity correction surgery. OUTCOME MEASURES 1) Content validity ratio quantifying consensus from nominal group technique sessions among experienced deformity surgeons; 2) multivariate cluster analysis of radiographic factors describing SK deformities; 3) intraclass correlation coefficients for determining inter-rater reliability of proposed SK classification. METHODS Using multiple rounds of nominal group technique, an established consensus-building methodology, 15 deformity surgeons identified 3 radiographic parameters most critical to characterize and treat a SK deformity: apex level, last-touched sagittal vertebra (LTSV), and pelvic incidence (PI). A classification system was then formulated and agreed upon: 1. Apex: Thoracic (above T10) versus thoracolumbar (T10 and below) 2. LTSV: T12, L1, L2, L3, L43. Pelvic Incidence: (−) 50° EXAMPLE: T / L3 / -A multicenter radiographic database of 223 SK patients was then used in a multivariate cluster analysis to evaluate whether these variables did indeed predict groupings within the SK population. To calculate interrater reliability, seven surgeons classified a series of 28 SK patients on two occasions, and intraclass correlation coefficients (ICC) were calculated. RESULTS Each of the classification parameters had a content validity ratio of 1.0, indicating perfect consensus achieved through the nominal group technique. Cluster analysis identified that the three classification variables did indeed predict distinct groupings within the SK patients (r=0.5, p CONCLUSIONS We propose a novel, modular and simple classification for Scheuermann's kyphosis (SK) that achieves more complete identification of SK subtypes, and can organize decision-making regarding deformity correction. The three classification variables of apex level, LTSV and PI identified distinct groupings within the SK population and demonstrated near-perfect reliability. Classification formulation and reliability confirmation are essential steps towards establishing a much-needed common language for SK to predict outcomes and guide treatment. FDA DEVICE/DRUG STATUS This abstract does not discuss or include any applicable devices or drugs.