Development of validated computer-based preoperative predictive model for proximal junction failure (PJF) or clinically significant PJK with 86% accuracy based on 510 ASD patients with 2-year follow-up

Justin K. Scheer, Joseph A. Osorio, Justin S. Smith, Frank Schwab, Virginie Lafage, Robert A. Hart, Shay Bess, Breton Line, Bassel G. Diebo, Themistocles S. Protopsaltis, Amit Jain, Tamir Ailon, Douglas C. Burton, Christopher I. Shaffrey, Eric Klineberg, Christopher P. Ames

Research output: Contribution to journalArticle

Abstract

Study Design. A retrospective review of large, multicenter adult spinal deformity (ASD) database. Objective. The aim of this study was to build a model based on baseline demographic, radiographic, and surgical factors that can predict clinically significant proximal junctional kyphosis (PJK) and proximal junctional failure (PJF). Summary of Background Data. PJF and PJK are significant complications and it remains unclear what are the specific drivers behind the development of either. There exists no predictive model that could potentially aid in the clinical decision making for adult patients undergoing deformity correction. Methods. Inclusion criteria: age -18 years, ASD, at least four levels fused. Variables included in the model were demographics, primary/revision, use of three-column osteotomy, upper-most instrumented vertebra (UIV)/lower-most instrumented vertebra (LIV) levels and UIV implant type (screw, hooks), number of levels fused, and baseline sagittal radiographs [pelvic tilt (PT), pelvic incidence and lumbar lordosis (PI-LL), thoracic kyphosis (TK), and sagittal vertical axis (SVA)]. PJK was defined as an increase from baseline of proximal junctional angle -208 with concomitant deterioration of at least one SRS-Schwab sagittal modifier grade from 6 weeks postop. PJF was defined as requiring revision for PJK. An ensemble of decision trees were constructed using the C5.0 algorithm with five different bootstrapped models, and internally validated via a 70 : 30 data split for training and testing. Accuracy and the area under a receiver operator characteristic curve (AUC) were calculated. Results. Five hundred ten patients were included, with 357 for model training and 153 as testing targets (PJF: 37, PJK: 102). The overall model accuracy was 86.3% with an AUC of 0.89 indicating a good model fit. The seven strongest (importance ≥0.95) predictors were age, LIV, pre-operative SVA, UIV implant type, UIV, pre-operative PT, and pre-operative PI-LL. Conclusion. A successful model (86% accuracy, 0.89 AUC) was built predicting either PJF or clinically significant PJK. This model can set the groundwork for preop point of care decision making, risk stratification, and need for prophylactic strategies for patients undergoing ASD surgery.

Original languageEnglish (US)
Pages (from-to)E1328-E1335
JournalSpine
Volume41
Issue number22
DOIs
StatePublished - Nov 15 2016

Fingerprint

Kyphosis
Spine
Area Under Curve
Lordosis
Point-of-Care Systems
Demography
Decision Trees
Incidence
Osteotomy
Decision Making
Thorax
Databases

Keywords

  • Adult spinal deformity
  • Predictive modeling
  • Proximal junctional failure
  • Proximal junctional kyphosis
  • Sagittal malalignment
  • Scoliosis

ASJC Scopus subject areas

  • Orthopedics and Sports Medicine
  • Clinical Neurology

Cite this

Development of validated computer-based preoperative predictive model for proximal junction failure (PJF) or clinically significant PJK with 86% accuracy based on 510 ASD patients with 2-year follow-up. / Scheer, Justin K.; Osorio, Joseph A.; Smith, Justin S.; Schwab, Frank; Lafage, Virginie; Hart, Robert A.; Bess, Shay; Line, Breton; Diebo, Bassel G.; Protopsaltis, Themistocles S.; Jain, Amit; Ailon, Tamir; Burton, Douglas C.; Shaffrey, Christopher I.; Klineberg, Eric; Ames, Christopher P.

In: Spine, Vol. 41, No. 22, 15.11.2016, p. E1328-E1335.

Research output: Contribution to journalArticle

Scheer, JK, Osorio, JA, Smith, JS, Schwab, F, Lafage, V, Hart, RA, Bess, S, Line, B, Diebo, BG, Protopsaltis, TS, Jain, A, Ailon, T, Burton, DC, Shaffrey, CI, Klineberg, E & Ames, CP 2016, 'Development of validated computer-based preoperative predictive model for proximal junction failure (PJF) or clinically significant PJK with 86% accuracy based on 510 ASD patients with 2-year follow-up', Spine, vol. 41, no. 22, pp. E1328-E1335. https://doi.org/10.1097/BRS.0000000000001598
Scheer, Justin K. ; Osorio, Joseph A. ; Smith, Justin S. ; Schwab, Frank ; Lafage, Virginie ; Hart, Robert A. ; Bess, Shay ; Line, Breton ; Diebo, Bassel G. ; Protopsaltis, Themistocles S. ; Jain, Amit ; Ailon, Tamir ; Burton, Douglas C. ; Shaffrey, Christopher I. ; Klineberg, Eric ; Ames, Christopher P. / Development of validated computer-based preoperative predictive model for proximal junction failure (PJF) or clinically significant PJK with 86% accuracy based on 510 ASD patients with 2-year follow-up. In: Spine. 2016 ; Vol. 41, No. 22. pp. E1328-E1335.
@article{fedee206ee6240c0906c680f04da0353,
title = "Development of validated computer-based preoperative predictive model for proximal junction failure (PJF) or clinically significant PJK with 86{\%} accuracy based on 510 ASD patients with 2-year follow-up",
abstract = "Study Design. A retrospective review of large, multicenter adult spinal deformity (ASD) database. Objective. The aim of this study was to build a model based on baseline demographic, radiographic, and surgical factors that can predict clinically significant proximal junctional kyphosis (PJK) and proximal junctional failure (PJF). Summary of Background Data. PJF and PJK are significant complications and it remains unclear what are the specific drivers behind the development of either. There exists no predictive model that could potentially aid in the clinical decision making for adult patients undergoing deformity correction. Methods. Inclusion criteria: age -18 years, ASD, at least four levels fused. Variables included in the model were demographics, primary/revision, use of three-column osteotomy, upper-most instrumented vertebra (UIV)/lower-most instrumented vertebra (LIV) levels and UIV implant type (screw, hooks), number of levels fused, and baseline sagittal radiographs [pelvic tilt (PT), pelvic incidence and lumbar lordosis (PI-LL), thoracic kyphosis (TK), and sagittal vertical axis (SVA)]. PJK was defined as an increase from baseline of proximal junctional angle -208 with concomitant deterioration of at least one SRS-Schwab sagittal modifier grade from 6 weeks postop. PJF was defined as requiring revision for PJK. An ensemble of decision trees were constructed using the C5.0 algorithm with five different bootstrapped models, and internally validated via a 70 : 30 data split for training and testing. Accuracy and the area under a receiver operator characteristic curve (AUC) were calculated. Results. Five hundred ten patients were included, with 357 for model training and 153 as testing targets (PJF: 37, PJK: 102). The overall model accuracy was 86.3{\%} with an AUC of 0.89 indicating a good model fit. The seven strongest (importance ≥0.95) predictors were age, LIV, pre-operative SVA, UIV implant type, UIV, pre-operative PT, and pre-operative PI-LL. Conclusion. A successful model (86{\%} accuracy, 0.89 AUC) was built predicting either PJF or clinically significant PJK. This model can set the groundwork for preop point of care decision making, risk stratification, and need for prophylactic strategies for patients undergoing ASD surgery.",
keywords = "Adult spinal deformity, Predictive modeling, Proximal junctional failure, Proximal junctional kyphosis, Sagittal malalignment, Scoliosis",
author = "Scheer, {Justin K.} and Osorio, {Joseph A.} and Smith, {Justin S.} and Frank Schwab and Virginie Lafage and Hart, {Robert A.} and Shay Bess and Breton Line and Diebo, {Bassel G.} and Protopsaltis, {Themistocles S.} and Amit Jain and Tamir Ailon and Burton, {Douglas C.} and Shaffrey, {Christopher I.} and Eric Klineberg and Ames, {Christopher P.}",
year = "2016",
month = "11",
day = "15",
doi = "10.1097/BRS.0000000000001598",
language = "English (US)",
volume = "41",
pages = "E1328--E1335",
journal = "Spine",
issn = "0362-2436",
publisher = "Lippincott Williams and Wilkins",
number = "22",

}

TY - JOUR

T1 - Development of validated computer-based preoperative predictive model for proximal junction failure (PJF) or clinically significant PJK with 86% accuracy based on 510 ASD patients with 2-year follow-up

AU - Scheer, Justin K.

AU - Osorio, Joseph A.

AU - Smith, Justin S.

AU - Schwab, Frank

AU - Lafage, Virginie

AU - Hart, Robert A.

AU - Bess, Shay

AU - Line, Breton

AU - Diebo, Bassel G.

AU - Protopsaltis, Themistocles S.

AU - Jain, Amit

AU - Ailon, Tamir

AU - Burton, Douglas C.

AU - Shaffrey, Christopher I.

AU - Klineberg, Eric

AU - Ames, Christopher P.

PY - 2016/11/15

Y1 - 2016/11/15

N2 - Study Design. A retrospective review of large, multicenter adult spinal deformity (ASD) database. Objective. The aim of this study was to build a model based on baseline demographic, radiographic, and surgical factors that can predict clinically significant proximal junctional kyphosis (PJK) and proximal junctional failure (PJF). Summary of Background Data. PJF and PJK are significant complications and it remains unclear what are the specific drivers behind the development of either. There exists no predictive model that could potentially aid in the clinical decision making for adult patients undergoing deformity correction. Methods. Inclusion criteria: age -18 years, ASD, at least four levels fused. Variables included in the model were demographics, primary/revision, use of three-column osteotomy, upper-most instrumented vertebra (UIV)/lower-most instrumented vertebra (LIV) levels and UIV implant type (screw, hooks), number of levels fused, and baseline sagittal radiographs [pelvic tilt (PT), pelvic incidence and lumbar lordosis (PI-LL), thoracic kyphosis (TK), and sagittal vertical axis (SVA)]. PJK was defined as an increase from baseline of proximal junctional angle -208 with concomitant deterioration of at least one SRS-Schwab sagittal modifier grade from 6 weeks postop. PJF was defined as requiring revision for PJK. An ensemble of decision trees were constructed using the C5.0 algorithm with five different bootstrapped models, and internally validated via a 70 : 30 data split for training and testing. Accuracy and the area under a receiver operator characteristic curve (AUC) were calculated. Results. Five hundred ten patients were included, with 357 for model training and 153 as testing targets (PJF: 37, PJK: 102). The overall model accuracy was 86.3% with an AUC of 0.89 indicating a good model fit. The seven strongest (importance ≥0.95) predictors were age, LIV, pre-operative SVA, UIV implant type, UIV, pre-operative PT, and pre-operative PI-LL. Conclusion. A successful model (86% accuracy, 0.89 AUC) was built predicting either PJF or clinically significant PJK. This model can set the groundwork for preop point of care decision making, risk stratification, and need for prophylactic strategies for patients undergoing ASD surgery.

AB - Study Design. A retrospective review of large, multicenter adult spinal deformity (ASD) database. Objective. The aim of this study was to build a model based on baseline demographic, radiographic, and surgical factors that can predict clinically significant proximal junctional kyphosis (PJK) and proximal junctional failure (PJF). Summary of Background Data. PJF and PJK are significant complications and it remains unclear what are the specific drivers behind the development of either. There exists no predictive model that could potentially aid in the clinical decision making for adult patients undergoing deformity correction. Methods. Inclusion criteria: age -18 years, ASD, at least four levels fused. Variables included in the model were demographics, primary/revision, use of three-column osteotomy, upper-most instrumented vertebra (UIV)/lower-most instrumented vertebra (LIV) levels and UIV implant type (screw, hooks), number of levels fused, and baseline sagittal radiographs [pelvic tilt (PT), pelvic incidence and lumbar lordosis (PI-LL), thoracic kyphosis (TK), and sagittal vertical axis (SVA)]. PJK was defined as an increase from baseline of proximal junctional angle -208 with concomitant deterioration of at least one SRS-Schwab sagittal modifier grade from 6 weeks postop. PJF was defined as requiring revision for PJK. An ensemble of decision trees were constructed using the C5.0 algorithm with five different bootstrapped models, and internally validated via a 70 : 30 data split for training and testing. Accuracy and the area under a receiver operator characteristic curve (AUC) were calculated. Results. Five hundred ten patients were included, with 357 for model training and 153 as testing targets (PJF: 37, PJK: 102). The overall model accuracy was 86.3% with an AUC of 0.89 indicating a good model fit. The seven strongest (importance ≥0.95) predictors were age, LIV, pre-operative SVA, UIV implant type, UIV, pre-operative PT, and pre-operative PI-LL. Conclusion. A successful model (86% accuracy, 0.89 AUC) was built predicting either PJF or clinically significant PJK. This model can set the groundwork for preop point of care decision making, risk stratification, and need for prophylactic strategies for patients undergoing ASD surgery.

KW - Adult spinal deformity

KW - Predictive modeling

KW - Proximal junctional failure

KW - Proximal junctional kyphosis

KW - Sagittal malalignment

KW - Scoliosis

UR - http://www.scopus.com/inward/record.url?scp=84962406457&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84962406457&partnerID=8YFLogxK

U2 - 10.1097/BRS.0000000000001598

DO - 10.1097/BRS.0000000000001598

M3 - Article

VL - 41

SP - E1328-E1335

JO - Spine

JF - Spine

SN - 0362-2436

IS - 22

ER -