Can proximal junctional failure be predicted after adult spinal deformity surgery by experienced deformity surgeons based on patient, surgical and radiographic parameters?

Abstract

Objective: This study aimed to assess the ability of experienced spine surgeons to predict the occurrence of proximal junctional failure (PJF) within the first 2 years following ASD surgery, based on patient, surgical, and pre-discharge radiographic evaluation.

Methods: This case-based survey sent to spinal deformity surgeons included ASD patients who underwent correction-fusion surgery with a minimum follow-up of one year. Twenty patients who underwent posterior fusion and required PJF revision were identified. Another 20 patients with similar characteristics but without PJF served as a control group. Detailed patient vignettes included demographics, frailty assessment, medical history, DEXA, and Health-Related Quality of Life scores. Surgical details covered instrumentation, procedures, approach, graft use, osteotomies, cage incorporation, and PJF prophylaxis. Pre-operative and post-operative radiographs were provided. A panel of 21 experts in spine deformity surgery was then tasked to review these vignettes. The experts were instructed to offer qualitative predictions regarding the risk of failure, choosing from six possibilities ranging from "Extremely likely" to "Extremely unlikely." Metrics for prediction quality (sensitivity, specificity, precision, accuracy) were analyzed. Furthermore, they were asked to provide insights into the factors influencing their predictions.

Results: The median follow-up for these cases was 2 years. The median time to failure was 5.5 months (Range: 2 to 20 months). Of 782 predictions, only 136 were extreme (i.e., either "Extremely likely" or "Extremely unlikely" categories); most responses were uncertain, with 42.8% expressing predictions of being somewhat likely or unlikely to fail. Overall, prediction accuracy varied between experts (37% to 65%), precision ranged from 40% to 69%, and sensitivity ranged from 35% to 100%. Aggregating consensus instead of individual answers, the prediction accuracy was 61%, precision 67%, and sensitivity 50%. Only four experts demonstrated accuracy above the consensus results, and one exhibited better precision. Factors influencing predictions included post-operative sagittal alignment for success and various patient factors (age, BMI, frailty, bone quality) for PJF.

Conclusions: Experienced spinal deformity surgeons have a poor ability to predict PJK and PJF. Expert predictions varied, with consensus achieving 61% accuracy. Factors influencing predictions included post-operative sagittal alignment for success and patient-related factors for PJF. The findings highlight the complexity of predicting PJF, emphasizing the need for improved risk assessment tools in ASD surgery planning.

Keywords: Adult spinal deformity; Mechanical complications; Proximal junctional failure; Proximal junctional kyphosis; Survey.