Validity of Classification Systems for Guiding Surgical Intervention in Degenerative Lumbar Spondylolisthesis: A Retrospective Study
Rajib Ahmed, Ramy Sherif, Michael McCarthy, Francis BrooksBackground:
Degenerative spondylolisthesis of the lumbar spine (DSLS) is a common spinal pathology that often requires surgery. There is no standardized method for determining the most appropriate surgical approach. Recent attempts to develop classification systems, such as the degenerative spondylolisthesis instability classification (DSIC) and the University of California San Francisco Degenerative Spondylolisthesis (UCSF DS) scoring system, aim to bridge this gap by providing frameworks to predict surgical interventions based on clinical and radiographic data. This study evaluates the utility of these systems in guiding surgical decisions.
Methods:
Data from 82 DSLS cases that underwent surgical treatment were retrospectively reviewed using a comprehensive database. Two classification systems, UCSF DS and DSIC, were applied to each case to predict the appropriate surgical intervention. These predictions were then compared with the actual surgical procedures performed. Predictive accuracy for each system was measured using F1 scores and area under the receiver operating characteristic curve (AUC). McNemar test, Cohen kappa, and χ 2 analysis were used to compare system performance and estimation patterns.
Results:
The UCSF DS system outperformed the DSIC system, achieving an F1 score of 0.438 compared with 0.371, and an AUC of 0.62 compared with 0.58. Although overall prediction accuracy did not differ significantly between systems, χ
2
analysis revealed a highly significant difference in estimation patterns (χ²=39.99,
Conclusion:
The UCSF DS classification system demonstrates promise as a predictive tool for guiding surgical interventions in DSLS cases, with a higher match rate and a tendency toward less invasive recommendations compared with DSIC. However, this study assessed only agreement between predicted and actual surgical interventions; it did not evaluate clinical outcomes. Further prospective studies incorporating patient outcomes and complication data are needed to validate the clinical utility of these classification systems.