Summary: The Lonstein‐Carlson Progression Factor does not predict scoliosis curve progression in an independent replication study
Introduction: A risk of progression model for adolescent idiopathic scoliosis (AIS) was created in pivotal research by Lonstein and Carlson (L‐C) in 1984. This study suggested a formula and nomogram for risk of progression in AIS curves 20‐29o in those who were skeletally immature using Risser sign, chronologic age and Cobb angle as inputs. To our knowledge, the formula and nomogram have not been formally evaluated in a second cohort of AIS patients. Our objective was to test the performance of the L‐C model in an independent population.
Methods: Using the same methods as the original study, we compared the L‐C study of 268 patients with 315 similar AIS patients drawn from a wide distribution in North America. We calculated the progression factor versus the incidence of progression as defined in the original L‐C study.
Results: The demographic and clinical features of the L‐C cohort and the present study cohort were not statistically different. As shown in the plot below, there was little correlation between the progression score and the observed risk of progression (R=0.24).
Conclusion: This replication study shows that the L‐C Progression Factor model is not generalizable. Furthermore, any algorithm that relies so heavily on the Cobb angle is not truly predicting the risk of progression, rather it is an observation of how much progression has already occurred.
Significance: The L‐C study used rigorous and state‐of‐the‐art methods, but widespread use (and some misuse) of the data have occurred prior to any validation and replication studies. It is clear that biomarkers other than Risser sign, age and Cobb angle are necessary to provide more accurate parameters for calculating a risk of progression.