29. Understanding surgeon decision-making and current surgical practice regarding the DSIC scheme and DLS treatment: a CSORN study

Abstract

BACKGROUND CONTEXT Degenerative lumbar spondylolisthesis (DLS) is a spinal pathology involving the slip of one or more lumbar spinal vertebrae. It is most common in women and African-Americans above the age of 50, and usually involves the anterior slip of the fourth lumbar vertebra of up to 30% of its vertebral body width due to the more sagittally-oriented facet joints, and therefore, high force concentration at that level. Despite its worldwide prevalence of about 6%, the classification and treatment of DLS is not standardized. In 2015, Simmonds et al published the largest to-date systematic review of best available evidence and proposed a degenerative spondylolisthesis instability classification (DSIC) scheme. This classification scheme proposes that the stability of the patient's pathology be determined by a surgeon based on quantitative and qualitative clinical and radiographic parameters. PURPOSE This study aims to determine whether surgeons are following DSIC scheme treatment recommendations. It also aims to determine which radiographic and/ or clinical parameters and/ or surgeon characteristics are most influencing classification and treatment decisions. The goal of this study is to shed light on current surgeon treatment decisions and to better understand the factors most taken into account. In order for the treatment of DLS to become standardized, and therefore uniform and cost effective, we must first understand how current treatment decisions are made. STUDY DESIGN/SETTING This study used data from CSORN collected by surgeons across Canada between 2015 and 2018. Statistical analysis was used to determine which factors most influenced the classification and treatment decision of the surgeons. PATIENT SAMPLE A total of 224 patients. METHODS The patient data was collected by CSORN between 2015 and 2018 and included 224 DLS surgery cases. The data were cleaned and then analyzed using recursive feature elimination analysis, stability selection analysis, and principal component analysis to determine the correlations between the DSIC scheme parameters and the DSIC scheme grade, between the DSIC scheme parameters and the surgical treatment, and between selected parameters and the surgical treatment of DLS. Machine learning models (support vector machine and logistic regression) built using Python-based libraries were used to test the strength of the correlations found. RESULTS For the DSIC scheme classification and associated treatment, most cases of DLS were classified as being Type II (potentially unstable) (52%). However, most cases (61%) were treated with a 360° fusion regardless of their stability. In fact, only 18% of cases were classified as being Type III and thus demanding the TLIF/PLIF procedure. For Type I cases, the surgical treatment corresponding to Type I was employed 41% of the time. The use of appropriate surgical treatment was much lower for Type II (17%). For Type III, 91% of cases were treated with the recommended procedure. Dynamic translation was the parameter found to be the most correlated with the DSIC grade, with an r value of 0.475. Meyerding Grade and Disc Angle Standing were the two parameters most correlated with treatment decision-making, with respective first principal components of 0.374 and 0.332. CONCLUSIONS This study presented the demographics of DLS classification and treatment, which demonstrated that most surgeons are choosing the most extensive surgical procedure, 360° fusion, even when not indicated by the DSIC scheme. The study also found important correlations between patient and surgeon parameters and DLS classification and treatment, eg, dynamic translation influencing the classification and Meyerding Grade influencing the treatment. A larger data set would likely provide even better correlation results. Further efforts are needed to develop a new scoring algorithm that would serve to classify the stability of DLS based on the correlations found in this study. This new scoring algorithm would be extremely clinically significant as it would be more comprehensible and reproducible than the DSIC scheme and able to guide the treatment of DLS. FDA DEVICE/DRUG STATUS This abstract does not discuss or include any applicable devices or drugs.