Future Data Points to Implement in Adult Spinal Deformity Assessment for Artificial Intelligence Modeling Prediction: The Importance of the Biological Dimension

Abstract

Adult spinal deformity (ASD) surgery is still associated with high surgical risks. Machine learning algorithms applied to multicenter databases have been created to predict outcomes and complications, optimize patient selection, and improve overall results. However, the multiple data points currently used to create these models allow for 70% of accuracy in prediction. We need to find new variables that can capture the spectrum of probability that is escaping from our control. These proposed variables are based on patients' biological dimensions, such as frailty, sarcopenia, muscle and bone (tissue) sampling, serological assessment of cellular senescence, and circulating biomarkers that can measure epigenetics, inflammaging, and -omics. Many of these variables are proven to be modifiable and could be improved with proper nutrition, toxin avoidance, endurance exercise, and even surgery. The purpose of this manuscript is to describe the different future data points that can be implemented in ASD assessment to improve modeling prediction, allow monitoring their response to prerehabilitation programs, and improve patient counseling.

Emerging Technologies within Spine Surgery

“…It uses multiple complex models in parallel to examine indirect relationships and generate accurate predictions. It will expand our considered measures into previously unexplored areas such as biomarkers and epigenetics [62]. However, concerns regarding the utilization of AI within spine surgery exist.…”

Advancing spine care through AI and machine learning: overview and applications

“…In the spine domain, omics and multimodal approaches have been overlooked. However, it has been shown that the variables currently used (EHR, images, questionnaires) allow for around 70% accuracy in prediction suggesting that more variables coming from the biological field need to be investigated (87,88). There have been some attempts to implement multimodal models but without including omics and using a few modalities only.…”

Personalizing adult spinal deformity surgery through multimodal artificial intelligence

“…These may include biomarkers of frailty and senescence, radiomics, and objective markers of phenotype. 15 Smartphones and wearable devices are also a growing source of functional health data, providing a natural and continuous method for ASD patient metrics to be recorded and characterized within a “digital phenotype.” 16 , 17 …”