Clinical and radiomics feature-based outcome analysis in lumbar disc herniation surgery

Lumbar disc herniation (LDH) is a condition where the intervertebral disc is displaced, leading to compression of the spinal nerve root and causing pain, numbness, and muscle weakness. This study investigates if radiomics features from MRI images, combined with clinical features, can improve the prediction of outcomes after LDH surgery. The study uses machine learning and deep learning to analyze radiomics and clinical features, aiming to enhance the accuracy of predicting patient outcomes.

• 1
  The inclusion of radiomics features alongside clinical variables led to a slight enhancement in predictive accuracy for surgical outcomes.
• 2
  Age and preoperative CRP levels were identified as the most significant clinical predictors, while GLCM, first-order statistics, and NGTDM were the most important radiomics features.
• 3
  The study demonstrated the potential of combining diverse data types in clinical predictive models, although the benefits should be considered with caution.