Lumbar Spine Degenerative Classification

In the year 2020, the World Health Organization reported that lower back pain constituted the primary cause of disability on a global scale, impacting 619 million individuals. Magnetic resonance imaging (MRI) offers a comprehensive perspective of the lumbar spine, encompassing vertebrae, discs, and nerves, thereby allowing radiologists to evaluate both the presence and severity of these conditions. This paper classifies five degenerative conditions regarding the lumbar spine, which are the following: Left Neural Foraminal Narrowing, Right Neural Foraminal Narrowing, Left Subarticular Stenosis, Right Subarticular Stenosis, and Spinal Canal Stenosis. Each study from the imaging dataset is represented by severity scores categorized as either Normal/Mild, Moderate, or Severe, and every one of these five conditions are classified over the five intervertebral disc levels: L1/L2, L2/L3, L3/L4, L4/L5, and L5/S1. We will use DenseNet, a convolutional neural network architecture known for its effectiveness in feature propagation and optimization of parameters, to overcome this classification challenge.