Automatic Detection, Classification, and Grading of Lumbar Intervertebral Disc Degeneration Using an Artificial Neural Network Model

Abstract

Background and objectives: Intervertebral disc degeneration (IDD) is a common cause of symptomatic axial low back pain. Magnetic resonance imaging (MRI) is currently the standard for the investigation and diagnosis of IDD. Deep learning artificial intelligence models represent a potential tool for rapidly and automatically detecting and visualizing IDD. This study investigated the use of deep convolutional neural networks (CNNs) for the detection, classification, and grading of IDD.

Methods: Sagittal images of 1000 IDD T2-weighted MRI images from 515 adult patients with symptomatic low back pain were separated into 800 MRI images using annotation techniques to create a training dataset (80%) and 200 MRI images to create a test dataset (20%). The training dataset was cleaned, labeled, and annotated by a radiologist. All lumbar discs were classified for disc degeneration based on the Pfirrmann grading system. The deep learning CNN model was used for training in detecting and grading IDD. The results of the training with the CNN model were verified by testing the grading of the dataset using an automatic model.

Results: The training dataset of the sagittal intervertebral disc lumbar MRI images found 220 IDDs of grade I, 530 of grade II, 170 of grade III, 160 of grade IV, and 20 of grade V. The deep CNN model was able to detect and classify lumbar IDD with an accuracy of more than 95%.

Conclusion: The deep CNN model can reliably automatically grade routine T2-weighted MRIs using the Pfirrmann grading system, providing a quick and efficient method for lumbar IDD classification.

Keywords: MRI; automation model; computer neural network; deep learning; diagnostic performance; intervertebral disc degenerations; lumbar spine.

Figure 1

The Pfirrmann grading system: grade I (a), grade II (b), grade III (c), grade IV (d), and grade V (e).

Figure 2

Dataset preparation: sagittal T2-weighted MRI of lumbar spine (a), annotation by a radiologist (b), image generation augmented by horizontal flip, cropping (zoom 0–20%), rotation (between −15° and 15°), shear (±15° horizontal and ±15° vertical) (c).

Figure 3

Deep learning architecture model used convolutional neural network (CNN).

Figure 4

YOLOv5 architectural size of the neural network model: Small-YOLOv5s (a), Medium-YOLOv5m (b), Large-YOLOv5l (c), and Extra-large-YOLOv5x (d).

Figure 5

Results of automatic detection and grading of lumbar degenerative discs using convolutional neural network (CNN) with YOLOv5. Prediction based on the Pfirrmann grading system: patients with grade I–III (a), patients with grade II (b), patients with grade IV (c), and patients with grade I–V (d). The picture demonstrates the grading and the mean average precision.

Figure 6

The IDD distribution of grades and shape information of bounding boxes of the dataset after using the augmentation technique.

Figure 7

Performance training analysis plot graphs of the CNN network with a large YOLO-V5 model.

Figure 8

Confusion matrix representation and accuracy of the training model. (Grade I accuracy 0.98, grade II 1.0, grade III 0.99, grade IV 0.99, grade V 1.0). A column represents an instance of the actual class, whereas a row represents an instance of the predicted class.

Figure 9

F1 score graph showing the relationship between F1 and the confidence curve.

Figure 10

Recall–precision graph showing the relationship between recall and precision.

Figure 11

Precision–confidence graph showing the relationship between precision and confidence.

Figure 12

F1 score graph showing the relationship between the F1 and confidence curves.