A Deep Learning Model for the Accurate and Reliable Classification of Disc Degeneration Based on MRI Data

Abstract

Objectives Although magnetic resonance imaging–based formalized grading schemes for intervertebral disc degeneration offer improved reproducibility compared with purely subjective ratings, their intrarater and interrater reliability are not nearly good enough to be able to detect small to medium effects in clinical longitudinal studies. The aim of this study thus was to develop a method that enables automatic and therefore reproducible and reliable evaluation of disc degeneration based on conventional clinical image data and Pfirrmann's grading scheme. Materials and Methods We propose a classifier based on a deep convolutional neural network that we trained on a large, manually evaluated data set of 1599 patients (7948 intervertebral discs). To improve upon the status quo, we focused on the quality of the training data and performed extensive hyperparameter optimization. We assessed the potential benefits of optimizing loss functions beyond common cross-entropy loss, such as soft kappa loss, ordinal cross-entropy loss, or regression losses. We furthermore experimented with ways to mitigate class imbalance by pooling classes or using class-weighted loss functions. During model development and hyperparameter optimization, we used a fixed 90%/10% training/validation set split. To estimate real-world prediction performance, we performed 10-fold cross-validation. Results The evaluated image data results in a Gaussian degeneration grade distribution, and thus grades 1 and 5 are slightly underrepresented in the training set. Our default cross-entropy–based classifier achieves a reliability of κ = 0.92 (Cohen κ), an average sensitivity of 90.2%, and an average precision of 92.5%. In 99.2% of validation cases, the network's prediction deviates at most 1 Pfirrmann grades from the ground truth. Framed as an ordinal regression problem, the mean absolute error between the ground truth and the prediction is 0.08 Pfirrmann grade with a correlation of r = 0.96. The results of the 10-fold cross validation confirm those performance estimates, indicating no substantial overfitting. More sophisticated loss functions, class-based loss weighting, or class pooling did not lead to improved classification performance overall. Conclusions With a reliability of κ > 0.9, our system clearly outperforms average human interrater as well as intrarater reliability. With an average sensitivity of more than 90%, our classifier also surpasses state-of-the-art machine learning solutions for automatically grading disc degeneration.

Immuno-Modulatory Effects of Intervertebral Disc Cells

“…Consistency in grading is essential for a clinician in order to have a clear idea of the patient’s condition, which has led to the development of deep learning models. Niemeyer et al (2021) have successfully managed to develop a deep learning model for the classification of discs based on MRI data that has an average sensitivity of 90%.…”

A comprehensive tool box for large animal studies of intervertebral disc degeneration

“…NP: nucleus pulposus, AF: annulus fibrosus. Human and sheep MRIs were kindly provided by Frank Niemeyer 169 and Marion Fussilier, 178 respectively storing at 4 C and sealing to prevent dehydration. 52 The loading history of the sample influences the reproducibility of the biomechanical evaluations, for example, due to variations in hydration.…”

“…NP: nucleus pulposus, AF: annulus fibrosus. Human and sheep MRIs were kindly provided by Frank Niemeyer 169 and Marion Fussilier, 178 respectively…”

“…In addition to Modic changes, MR imaging, especially advanced techniques, assists in improving the in‐depth knowledge of IVD degeneration while facilitating the 3Rs principles. Furthermore, considering that the degenerative changes scored on MRI are a continuum rather than the predefined five grades, the subjectivity of a five‐grade scoring system could be further improved with the aid of artificial intelligence and machine learning currently available for human applications 168,169 . Sensitivity of standard signal intensity‐based MRI is limited in early grades of degeneration and may lead to a gap in the recognition of early tissue alterations occurring in large animal models.…”

Analysis of the Coupling and Coordination Relationship between the Evolution of Enterprise Spatial Structure and Economic Development Based on the Deep Learning Model

“…According to the high-dimensional complex deep self-encoding network signal processing flow, it is divided into a data security input layer, a time series hidden layer, and a self-encoding network output layer. e perceptron layer that initially senses the incoming signal is called the input layer [5][6][7][8][9].…”