Poster Presentation Sydney Spinal Symposium 2019

Decay Variance - A new MRI technique for Diagnosing Disc Degeneration (#32)

Kyle A Sheldrick 1 , Uphar Chamoli 1 , Koichi Masuda 2 , Kenji Kato 2 , Shingo Miyazaki 2 , Ashish D Diwan 1
  1. Spine Service, St George and Sutherland Clinical School, University of New South Wales, Kogarah, NSW, Australia
  2. University of California San Diego, San Diego, California, USA


In most cases of low back pain no specific cause can be identified. Degeneration of the intervertebral disc is one putative cause, but there is little correlation between grading schemes of the appearance of the intervertebral disc on MRI and back pain. We attempted to develop a more accurate MRI technique for diagnosing disc degeneration based on known histochemical changes from the literature.


Using an animal puncture model in New Zealand White Rabbits we induced degeneration in 50 non-contiguous discs in 25 rabbits, while leaving the intervening disc as a control for a total of 75 discs. After 16 weeks, rabbits underwent a multi-echo-spin-echo (T2) MRI acquisition and were euthanised for histological grading of disc degeneration. The degree of histological degeneration was compared to the MRI findings in three ways: the average T2 signal intensity in each disc from a single echo (T2-Weighted-Imaging) as is used in clinical practice, the calculated T2 time by a non-linear-least-squares curve fitting operation (Quantitative T2) as is used in research, and our novel "Decay Variance" technique. 


A well defined nucleus pulposus with a consistent rate of decay (low decay variance) was visualised on Decay Variance Maps in healthy discs. This was progressively lost with increasing degrees of histological degeneration. Histology scores were more strongly correlated with Decay variance (r=0.82, P <.01) than with T2 signal intensity (r=0.32, P <.01) or quantitative T2  times(r=0.39, P <.01). Decay Variance had superior sensitivity and specificity, and overall pairwise accuracy (by area under ROC curve) for the diagnosis of degeneration than T2 Signal Intensity or Quantitative T2 mapping in our animal model.


Our technique had superior performance for the diagnosis of disc degeneration in a rabbit model. We cannot yet draw conclusions about correlation with symptoms in humans. Human studies are planned for 2019.