Neural Network Based Spinal Age Estimation Using Lumbar Spine Magnetic Resonance Images (MRI) Atif Khan¹*, Daciana Iliescu¹, Evor Hines¹, Charles Hutchinson², Robert Sneath³ ¹School of Engineering, University of Warwick, Coventry, United Kingdom ²Warwick Medical School, University of Warwick, Coventry, United Kingdom ³University Hospital Coventry and Warwickshire, NHS Trust, Coventry, United Kingdom {Atif.Khan, D.D.Iliescu, E.L.Hines, C.E.Hutchinson}@warwick.ac.uk, robert.sneath@uhcw.nhs.uk Abstract— A human spine is a complicated structure of bones, joints, ligaments and muscles which all undergo a process of change with the age. This paper describes the existence of a pattern in degenerative process of human spine. Unveiling this pattern will be helpful in reassuring patients that the results of their scan are not unusual or indicative of any disease. A model based on artificial neural networks was formed with the help different spinal features such as vertebral height, disc height, disc signal and para-spinal muscles etc. These features of the lumbar spine vary with the age. Proposed model puts the degenerative changes of the lumbar spine into the context of a normal ageing process to estimate patient “spinal age”. This research work will provide a more concrete view of spinal growth in human being with the help of statistical features. This work will be helpful in drawing a borderline between normal, under and over growth of the human spine with respect to the person’s age. Keywords-: lumbar spine; degenerative changes; neural network; spinal age; magnetic resonance imaging. I. INTRODUCTION A human spine is a complicated and key component of human being. During the normal ageing process, spine undergoes progressive and regressive changes which presumably follow some pattern. There is much work to be done to uncover the pattern of human spine growth with reference to the normal ageing process. This research focuses explicitly on the study of progressive and degenerative changes occurring in human spine with the normal ageing process. The notable research carried out in this regards was to examine the pattern and prevalence of lumbar spine MRI changes within a southern Chinese population and their relationship to back pain with the help of 1043 MRI data samples [1]. Proposed research work concentrates on the identification and classification of age- related variations in "human spine" with the help of Magnetic Resonance Image (MRI) Scans of the lumbar spine area, belonging to different age groups. Back pain is usually associated with the spine disorder. You can sprain ligaments, strain muscles, rupture disks, and irritate joints, all of which can lead to back pain. While sports injuries or accidents can cause back pain, sometimes the simplest of movements, for example: picking up a pencil from the floor; can have painful results. In addition to that, arthritis, poor posture, obesity and psychological stress can cause or complicate back pain. Back pain can also directly result from disease of the internal organs, such as kidney stones, kidney infections, blood clots or bone loss [2]. Back pain is the second most common reason for visits to the doctor’s clinic, outnumbered only by the upper-respiratory infections. [3,4,5] Back pain is one of the most common reasons for missed work. One-half of all working Americans admit to having back pain symptoms each year [3,6]. The evidence from Britain and elsewhere is that back pain is becoming a bigger problem [7,8,9]. The total number of days lost in Britain for back incapacity obtained through sickness and invalidity benefit has risen dramatically in recent years [10,11]. The British experience is by no means unusual, with other developed countries seeing an even steeper rise to higher levels [12,13,14]. The data used in this research was collected from University Hospital Coventry and Warwickshire, United Kingdom. Data was in the form of magnetic resonance images (MRI) of the lumbar spine area. These MRI scans were of patients from different age groups. Scoring criteria was set under the expertise of an orthopedic surgeon and radiologist. Since the data was in raw form, some data pre- processing was done. The final model was designed and built using neural networks. Overall goal of this research was to uncover the growth/degeneration pattern of human spine with the normal ageing process and pinpointing the causative factors of the back pain. Before going deep into the analysis of age related spinal variations, one need to prove that there do exists some pattern in spinal ageing. This was achieved using artificial neural network. A model was built which estimates the patient age on the basis of different physical features of the lumbar spine. This will help spine specialists to see the deviation shown by a patient’s spine from where it is supposed to be. This research work tends to provide a better overview to the spine specialists as well as to the patients about the behavior shown by their spine.