Retrospective Review of Magnetic Resonance Imaging of the Lumbosacral Spine: Are We Overinvestigating?

Abstract

Objective
Lower back pain (LBP) is a worldwide health problem, and magnetic resonance imaging (MRI) is a common modality used to aid in its diagnosis. Although specific guidelines for assessing the necessity of MRI usage exist, the use of MRI as the initial imaging method for LBP seems to be more common than necessary in general practice.

Methods
We conducted a retrospective chart review of 313 patients who had undergone MRI of the lumbosacral spine during 2014–2015. We recorded and compared various factors, including age, sex, body mass index, current smoking status, race, symptoms, MRI findings, and progression to surgery within the next year. All rates were compared according to whether the MRI results showed radiographically significant findings (MRI-positive) or not (MRI-negative) using the chi-square or Fisher exact tests (if the expected cell count was <5). All analyses were performed using SAS version 9.4.

Results
There were no statistically significant differences in the rates of each symptom between the MRI-positive and MRI-negative groups, which accounted for 58.5% (183 of 313) and 41.5% (130 of 313) of the MRIs, respectively. The difference in the rate of surgery in the next year (18% among MRI-positive patients and 8.5% among MRI-negative patients) was found to be statistically significant (p<0.05).

Conclusion
Based on our findings, 41.5% of patients underwent lumbar MRI unnecessarily and 81% of patients with positive MRIs did not have surgery within the next year. Further physician training is needed to avoid unnecessary investigations and expenditures.

Full text

www.e-neurospine.org 383
Original Article
Corresponding Author
Suchit Khanduja
https://orcid.org/0000-0001-7257-8424
Department of Anesthesiology, Henry
Ford Hospital, 2799 W Grand Blvd,
Detroit, MI 48202, USA
Tel: +1-248-525-0801
Fax: +1-313-916-2600
E-mail: drsuchit@gmail.com
Received: April 17, 2018
Revised: July 2, 2018
Accepted: August 30, 2018
Retrospective Review of Magnetic
Resonance Imaging of the Lumbosacral
Spine: Are We Overinvestigating?
Suchit Khanduja, Vivek Loomba, Joseph Salama-Hannah, Aman Upadhyay,
Neha Khanduja, Gaurav Chauhan
Department of Anesthesiology, Henry Ford Hospital, Detroit, MI, USA
Objective: Lower back pain (LBP) is a worldwide health problem, and magnetic resonance
imaging (MRI) is a common modality used to aid in its diagnosis. Although specic guide-
lines for assessing the necessity of MRI usage exist, the use of MRI as the initial imaging
method for LBP seems to be more common than necessary in general practice.
Methods: We conducted a retrospective chart review of 313 patients who had undergone
MRI of the lumbosacral spine during 2014–2015. We recorded and compared various fac-
tors, including age, sex, body mass index, current smoking status, race, symptoms, MRI
ndings, and progression to surgery within the next year. All rates were compared accord-
ing to whether the MRI results showed radiographically signicant ndings (MRI-positive)
or not (MRI-negative) using the chi-square or Fisher exact tests (if the expected cell count
was < 5). All analyses were performed using SAS version 9. 4.
Results: ere were no statistically signicant dierences in the rates of each symptom be-
tween the MRI-positive and MRI-negative groups, which accounted for 58.5% (183 of 313)
and 41.5% (130 of 313) of the MRIs, respectively. e dierence in the rate of surgery in
the next year (18% among MRI-positive patients and 8.5% among MRI-negative patients)
was found to be statistically signicant (p<0.05) .
Conclusion: Based on our ndings, 41. 5% of patients underwent lumbar MRI unnecessar-
ily and 81% of patients with positive MRIs did not have surgery within the next year. Fur-
ther physician training is needed to avoid unnecessary investigations and expenditures.
Keywords: Lower back pain, Spinal surgery, Lumbar magnetic resonance imaging
INTRODUCTION
Lower back pain (LBP) is a worldwide health problem in
adults.1 This is the fifth most common reason for all doctor vis-
its in the United States.2 Despite technical progress in imaging,
the precise cause of the pain can only be determined in less
than 50% of cases.3 Magnetic resonance imaging (MRI) is a
common imaging technique used to aid in the diagnosis of
LBP. There are specific guidelines, such as failure of conserva-
tive treatment for 6 weeks and worsening of condition thereaf-
ter, abnormal electromyography or nerve conduction study
etc.,4 to assess the necessity for MRI usage. Further, the guide-
lines recommend the clinician not to routinely obtain imaging
in patients with nonspecific LBP5 since this costly diagnostic
modality has added a considerable burden on health systems of
many countries. Often times obtaining a lumbar spine MRI can
be of little value in making a definite diagnosis. This is because
a large number of spinal MRI findings can be present in asymp-
tomatic individuals.2 It is also true that a large number of symp-
tomatic individuals can have nonspecific findings on MRI. Yet
there is a widespread belief amongst physicians that an MRI is
needed to diagnose LBP. Associating LBP with lumbar MRI
findings can have many challenges. First of all, there is no accu-
rate definition of LBP. Secondly, symptoms can range from a
Neurospine 2018;15(4):383-387.
https://doi.org/10.14245/ns.1836110.055
N
eurospine
eISSN 2586-6591 pISSN 2586-6583
This is an Open Access article distributed under
the terms of the Creative Commons Attribution
Non-Commercial License (http://creativecom-
mons.org/licenses/by-nc/4.0/) which permits
unrestricted non-commercial use, distribution,
and reproduction in any medium, provided the
original work is properly cited.
Copyright © 2018 by the Korean Spinal
Neurosurgery Society

MRI of Lumbosacral SpineKhanduja S, et al.
https://doi.org/10.14245/ns.1836110.055
384 www.e-neurospine.org
purely dull aching back, solely leg pain or mixed with fluctua-
tions between both symptoms. The symptoms can also be con-
fused with isolated nerve compression. There are many MRI
findings that are primarily associated with LBP such as Modic
changes, vertebral compression fractures and degenerative disc
degeneration. Many findings are associated with radicular pain
such as spinal stenosis, disc extrusions and compressive neu-
ro pathy.3 The above stated observation makes the differentia-
tion of symptoms and their etiology complex since there may
be multiple symptoms in the patient and more than one posi-
tive MRI finding. To best of our knowledge there is hardly any
study to find if the recommended guidelines had any impact on
the usage of MRI in LBP. Therefore in this study we present a
retrospective review of records of patients who had undergone
MRI of lumbosacral spine for LBP.
MATERIALS AND METHODS
After obtaining Institutional Review Board approval from the
Hospital Ethics Committee we performed a retrospective chart
review of 313 patients who had undergone MRI of lumbosacral
spine for LBP in the Henry Ford Health system between 2014–
2015. The data was obtained from electronic health records,
specifically from Epic and the Care Plus Next Generation data-
bases. There was no exclusion criteria set except that all the pa-
tients were adults. We recorded and compared various factors
including chief complaint, age, sex, body mass index (BMI),
current smoking history, race, symptoms, MRI findings, and
progression to surgery within the next 1 year.
While radiologists in our center give thorough reports in re-
gards to MRI findings, it is up to the ordering physician to de-
termine the importance of the MRI findings in the given clini-
cal scenario. The findings that were considered positive were as
follows: severe degenerative changes, lumbar disc herniation,
spinal canal stenosis, spondylolisthesis, severe facet hypertro-
phy, nerve compression, spinal tumors, spinal Infection, fresh
vertebral fracture, and spinal deformity (kyphosis/scoliosis).
Mild to moderate degeneration, annular tears, modic changes,
and mild disc bulging were not considered to be positive as of-
ten times these findings can be present in asymptomatic indi-
viduals.2 All continuous variables were described using means,
standard deviations medians, minimums, and maximums, while
all categorical variables were described using counts and per-
centages. All rates were compared between groups (MRI posi-
tive and MRI negative) using chi-square or Fisher exact tests (if
expected cell counts are <5). Statistical significance was set at
Table 1. The descriptive statistics for all variables in the datas-
et of 313 patients
Vari ab le Val ue
Sex
Female 57.5
Male 42.5
Race
Black 65.6
White 33.1
Hispanic/Latino 1.3
Current smoking
No 80.9
Ye s 19.1
Body mass index (kg/m2)
Mean±SD 29.9 ± 7.9
Median (range) 28.1 (14.3–67.2)
Back pain
No 23.0
Ye s 77.0
Leg pain
No 98.7
Ye s 1.3
Back and leg pain
No 80.5
Ye s 19.5
Back injury
No 100
Ye s 0
Leg weakness
No 90.1
Yes 9.9
Bladder/bowel dysfunction
No 99.7
Ye s 0.3
Other symptoms
No 99.7
Ye s 0.3
MRI (binary)
Negative 41.5
Positive 58.5
MRI: normal
No 58.5
Ye s 41.5
(Continued to the next page)

MRI of Lumbosacral SpineKhanduja S, et al.
https://doi.org/10.14245/ns.1836110.055
www.e-neurospine.org 385
p < 0.05. All analyses were performed using SAS 9.4 (SAS Insti-
tute Inc., Cary, NC, USA).
RESULTS
The descriptive statistics for all variables in the analysis, pro-
portion of positive MRI findings by symptom, and departments
seeking MRI investigation of patients with LBP are shown in
Tables 1-3 respectively. Females constituted about 57.5% of the
population of the sample. The average age of patients was
63 ± 15.2 years Isolated LBP was the most common symptom
and reason of getting the MRI done (77%) followed by both
lower back and leg pain (19.5%). In percentage, 58.5% of the MRI
results showed radiographically significant findings whereas,
41.5% of MRIs were negative. The most common findings of
MRI positive patients were lumbar stenosis (29.5%), nerve com-
pression (26.8%), and lumbar disc herniation (24.3%). There
Vari ab le Val ue
MRI: lumbar disc herniation
No 75.7
Ye s 24.3
MRI: spinal stenosis
No 70.5
Ye s 29.5
MRI: Spondylolisthesis
No 99.4
Ye s 0.6
MRI: facet hypertrophy
No 97.1
Ye s 2.9
MRI: nerve compression
No 73.2
Ye s 26.8
MRI: spinal tumor
No 94.9
Ye s 5.1
MRI: spinal infection
No 98.1
Ye s 1.9
MRI: vertebral fracture
No 94.3
Ye s 5.7
MRI: spinal deformity
No 96.8
Ye s 3.2
Surgery
No 85.5
Ye s 14.5
Departments
Emergency Medicine 39.6
Internal Medicine 22.5
Neurosurgery 19.8
Orthopaedic Surgery 3.8
Hematology/Oncology 4.2
Urology 1.0
Neurology 4.1
Family Medicine 4.1
Vascular Surgery 0.3
Pain Clinics 0.3
Critical Care Surgery 0.3
Values are presented as percentage unless otherwise indicated.
SD, standard deviation; MRI, magnetic resonance imaging.
Table 1. Continued
Table 2. Percentage of negative and positive clinical findings by
symptom in 130 MRI negative and 183 MRI positive patients
Vari ab le MRI nega-
tive (n=130)
MRI positive
(n = 183) p-value
Back pain 0.108
No 18.5 26.2
Ye s 81.5 73.8
Leg pain 0.644
No 99.2 98.4
Ye s 0.8 1.6
Back/leg pain 0.209
No 83.9 78.1
Ye s 16.2 21.9
Back injury
No 100 100
Ye s 0 0
Leg weakness 0.962
No 90.0 90.2
Ye s 10.0 9.8
Bladder/bowel dysfunction 1.000
No 100 99.5
Ye s 00.5
Other symptoms 0.415
No 99.2 100
Ye s 0.8 0
Values are presented as percentage.
MRI, magnetic resonance imaging.

MRI of Lumbosacral SpineKhanduja S, et al.
https://doi.org/10.14245/ns.1836110.055
386 www.e-neurospine.org
were no statistically significant differences in the rates of each
symptom between the 2 groups (MRI positive and MRI nega-
tive). Eighty-one percent of the population did not have a cur-
rent smoking history. The average BMI of the patients who ob-
tained an MRI was 29.9 ±7.9, throwing out the notion that obese
individuals constitute the majority of the population that suffers
with LBP. Fig. 1 depicts the difference in rates of surgery in next
1 year among 18.0% positive MRI patients (33 of 183) and 8.5%
(11 of 130) among MRI-negative patients. The difference was
found to be statistically significant at p< 0.05.
DISCUSSION
The aim of our study was to deduce when an MRI of the lum-
bosacral spine is deemed necessary among various physicians.
We collected data on chief complaints, imaging findings, vari-
ous demographic data and those whom ultimately obtained
surgical intervention as well. The study revealed that 41.5% of
the MRIs of lumbosacral spine performed in our center were
negative. Among the MRIs that were positive, 81% did not have
any surgical intervention within the next 1 year. Isolated back
pain (77%), followed by back/leg pain (19.5%) were the most
common reasons for obtaining an MRI. Based on percentages,
lumbar spinal stenosis was the most common positive MRI
finding (30.7%), followed by disc herniation (24.3%). We also
found that of the multitude of departments included in our
health system, the Emergency Department (ED) had the high-
est rate of ordering MRIs with negative findings. There in fact
Fig. 1. Percentage of patents undergone surgery (light bar)
and having no surgery (dark bar) in magnetic resonance im-
aging (MRI) positive and MRI negative cases. *p< 0.05.
MRI negative MRI positive
100
90
80
70
60
50
40
30
20
10
0
(%)
Table 3. Percentage of MRI negative (130) and MRI positive
(183) referred by different departments
Department MRI negative MRI positive
Emergency Medicine 45.0 35.4
Internal Medicine 20.9 23.8
Neurosurgery 16.3 22.6
Orthopedic Surgery 3.1 4.3
Hematology/Oncology 3.9 4.3
Urology 0.6 1.2
Neurology 3.9 4.3
Family Medicine 3.9 4.3
Vascular Surgery 0.8 0
Pain Clinics 0.8 0
Critical Care Surgery 0.8 0
Values are presented as percentage.
MRI, magnetic resonance imaging.
have been guidelines laid out by the American College of Radi-
ology (ACR) for ordering imaging in lower backache which
may not be rigorously followed.5 In our study, we tried to mini-
mize the false positive rate of MRI results by excluding findings
such as mild to moderate degeneration, mild lumbar stenosis
which are incidental findings in a large portion of patients. Our
emphasis was on more critical findings such as severe degener-
ation, severe stenosis, or nerve compression which can warrant
medical or surgical intervention. In 2007, the American College
of Physicians and the American Pain Society published guide-
lines related to diagnostic imaging for backache.6 One of the
guidelines states that diagnostic imaging should be performed
only when a severe neurological deficit is noted or when on
history and physical examination, an underlying disease is sus-
pected. Also, physicians should only order an MRI if they feel a
patient is a candidate for neuraxial injection and or surgical in-
tervention.
While there is no radiation exposure associated with MRI
imaging, it can be of potential harm in that they can lead to spine
surgeries with no better outcomes.7 Carragee et al.8 performed
MRIs at baseline (no symptoms of LBP) and then a repeat MRI
if a patient developed an episode of LBP. The sample had 200
patients who were followed for 5 years. Eighty-four percent of the
patients who had recurrence of pain over next 5 years did not
have any changes in MRI findings.
There are several reasons postulated for the overutilization of
MRIs in the management of low back ache. First, despite the
presence of ACR guidelines for back pain management, there
are very few practitioners whom are aware of these guidelines.
Hence many physicians on the front line, such as ED physicians,
resort to an MRI as the first line in diagnostic imaging. Since
the MRI does not involve exposure to radiation, it is considered
safer than other diagnostic modalities by some healthcare pro-
viders. Variability in resident training and a lack of emphasis on
*

MRI of Lumbosacral SpineKhanduja S, et al.
https://doi.org/10.14245/ns.1836110.055
www.e-neurospine.org 387
complete history and physical examination may also lead to
over use of MRIs. There is also the continued stress of medical
liability on physicians, which may influence a physician’s deci-
sion making process when encountered with the new onset of
low back pain. This could be postulated as to the reason our
study showed a higher propensity for ED physicians in our sys-
tem to order MRIs.
All of the above stated studies and observations highlight the
problems associated with overuse of MRIs for low back pain.
We need to reinforce the importance of more conservative ways
of managing low back pain, such as exercise, yoga and physical
therapy. Physicians, and especially trainees, must be sensitized
to the importance of a detailed history and physical examina-
tion as well as be exposed to the ACR guidelines early on. Pa-
tient education plays an important role in this scenario as well.
Patients need to be educated of the benefits of introducing phys-
ical therapy as an initial means of treatment for pain. Many times
preconceived notions and fear avoidance beliefs regarding phys-
ical therapy deter patients from participating. Many times a pre-
occupation with incidental findings on an MRI may impair
healing patients and hence frequent reassurance and education
is needed. A perfect example is that of Virginia mason health
system in Seattle, WA.9 In 2004 Virginia mason was given no-
tice by insurance companies that their specialty practices were
costing 2 times the local practices in regards to imaging. Studies
in the hospital revealed that the physicians were not practicing
evidence based medicine. The physicians in the hospital system
had gotten into the habit of ordering MRIs unnecessarily. The
hospital changed numerous policies and brought physical ther-
apy to the forefront in managing back pain. Consequently, the
percentage of patients with LBP who got an MRI decreased from
15 to 10. The cost of care reduced substantially and decreased
the need for extra physicians in this systems pain clinic. It also
resulted in only 6 of patients losing time from work.
Our current study was a retrospective study of a segment of
population visiting Henry Ford Health system and is not indic-
ative of the overall population of the area. Also, the practices
are more indicative of our health system and the analysis and
demographics could be different in neighboring hospitals. Our
study does not differentiate between early and late MRIs i.e.,
MRI that was performed before conservative management and
the ones performed after conservative management had failed.
CONCLUSION
In short, there is still a possible overuse of MRI in our health
system. Both physician and patient education could help de-
crease the rate of MRIs performed for LBP and hence possibly
decrease a part of the financial burden on our health care system.
CONFLICT OF INTEREST
The authors have nothing to disclose.
REFERENCES
1. Deyo RA, Weinstein JN. Low back pain. N Engl J Med 2001;
344:363-70.
2. Chou R, Qaseem A, Owens DK, et al. Diagnostic imaging
for low back pain: advice for high-value health care from the
American College of Physicians. Ann Intern Med 2011;154:
181-9.
3. Suri P, Boyko EJ, Goldberg J, et al. Longitudinal associations
between incident lumbar spine MRI findings and chronic
low back pain or radicular symptoms: retrospective analysis
of data from the longitudinal assessment of imaging and
disability of the back (LAIDBACK). BMC Musculoskelet
Disord 2014;15:152.
4. MRI lumbar spine - MRI and back pain. In: NIA clinical
guidelines for medical necessity review. Scottsdale (AZ): Ma-
gellan Health Inc.; 2016. p. 92-6.
5. Davis PC, Wippold FJ 2nd, Brunberg JA, et al. ACR Appro-
priateness Criteria on low back pain. J Am Coll Radiol 2009;
6:401-7.
6. Chou R, Qaseem A, Snow V, et al. Diagnosis and treatment
of low back pain: a joint clinical practice guideline from the
American College of Physicians and the American Pain So-
ciety. Ann Intern Med 2007;147:478-91.
7. Yu L, Wang X, Lin X, et al. The use of lumbar spine magnet-
ic resonance imaging in Eastern China: appropriateness and
related factors. PLoS One 2016;11:e0146369.
8. Carragee E, Alamin T, Cheng I, et al. Are first-time episodes
of serious LBP associated with new MRI findings? Spine J
2006;6:624-35.
9. Flynn TW, Smith B, Chou R. Appropriate use of diagnostic
imaging in low back pain: a reminder that unnecessary im-
aging may do as much harm as good. J Orthop Sports Phys
Ther 2011;41:838-46.