Low back pain is a leading cause of disability worldwide and an economic burden on health resources. Modic
changes seen on magnetic resonance images are common amongst patients with chronic low back pain.
Radiofrequency ablation of the basivertebral nerve and antibiotics for low-grade disc infection are potential
treatments that both rely on the presence of Modic type 1 changes. Accurate and standardised reporting of
Modic changes to select patients most likely to benefit and to ensure good antimicrobial stewardship is needed
but is challenging. Examples are given of varying, incomplete and/or inaccurate Modic change findings. Including
the extent of Modic change and likely pathophysiological interpretation would be helpful parameters to report.
Keywords: Modic Changes; Spine Infection; MRI
Abbreviations: CLBP: Chronic Low Back Pain; BVNA: Basivertebral Nerve; MRI: Magnetic Resonance Images;
MCs: Modic changes
The World Health Organization estimates that 619 million people
are living with low back pain and around 80% of people experience
back pain at some point during their lifetime. It is a leading cause
of disability worldwide and an economic burden on health resources
that is rising with increasing life expectancy [1,2]. In the United
States, overall healthcare costs associated with back pain have risen
to more than 134.5 billion dollars a year [3-5]. In over 80% of cases,
the cause of pain is unexplained and treatment is frequently inadequate
or ineffective [6,7]. Symptoms may be influenced by psychosocial
issues such as depression, catastrophizing and beliefs about pain
[8,9]. After six months of conservative treatment, therapeutic options
for chronic low back pain (CLBP) include intradiscal electrothermal
therapy, intradiscal steroid injection, intradiscal biacuplasty or various
forms of surgical fusion, all of which have variable results with no
evidence of long-term benefit [10].
Radiofrequency ablation of the basivertebral nerve (BVNA) and
antibiotics for low-grade disc infection are two potential treatments
for CLBP that are linked to inflammation seen as active endplate
changes on magnetic resonance images (MRI). Clinical trials have
demonstrated improved disability scores following BVNA [11-15]
and following treatment with oral antibiotics [16,17]. Intradiscal injection
of local antibiotic, as opposed to oral antibiotic therapy, is currently
being compared against a placebo in the Modic Trial (Clinical-
Trials.gov Identifier: NCT04238676). For both BVNA and treatment
with antibiotics, patients are selected on the basis of Modic changes
(MCs) seen on MRI scans [18,19] which are strongly associated with
low back pain [20-23]. More than 46% of patients with low back pain
have MCs compared to 6% of the general population [24-26]. Pain
relief following BVNA correlates with Modic type 1 (MCI) and type
2 (MCII) changes [12]. Benefit from antibiotics has been associated
with MCI but not MCII alone or Modic type 3 (MCIII) [16,17,27]. Thus,
correct identification and interpretation of MCs is of paramount importance
to select patients most likely to benefit from treatment with
either BVNA or antibiotics.
At present, in the UK, MRIs are not routinely performed (or encouraged)
as part of the assessment of CLBP [28]. Furthermore, appearance
of MCs varies according to the equipment and sequences
used. Fields, et al. [29] have published methodological guidelines to
facilitate comparable measurement of MCs. The ability to distinguish
between MCI and MCII is influenced by field strength and sequence
parameters; 1.5 Tesla is probably the better choice for detecting MCI
which is best visualized on fluid sensitive (STIR) sequences in conjunction
with T1 sequences [26]. However, both the presence and
type of MCs are variably reported [30-34]. Accurate and standardised
reporting of MCs is needed to optimise patient selection for either
BVNA or antibiotic therapy and to ensure good antimicrobial stewardship.
Here, we describe examples of variation, incomplete and/
or inaccurate Modic change findings on MRI that may represent challenges
to this aim.
Images from MRI scans and extracts from MRI reports have been
selected from patients who expressed an interest in participating in
the Modic Trial (ClinicalTrials.gov Identifier: NCT04238676) at Oxford
University Hospital NHS Foundation Trust (OUH), Oxford, UK to
illustrate the extent of variation in MRI reporting of MCs in patients
with CLBP. This study has been approved by the OUH institutional review
board as part of quality improvement in radiological reporting
(application number 9320).
Figure 1 shows classic examples of MCI, MCII and mixed MCI+MCII
found in this CLBP cohort. Figure 2 shows images from MRI scans
and the corresponding radiological reports that exemplify unreported
MCs, incomplete reporting of MCs and inaccurate reporting of MCs.
The image in Figure 2a shows clear MCI at L1/2 but the corresponding
report does not mention MCs. The report corresponding to the
image in Figure 2b states that ‘there is Modic endplate change at L5/
S1’. This patient has MCII at L5/S1 and would be unlikely to benefit
from treatment with antibiotics but may be suitable for BVNA. The
report corresponding to the image in Figure 2c states that the patient
has MCI when the patient has MCII in the inferior endplate of L5 and
a Schmorl’s node in the inferior endplate of L4. The signal changes
surrounding the Schmorl’s node may have been interpreted as MCI.
Figure 3 demonstrates variation in the interpretation of MCs depending
on the level of experience and specialist knowledge of the viewer.
Figure 4 illustrates two cases of MCI-associated CLBP that would be
unsuitable for intradiscal injection due to inaccessible discs. Table 1
gives typical examples of inconsistency and variability of terms used
in reporting MCs in patients with CLBP.
Figure 1
Figure 2
Figure 3
Figure 4
Table 1: Examples of inconsistency and inadequacy of terms used to
report MC in patients with CLBP.
Note: MC: Modic Changes; CLBP: Chronic Low Back Pain; MCI: Modic
Type 1 Changes; MCII: Modic Type 2 Changes
MCs are observed in about 6% of the general population but 46%
of patients with low back pain [25]. For those individuals with MCs,
pain is commonly continuous and it disrupts sleep [33,35]. The causal
role represented by MCs in CLBP, however, remains controversial due
to their multi-factorial aetiology. Their pathogenesis may be degenerative,
familial, infective, traumatic, or due to autoimmunity or instability
[30] and the absence of standardised radiological nomenclature
contributes to difficulties in understanding their significance [32]. In
order for MCs to usefully inform treatment choice, particularly for
emerging therapies, it is imperative that accurate and precise reporting
of Modic changes should form part of any standard report using
the appropriate Modic type nomenclature. In addition, disc level(s)
and endplate(s) affected should be clearly specified and distinguished
from other entities such as Schmorl’s nodes [36]. Quantifying the extent
of MCs is likely to be more useful than subjective descriptions
such as ‘early’, ‘mild’, ‘small’, ‘minor’, ‘moderate’ or ‘severe’. We suggest
that radiological reporting of MCs should include type, site, extent
and likely pathophysiological interpretation. Examples are provided
in Table 2.
MCI are active changes characterised anatomically by disruption
and fissuring of the endplate, the presence of extracellular water, micro-
fractures of the trabeculae and vascularized tissue in the adjacent
bone marrow [18,19]. They are seen as hypo-intense signal intensities
on T1-weighted spin-echo sequences and hyper-intense signal
intensities on T2-weighted sequences [37].
Table 2: Suggested terminology to use for reporting MCI, MCII and no
MC for patients with CLBP.
Note: MC: Modic Changes; CLBP: Chronic Low Back Pain; MCI: Modic
Type 1 Changes; MCII, Modic Type 2 Changes
Histologically, MCI represents inflammation and bone marrow
oedema thought to be due to low-grade disc infection [38]. Treatment
with antibiotics, therefore, is likely to be appropriate for patients with
MCI changes. MCII are inactive changes characterised by marrow
ischemia and fatty marrow replacement [18,19]. They are seen as hyper-
intense signal on both T1- and T2-weighted sequences [37] and
are more frequently found amongst individuals with degenerative
disc disease [39-41]. Disc degeneration is associated most commonly
with MCII but may also be associated with MCI. Mounting evidence
that vertebral endplates play a significant role in CLBP [42,43] has led
to a general consensus that BVNA is likely to provide clinical benefit to
patients with either MCI or MCII [44,45]. It is becoming an increasingly
established intervention for CLBP [14,46,47] despite the fact that
ablation destroys tissue and long-term follow-up is awaited [48]. MCI
has been associated with greater pain intensity than MCII [49-53],
with the pain directly correlating to the extent of MC [54]. There is,
however, inconsistency in reports of a potential association between
resolution of oedema (i.e. MCI) and symptom relief [17,35,55-57].
The most recent study suggests that any reduction in MCI may not be
not clinically relevant but the volume of MCI was not quantified [57].
Further studies are needed to inform CLBP treatment pathways associated
with mechanical degeneration and low- grade infection.
MCs are common MRI findings amongst those with CLBP but reporting
of MCs often remains variable and unstandardised. We suggest
that encouraging increased standardisation of Modic reporting is
essential for research into CLBP and selection of patients most likely
to benefit from emergent therapies. Inclusion of type, site and extent
of MCs, as well as the likely pathophysiological interpretation, are useful
parameters that could help and influence research opportunities.
The Modic trial was funded by Persica Pharmaceuticals Ltd. The
observations reported here were collated independently of Persica
but the authors would like to express their gratitude to Lloyd
Czaplewski, Sarah Guest and Duncan McHale for their critical appraisal
of the first draft of this manuscript.
ISSN: 2574 -1241
Abstract
PDF
