axSpA International Outcome Assessment (AXIOMA)

In spite of its popularity as a research field, it is not known whether
biological drugs such as TNF-inhibitors and IL-17 inhibitors slow the
progression of structural damage in axial spondyloarthritis (axSpA) and can
thus be declared *disease modifying*. Lack of clarity here may either imply the
absence of a true pathophysiological effect or a metrological artifact (or in
theory a combination of both). The current gold standard for the assessment of
structural damage is conventional radiography (CR); however, it has several
limitations: only the anterior vertebral corners are assessed, the thoracic
spine cannot be reliably assessed due to overprojection of soft tissues and
bones, it only allows a 2 dimensional assessment, sensitivity to change is
limited, damage progression cannot be reliably captured at 1 year and therefore
can only be measured at 2 years, among others. Nevertheless, a better
assessment method has been lacking.
Recently, low-dose computed tomography (ldCT), which is the radiation-poor so
more acceptable alternative to conventional CT, has entered the field of axSpA.
Major advantages of CT are the lack of overprojection, giving the possibility
to assess the whole spine; the ability to scroll through the slices and the
3-dimensionality of the images allowing to assess syndesmophytes from different
planes, namely coronal and sagittal, and thereby assessing eight quadrants per
vertebral unit instead of two corners. Our team has released the
CT-Syndesmophyte-Score (CTSS) and has proved its validity.[1] ldCT, covering
the whole spine, detects 5 times more progression in the form of new and
growing syndesmophytes, within a 2-year interval than CR, which is limited to
the cervical and lumbar spine. Most progression was found to take place in the
thoracic spine, which can only reliably be assessed with ldCT. ldCT allows a
more detailed assessment of the growth of syndesmophytes, which in turn
resulted in a more pronounced difference in the detection of grown
syndesmophytes as compared to CR: ldCT detected 8 times more grown
syndesmophytes than CR. These findings allow the application of ldCT in
axSpA-trials and observational studies while limiting sample-size (in
provisional sample size calculations we estimated a possible reduction of
40%).[2] Currently, our team is running another project involving ldCT. As in
early axSpA CR is very limited in detecting structural damage,[3] we are
assessing whether ldCT can already in the early disease capture some measurable
damage. This study is ongoing as part of the follow-up of the Spondyloarthritis
Caught Early (SPACE) cohort.[4]
In comparison with CR ldCT has interesting additional research options. It
provides access to facet joints and may shed light on initial processes leading
to ankylosis. We have already shown that the assessment of facet joint
ankylosis is feasible, reliable and sensitive to change.[5] One of our
hypotheses awaiting further testing is that an ldCT score that combines
syndesmophytes and facet joint ankylosis may not only outperform the modified
Stoke in Ankylosing Spondylitis Spine Score (mSASSS, the most adequate scoring
method for CR) but also the CTSS (e.g. more discriminative), and thus better
reflects irreversible (bone) changes and their consequences in patients with
axSpA. A CT facet ankylosis syndesmophytes spine score (CTFASSS) is being
developed in our Sensitive Imaging of Axial Spondyloarthritis (SIAS) cohort in
which the CTSS and the facet joint ankylosis score had been developed.[1] The
new scoring method will combine both syndesmophytes (development and growth)
and facet joint ankylosis into one outcome. Both the CTSS and CTFASSS require
further validation before they (or one of them) can eventually be used for
instance in clinical trials and that is one of the aims of this project.
Another option for further research is to limit the field of view to only the
thoracic spine, thus further reducing the already acceptably low radiation
level. Based on previous data we hypothesize that such a focused ldCT-scan will
not jeopardize discrimination too much, while preserving many of the
methodological advantages of ldCT.
A last - but urgently required - additional option for research is to reduce
the minimum time-interval for detecting real progression of damage from two
years to one year. We hypothesize that ldCT-scanning of the spine and scoring
with an appropriate method will give resolution here. In summary, this project
aims at developing and defining the role of ldCT in the assessment of
structural damage in axSpA. The main outcome of this project is an
appropriately validated ldCT-scoring method for structural damage in patients
with axSpA. Such a method should be ready for application in clinical trials.
As ldCT is gaining popularity in the field of axSpA, it becomes important to
get insight into the course of structural damage progression over time. So far,
progression of structural damage as measured with ldCT has only been
investigated over a period of 2 years. Having identified a cohort of patients
with baseline structural damage and thus at risk for the development of more
damage, creates an opportunity to get insight into the progression of damage
over time in axSpA, as measured with ldCT, and both at the spinal and SIJ
level. Having a longer follow-up with both clinical outcomes and a complete
imaging assessment, including conventional radiographs, ldCT, as well as spinal
and SIJ MRI, allows to investigate the relationships between outcomes over
time, including shedding light on the longitudinal relationship between several
outcomes.
By adding a complete axial MRI to the assessments performed will allow to
investigate a current hot topic in field of axSpA, which are the structural
lesions as assessed on MRI and their validity and predictive validity. As CT is
the gold standard technique to assess bone damage, we can investigate whether
structural lesions seen on MRI, e.g. erosions, are also seen on ldCT, or
whether MRI-structural lesions can predict the development of CT-detected
structural lesions. If so, MRI-structural lesions could be used as surrogate
markers of structural damage. But if not, the field of axSpA also gets clarity
in that assessing the effect of interventions on MRI structural lesions is not
the way to go. The doubts around MRI structural lesions have led to not include
them in the most recent update of the ASAS Core Set for axSpA.[6] At the same
time, the need to measure structural damage in a more sensitive way calls for
more research in this area.

HYPOTHESES
i. An ldCT-score that includes syndesmophyte formation and facet joint
ankylosis will outperform a score that only includes syndesmophytes (CTSS).
ii. ldCT has superior *psychometric properties* to CR in the assessment of
progression of structural damage and may lead to 40% reduction of required
sample size in controlled trials.
iii. In contrast to CR, ldCT has the ability to capture structural progression
reliably already within 1 year.
iv. Limiting ldCT*s field of view to the thoracic spine only will reduce
radiation exposure while preserving superior *psychometric properties*.


2. OBJECTIVES
Primary Objective: To validate the CTSS and CTFASSS in an independent cohort
according to the OMERACT filter (truth, discrimination and feasibility);

Secondary Objectives:
i. To investigate whether ldCT has superior psychometric properties (e.g.
discrimination) to CR in the assessment of progression of structural damage;
ii. To investigate whether an ldCT-score including syndesmophyte formation and
facet joint analysis (CTFASSS, under development) performs better than a score
that only includes syndesmophytes (CTSS);
iii. To estimate a potential reduction in the sample size required in
controlled trials when using an ldCT score instead of CR to assess progression
of structural damage;
iv. To determine whether ldCT can capture structural progression reliably at 1
year;
v. To investigate whether an ldCT score limited to the thoracic spine has
superior psychometric properties compared to CR with reduced radiation exposure
compared to an ldCT score of the whole spine.
vi. Beyond the main research questions of the study, additional questions can
be addressed in AXIOMA. Examples are: comparing bone mineral densities using
ldCT (Hounsfield units) with progression of structural damage; whether a
syndesmophyte captured by ldCT is predictive of a syndesmophyte later visible
on CR; assessing whether progression of spinal damage in ldCT is associated
with progression of damage at the sacro-iliac joints also assessed with ldCT;
relationship with major outcomes over time, e.g. longitudinal association of
structural damage and spinal mobility or functional impairment, relationship
between disease activity and structural progression as measured by ldCT.
vii. New objectives related to the extension of AXIOMA:
a. To gain insight into spinal and SIJ structural damage progression over time,
as measured by ldCT, in patients with axSpA;
b. To analyse the predictive validity of MRI spinal and SIJ structural lesions
(erosions, fatty lesions, ankylosis) using the ldCT structural lesions as the
gold standard;
c. To investigate whether a change in inflammation, as measured on MRI, is
associated with a change in structural damage, as measured on ldCT, both in the
spine and SIJ;
d. To investigate the longitudinal relationship between spinal and SIJ
structural damage and spinal mobility, including individual spinal mobility
measures and also including an assessment of the relationship between the
structural damage in each spinal segment structural damage and spinal mobility;
e. To investigate the longitudinal relationship between spinal and SIJ
structural damage and functional impairment;
f. To investigate the longitudinal relationship between spinal and SIJ
structural damage and overall functioning and health, and also health-related
quality of life;
g. To investigate the longitudinal relationship between disease activity and
spinal and SIJ structural damage progression;
h. To investigate the role of physical activity as a proxy for mechanical
stress on structural damage progression;
i. To investigate the impact of different coping strategies on structural
damage progression, as well as on clinical outcomes such as functional
impairment, global functioning and HRQoL;
j. To investigate whether a change in MRI spinal and SIJ structural lesions is
associated with a change in clinical outcomes, such as functional impairment,
spinal mobility and overall functioning and HRQoL;

Study design: The Axial Spondyloarthritis International Outcome Assessment
(AXIOMA) Cohort is a prospective multicenter observational study.
Duration: 48 months of follow-up for each patient (for the whole study a
recruitment period of 12 months is expected). Patients will be invited to a
visit at baseline, 1 year, 2 years and 4 years. At 6 and 18 and 36 months they
will be invited to fill in a questionnaire online/on paper.
Setting: Patients will be recruited and observed in the rheumatology outpatient
clinic.

11.4. Benefits and risks assessment, group relatedness
There are no direct benefits for the patient upon participation in this study.
If participating, patients will be contributing to a progress in the knowledge
in axSpA and, particularly, in the measurement of structural damage. This
knowledge will likely be used in future trials testing the effect of treatment
strategies on the inhibition of structural damage. Patients will be informed by
the treating rheumatologist in case of unexpected findings important for the
patient*s health.
As the study is observational, there are limited risks in participating in the
study. Patients will not receive any treatment intervention. Burden and risks
associated with participation in the study are related to time spent in the
assessments, additional imaging assessments performed particularly ldCT, and
radiation associated with imaging. An effective dose of approximately 2.0 mSv
is received by a patient for a whole spine ldCT (including SIJ). The dose that
was initially theoretically estimated and with the use of a phantom at 4mSV, as
earlier published.[1,2] However, the measurement of the radiation dose to which
study participants were actually exposed revealed to be substantially lower and
to be of 2.0 mSV per ldCT. To compare this with CR, dose values could be
derived from the literature. There is a wide variety of reported doses, not
always representing current state-of-the art and reported doses may be too
high. In 2008, CR of the cervical spine (0.2 mSv), thoracic spine (1.0 mSv),
lumbar spine (1.5 mSv) and sacroiliac joints (1.5 mSv) combinedly delivered a
radiation dose of 3.4 mSv).[10] A more realistic estimate was made for our
SIAS-study radiography protocol, including patients from the Leiden University
Medical Center (LUMC) using published methodology.[11] This yielded a total
dose of 0.4 mSv resulting from cervical and lumbar spine and pelvic
radiographs. Background radiation in Europe is approximately 2.6 mSv per
year.[12] The European Commission for radiation protection has produced
guidelines for radiation exposure in medical research for patients under the
age of 50 years.[13] The dose delivered by ldCT falls in category 2B, which
allows research if it is *aimed directly at the diagnosis, cure or prevention
of disease*. Using ldCT, smaller and possibly more syndesmophytes can be seen
compared with CR; therefore, it is likely that earlier identification of
progression is possible. This would increase the feasibility of, for example,
medication trials for the prevention of progression. Therefore, ldCT fulfils
the requirement set by the European Commission. Compared with the background
radiation and the added value of ldCT over CR, we consider ldCT as a viable
method for the assessment of structural damage in the spine. Moreover, the
development in software with the capability of reducing the radiation dose
further is rapidly growing and lower dosages may be expected in the near
future. No other risks are foreseen by the participation in this observational
study.