The Axial Set:
Why Spine, Pelvis, and Hips Are One Study
A clinical guide to minimum and optimal imaging coverage by diagnostic question
The acetabulum is part of the pelvis. The pelvis transmits load between the lumbar spine and the femoral heads. The sacroiliac joints sit at the junction of both. This is first-year anatomy — and it disappears, in practice, the moment an imaging order is written.
In clinical practice, the spine is ordered by one clinician, the pelvis by another, and the hips possibly by a third. The reports go to separate folders. They may be read by different specialists on different days. The patient, whose skeleton observed none of these administrative divisions, continues to produce symptoms that no single report fully accounts for.
The axial skeleton is a functional unit. The system that evaluates it is organised by region. When those two things are misaligned, the cost accumulates quietly — in delayed diagnoses, repeated studies, and management decisions built on structurally incomplete information.
What follows is a practical framework for closing that gap: eight clinical scenarios mapped to the minimum radiographic coverage that generates any useful information, the optimal coverage that generates a complete and actionable answer, and the specific clinical consequence of the distance between the two.
One Skeleton, One Process
Ankylosing spondylitis begins, in most patients, at the sacroiliac joints and propagates upward: lumbar spine, thoracolumbar junction, thoracic and cervical segments. Radiographically significant hip involvement — structural joint space narrowing or erosive change, distinct from peripheral arthritis — occurs in 17–24% of patients with axial SpA in prospective cohort data, with higher estimates when clinical involvement is included.¹ It is not a complication of spinal disease. It is the same inflammatory process expressed at a different point in the same anatomical chain.
DISH distributes by the same logic. Its flowing ossification is not a lumbar phenomenon that incidentally extends elsewhere — it is a systemic enthesopathy whose thoracic predominance, level count, and relationship to the sacroiliac joints are all diagnostically relevant. Three visible levels of bridging in the lumbar spine may be three of three, or three of nine. That distinction changes the diagnosis. It is not recoverable from a lumbar study alone.
Degenerative disease follows the same connectivity, mechanically rather than pathologically. Consider a scenario that most clinicians have encountered: a patient with lumbar foraminal stenosis at L3–L4 presenting with groin and anterior thigh pain, referred for hip imaging because that is where it hurts. AP pelvis shows mild joint space narrowing — enough to explain the pain, not enough to require intervention. Management is directed at the hip. Two years later, L3–L4 decompression resolves the pain that two courses of intra-articular injections did not. The lumbar spine was never imaged because the pain was in the hip. The hip was never the problem.
This is not a failure of any individual clinical decision. It is the predictable output of imaging organised by location rather than by question — and the table that follows is a structured attempt to make the distinction between those two things explicit.
Image 1: Axial chain diagram — C-spine → T-spine → L-spine → SI joints → pelvis → hips, with pathological propagation and mechanical load transfer vectors
Disease and mechanical load distribute along the axial chain. Segmental imaging captures one link. The diagnosis frequently lives in the relationship between links.
Study design is a clinical decision. The choice of which regions to include — and which projections — determines what questions the radiograph can answer before the first image is acquired.
Minimum, Optimal, and What the Gap Costs
The table maps eight common clinical scenarios to the minimum radiographic set that generates any useful structural information, the optimal set that generates a complete and actionable answer, and — in the final column — the specific clinical consequence of stopping at the minimum. Minimum does not mean adequate. In most scenarios below, minimum coverage produces a technically correct report that cannot support a clinical decision. The optimal set is the study designed for the question being asked, not for the region that hurts.
| Clinical scenario | Minimum set | Optimal set | Clinical cost of the gap |
| Suspected axial SpA — early or undifferentiated | AP pelvis with Ferguson projection for SI joint optimisation | Whole spine AP + lateral (C, T, L) · AP pelvis with Ferguson view · lateral hips if symptomatic | AP pelvis with equivocal bilateral sclerosis confirms nothing and excludes nothing on its own. Corner changes and syndesmophyte distribution in the lumbar and thoracic spine convert an ambiguous SI finding into a directed differential — shifting pre-test probability sufficiently to justify MRI referral on radiographic grounds. Without spinal context, the study cannot stage axial disease or meet the imaging limb of the ASAS classification criteria. |
| DISH vs ankylosing spondylitis — unresolved differential | AP + lateral lumbar · AP + lateral thoracic | Whole spine AP + lateral · AP pelvis with Ferguson view · lateral cervical | Level count, right-sided thoracic pattern, and disc height across all bridged levels require full spinal coverage. SI joint status — normal in DISH by definition, involved in virtually all established AS — is the decisive variable and requires dedicated pelvic views. Lateral cervical adds the final element: in late AS, cervical syndesmophytes are thin and marginal; in DISH with cervical extension, ossification is thick and anterolateral. Without lateral cervical, the differential remains open at the most proximal segment of the chain. |
| Chronic hip pain — source unclear | AP pelvis | AP pelvis · bilateral lateral hips · lateral lumbar · standing AP pelvis if leg length or alignment concern | AP pelvis shows medial and superior joint space but systematically underrepresents anterosuperior joint space loss — the earliest radiographic sign in cam-type femoroacetabular impingement. SI joint disease and lumbar radiculopathy (L3–L4, L4–L5) produce buttock and groin pain that overlaps clinically with true hip joint disease and cannot be excluded without lateral lumbar. A normal AP pelvis in unresolved hip pain is not reassurance. It is an incomplete evaluation presented as one. |
| Inflammatory arthropathy with axial involvement (PsA, IBD-SpA, reactive arthritis) | AP pelvis | Whole spine AP + lateral · AP pelvis with Ferguson view · lateral hips bilateral | These conditions are multi-region by nature. Asymmetric SI involvement, thoracic enthesopathy, hip erosions, and cervical syndesmophytes may each be independently present and independently significant for disease staging, treatment decision, and clinical trial eligibility. Pelvis alone captures one dimension of a process that typically spans several. |
| Degenerative coxarthrosis — pre-operative planning | AP pelvis · lateral hip (symptomatic side) | AP pelvis · bilateral lateral hips · lateral lumbar · standing full-length lateral (spine-pelvis-hip); EOS imaging where available | Contralateral hip disease becomes symptomatic post-operatively as load mechanics shift. Unrecognized lumbar stenosis or spondylolisthesis predicts worse functional recovery after THA independent of hip findings. Pelvic incidence — the fixed angle between a perpendicular to the sacral endplate midpoint and the line to the femoral head axes — determines functional acetabular cup orientation; patients with stiff spines and abnormal spinopelvic morphology cannot achieve a safe zone with standard positioning, with implications for dislocation risk. Standing lateral (or EOS) is the only plain film assessment that captures this in the functional upright posture. |
| Metabolic bone disease (osteoporosis, hyperparathyroidism, Paget’s disease) | Lateral thoracic · lateral lumbar | Whole spine lateral · AP pelvis · lateral hips bilateral · standing AP if deformity present | Vertebral fracture burden across the full spine determines fracture risk category and treatment intensity. Paget’s disease and hyperparathyroidism produce proximal femoral and pelvic changes that are part of the metabolic skeletal picture and alter management. Isolated thoracolumbar imaging systematically underestimates total disease burden and may miss the dominant site of involvement. |
| Juvenile SpA / JIA with suspected axial involvement | AP pelvis | Whole spine AP + lateral · AP pelvis · lateral hips bilateral · oblique SI views if SI joints equivocal on AP | The SI joint of a skeletally immature patient shows physiological irregularity and apparent widening that is normal for age — and indistinguishable from early inflammatory change without age-indexed context. Hip joint space narrowing in a growing patient carries different prognostic weight and treatment implications than in an adult. Spinal involvement in juvenile SpA may be subtle and multisegmental. Isolated pelvis cannot support age-appropriate axial assessment and risks false reassurance in the population where early diagnosis most changes long-term outcome. |
| Back pain with radiation to buttock or hip — source unclear | AP + lateral lumbar | AP + lateral lumbar · AP pelvis with Ferguson view · lateral hip (symptomatic side) | Lumbar imaging alone cannot exclude SI joint disease, hip pathology, or L4–L5 / L5–S1 radiculopathy as the pain source — all of which produce buttock and posterior hip pain in distributions that overlap clinically. A technically normal lateral lumbar with unresolved radiation to the hip is not a conclusion. It is the point at which the pelvis and hip should have been in the same study from the outset. |
Table 1. Minimum and optimal radiographic coverage by clinical scenario. Ferguson projection: 30° cephalad tube tilt on AP pelvis, which opens the SI joint space by correcting for the sacral inclination angle and substantially improves SI joint visualisation compared to standard AP. Dedicated oblique SI views provide additional characterisation in equivocal cases. EOS: low-dose biplanar slot-scanning system providing simultaneous AP and lateral full-body imaging in the standing position; where available, preferred over conventional standing radiographs for spinopelvic parameter measurement due to substantially lower radiation dose.
Three Scenarios in Detail
Three scenarios from the table illustrate the widest gap between what minimum coverage shows and what the clinical situation requires. The others are no less important — they simply compress into the table more completely.
Early axial SpA: equivocal is not negative, and the spine is part of the evaluation.
AP pelvis is the correct first imaging step in suspected axial spondyloarthritis. The sacroiliac joints are the canonical early target, and the Ferguson projection — a 30° cephalad tube tilt that opens the SI joint space by correcting for sacral inclination — substantially improves SI joint visualisation compared to a standard AP pelvis and should be requested specifically rather than assumed. NICE NG65 recommends plain radiography of the SI joints as first-line imaging in adults with suspected axial SpA, with MRI reserved for cases where plain films do not demonstrate sacroiliitis but clinical suspicion persists.²
The problem begins when the AP pelvis shows what most pelvic studies show in adults over 40: mild to moderate bilateral SI sclerosis. This finding is present in a substantial proportion of adults without inflammatory disease, confirms nothing, and excludes nothing on its own. Without spinal context, the study has produced genuine uncertainty — and the clinical response to genuine uncertainty is typically one of three: pursue MRI without clear radiographic indication, repeat imaging in six months with the same result, or — most commonly — defer and document ‘degenerative changes, clinical correlation recommended.’
Add lateral lumbar and lateral thoracic projections, and the picture changes. Anterior corner erosions at the thoracolumbar junction, vertebral body squaring at two levels, or early syndesmophytes in a patient with bilateral SI sclerosis shift the pre-test probability of axial SpA substantially. Together, these findings engage the imaging limb of the ASAS classification criteria³ and provide the radiographic basis for MRI referral with specificity — not as an administrative step, but as a directed investigation of a probable diagnosis. The spinal findings do not replace SI joint evaluation. They give it clinical meaning.
[ Image 2: AP pelvis with equivocal SI sclerosis alongside lateral lumbar with Romanus lesions — the combination that converts uncertainty into a directed differential ]
Chronic hip pain: the anatomy of referred pain is not intuitive, and the AP pelvis does not image all of it.
Ask a patient where their hip hurts and they will point to the lateral thigh, the groin, the anterior thigh, or the buttock — sometimes all four. These are not the same pain. Lateral thigh pain is more often trochanteric or iliotibial than intra-articular. Groin pain is the most reliable localiser for true hip joint disease. Anterior thigh pain follows the L3 dermatome. Buttock pain — one of the most common presentations in axial disease clinics — is produced by SI joint pathology, lumbar facet disease, L4–L5 or L5–S1 radiculopathy, and true hip disease, in overlapping distributions that cannot be separated by location alone.
Standard AP pelvic imaging shows the medial and superior hip joint space, the SI joints in an oblique frontal plane, and the femoral head contours. It does not show the anterosuperior hip compartment. In cam-type femoroacetabular impingement — the most common morphological subtype — anterosuperior joint space loss is the earliest radiographic sign, visible on a correctly angled lateral view and systematically underrepresented on AP. A patient with cam morphology and a normal-appearing AP pelvis has not been cleared. They have been partially imaged.
The lumbar spine is not in the field. A pelvic study that shows no significant hip joint pathology in a patient with persistent groin or buttock symptoms should not be read as reassuring — it should prompt the lateral lumbar that was not included, because L4–L5 foraminal stenosis and SI joint disease remain unexcluded. The study that was ordered is complete. The workup is not.
[ Image 3: AP pelvis vs AP + bilateral lateral hips + lateral lumbar — diagnostic yield at each step in chronic hip pain workup ]
Pre-operative hip arthroplasty: the surgery is planned for the hip; the outcome depends on the spine.
Hip replacement surgery that addresses only the symptomatic hip is planning the procedure, not the patient. Three anatomical relationships outside the primary surgical field have documented effects on post-operative outcomes — and all three require imaging that is not standard in a hip-focused pre-operative workup.
The contralateral hip. Advanced bilateral hip osteoarthritis is common, and the contralateral side is frequently underreported at the time of index surgery. Post-operative gait mechanics shift load to the non-operated side. Contralateral disease that was minimally symptomatic pre-operatively becomes the dominant source of pain and functional limitation within months of recovery. Bilateral AP pelvis with lateral views of both hips before arthroplasty is the study that allows an informed surgical and rehabilitation conversation — not a precautionary expansion of scope.
The lumbar spine. Lumbar spinal stenosis, spondylolisthesis, and significant foraminal stenosis each independently predict worse functional outcomes after total hip arthroplasty.⁴ This is not a minor complication factor — it changes the risk-benefit discussion before surgery, not after. Lateral lumbar imaging is standard pre-operative due diligence.
Pelvic incidence and spinopelvic balance. Pelvic incidence is the angle between a perpendicular to the midpoint of the sacral endplate and the line connecting that midpoint to the centre of the femoral head axes — a fixed morphological parameter that determines pelvic tilt and lumbar lordosis in the upright posture. It governs functional acetabular cup orientation after total hip arthroplasty: a cup placed correctly relative to the intraoperative pelvic position may be mechanically incorrect when the patient stands, because standing pelvic orientation differs from the supine surgical position. Patients with stiff spines and abnormal spinopelvic morphology cannot achieve a functional safe zone using standard surgical positioning.⁵ Standing full-length lateral imaging — or EOS biplanar imaging where available, which provides equivalent spinopelvic data at substantially lower radiation dose — is the only plain film assessment that captures pelvic incidence and lumbar lordosis simultaneously in the functional upright posture. It should be standard pre-operative imaging for any patient with known lumbar disease or reduced spinal mobility.
The Projection Problem: What Each View Actually Shows
The diagnostic value of any projection depends on the clinical question, the anatomical region, and what is being sought. The following observations apply specifically to the axial scenarios in the table above — not as a universal hierarchy of projections, but as a description of the consistent mismatch between what gets ordered and what the question requires.
AP pelvis: shows the medial and superior hip joint space, the SI joints in an oblique frontal plane, and the femoral head contours. It does not show the anterosuperior hip compartment, the posterior SI joint surface, or any sagittal relationship. For cam-type FAI, SI joint erosion morphology, and spinopelvic assessment, AP pelvis alone does not answer the question — it confirms that imaging was performed.
Ferguson projection: a 30° cephalad tube tilt that opens the SI joint space by correcting for sacral inclination. It is not a separate study — it is an AP pelvis done correctly for SI joint evaluation. Without it, the standard AP pelvis systematically underrepresents SI joint margin definition and early erosive change. It should be requested explicitly whenever SI joint assessment is the clinical purpose.
Lateral hip: shows the anterosuperior joint space — the earliest and most diagnostically sensitive site of loss in cam-type FAI and in the flexion-loading pattern of axial hip disease. In the axial scenarios in this article, the lateral hip frequently carries more diagnostic information than the AP pelvis that generated the referral.
Lateral lumbar: shows disc height at each level, endplate morphology, listhesis grade, foraminal caliber, and bridging ossification character. AP lumbar shows disc space width and coronal alignment. In most scenarios in the table, it is the lateral view that distinguishes between diagnoses. The AP confirms that a finding is present. The lateral determines what it is.
Lateral thoracic: shows bridging ossification morphology and laterality — the features that separate DISH from ankylosing spondylitis in the thoracic spine. Right-sided predominance, ossification thickness, and disc height across thoracic bridged levels are lateral view findings. AP thoracic alone cannot close the DISH–AS differential in the thoracic segment.
Lateral cervical: shows anterior ossification morphology at cervical levels — thick and anterolateral in DISH with cervical extension, thin and marginal in late AS. In the context of the DISH–AS differential, lateral cervical is the projection that characterises the most proximal segment of the chain and completes the distributional picture that no other single view provides.
Standing versus supine: weight-bearing changes listhesis grade, disc height, and pelvic orientation. Dynamic instability — spondylolisthesis present standing but absent or minimal supine — is not visible on supine imaging. For surgical planning, deformity assessment, and spinopelvic parameter measurement, standing lateral is a different study from supine lateral. EOS biplanar imaging, where available, is the preferred modality for standing full-body spinopelvic assessment, providing simultaneous AP and lateral acquisition at radiation doses substantially lower than conventional standing radiographs.
For the scenarios in this article, the lateral projection consistently carries more diagnostic information than the frontal view that prompted the referral. That asymmetry between what is ordered and what answers the question is the core practical problem this table addresses.
Building the Study: Practical Sequencing
A complete axial set does not require a single appointment, referral, or order. It requires a decision about what clinical question is being asked before the regions are selected.
Start with the question, not the painful region. ‘Rule out axial SpA’ and ‘characterise lumbar pain in a 60-year-old with known spondylosis’ require different studies. The first requires SI joint evaluation — with Ferguson projection — as a primary component. The second may not require pelvis at all. Ordering by region rather than by question is the mechanism that produces most incomplete axial studies and most of the repeated imaging that follows them.
Add the pelvis when the question is axial. AP pelvis with Ferguson projection is a low-incremental-dose addition to any lumbar or thoracic study. Effective dose for AP pelvis is approximately 0.4 mSv — roughly one quarter of a standard two-projection lumbar series, and well within the range of routine clinical justification. In any patient where SpA, DISH, or mixed inflammatory-degenerative disease is on the differential, omitting the pelvis is omitting the region most likely to determine the next step.
Include lateral views as a default component, not a request. Lateral lumbar and lateral thoracic are not supplementary projections. An AP-only lumbar series describes the coronal appearance of a structure whose most diagnostically relevant features — disc height, listhesis, bridging morphology, foraminal caliber — are sagittal. For the scenarios in this article, ordering AP without lateral is ordering the less informative half of the study.
Use standing or EOS imaging when deformity, instability, or surgical planning is involved. Supine imaging underestimates deformity and misses dynamic instability. Standing AP and lateral — or EOS where available — are the appropriate default for patients being assessed for axial surgery, progressive scoliosis, or sagittal imbalance. The radiation increment over supine imaging is modest with conventional equipment and negligible with EOS.
Design the study to answer the question once. The most common source of cumulative radiation exposure in axial imaging is not the comprehensive initial study — it is the sequence of partial studies, each inadequate, each generating a report that prompts another referral. A well-designed first study is, in most clinical scenarios, also the lower-lifetime-dose option.
[ Image 4: Study design decision tree — clinical question → appropriate axial set → next clinical or imaging step ]
What This Table Is, and What It Isn’t
The table in this article is a reference for thinking about study design — not a protocol, not a substitute for clinical judgment, and not a comprehensive imaging guide. Every clinical decision involves variables that no table can weigh: the patient’s age and radiation history, prior imaging and what it showed, pre-test probability, and what decision will be made on the basis of this study. Those variables belong to the clinician ordering the study and the radiologist interpreting it.
What the table provides is the structural logic of coverage: for a given diagnostic question, what minimum information is required, what complete information looks like, and what is at risk when the gap between the two is not closed. That logic does not change with specialty or career stage.
A student who internalises it will develop the habit of asking — before writing the order — what question this study is designed to answer. That question, asked consistently, is the single most effective intervention in axial imaging quality.
An experienced clinician will recognise specific patients in the table: the ones whose imaging record is technically complete and diagnostically fragmented, where no single study ever contained the variables needed to confirm or exclude the diagnosis that changed — or should have changed — management.
A radiologist will find in the table a framework for the conversation that the referral should have contained: not what region to image, but what question to answer, with which projections, at what level of coverage. That conversation, when it happens before the study is acquired rather than after, is where the most diagnostic value is generated — and where the relationship between ordering clinician and interpreting radiologist works best.
The table is the same for all three readers. A study designed for the question produces an answer that supports a decision. A study designed for the region produces a description that supports another study.
References
1. Dougados M, et al. Prevalence of comorbidities in rheumatoid arthritis and evaluation of their monitoring: results of an international, cross-sectional study (COMORA). Ann Rheum Dis. 2014;73(1):62–68. Hip involvement data from DESIR cohort: Dougados M, et al. Ann Rheum Dis. 2011;70(3):466–468.
2. National Institute for Health and Care Excellence. Spondyloarthritis in over 16s: diagnosis and management. NICE guideline NG65. 2017 (updated 2024). https://www.nice.org.uk/guidance/ng65
3. Rudwaleit M, et al. The development of Assessment of SpondyloArthritis international Society (ASAS) classification criteria for axial spondyloarthritis. Ann Rheum Dis. 2009;68(6):777–783.
4. Bedard NA, et al. What are the factors associated with achieving minimal clinically important difference after primary total hip arthroplasty? J Arthroplasty. 2019;34(11):2701–2707. See also: Clement ND, et al. Total hip and knee arthroplasty in patients with concomitant lumbar spinal stenosis. J Bone Joint Surg Br. 2013;95(12):1628–1634.
5. Vigdorchik JM, et al. Inability to achieve a functional safe zone of the acetabular component in patients with a stiff spine. J Arthroplasty. 2019;34(5):1001–1008. Innmann MM, et al. Spinopelvic mobility and its influence on prosthetic joint mechanics and dislocation after THA. EFORT Open Rev. 2020;5(6):318–324.
RheumaView™ · research.rheumaview.com · © 2026