Radiographs are collected in most arthroplasty studies, but collecting images is not the same as generating dependable imaging evidence. Differences in positioning, projection, equipment, timing and reader interpretation can make an apparently simple radiographic endpoint impossible to compare across sites.
CMO · Orthopedics & Spine
NHS orthopedic surgeon
in orthopedic surgery & Class III implant evidence
Chief Operating Officer · former TÜV SÜD Senior Reviewer · Senior Consultant Surgeon
in device evaluation & Notified Body review
Chief Data Officer · Biometrics & Data Systems
in clinical data management, EDC & biometrics
This page is for teams whose images must support a conclusion, not simply document the procedure.
The implant makes a claim that only imaging can substantiate.
The protocol includes endpoints derived from images.
Comparability between centers becomes the limiting factor.
Variation in positioning and projection is already present.
The sponsor needs assessment separated from site interpretation.
Findings that are notoriously inconsistently defined.
An early study where migration may be the central question.
Post-market evidence that has to withstand review.
The protocol says imaging will happen, without defining how.
A Notified Body or authority has challenged the imaging evidence.
→ Clinical evidence for arthroplasty medical devices — the parent hub. For the clinical side of the endpoint logic, see arthroplasty endpoints and PROMs.
Imaging strategy follows the same decision logic as endpoint strategy. The claim and the residual uncertainty define the imaging question; only then do modality, measurement method, timepoint and interpretation rule follow.
Is early fixation developing as expected?
Standard radiographs, RSA or CT-RSA.
Is cup position within the intended range?
Standardized pelvic radiographs or CT.
Is the planned component alignment achieved?
Long-leg and joint-specific radiographs.
Is fixation maintained over follow-up?
Standardized shoulder radiographs, or CT where justified.
Do fit and position match the preoperative plan?
CT-based planning and postoperative imaging.
How closely does achieved position match the plan?
System data combined with radiographic or CT confirmation.
As with clinical outcomes, precision of language is what prevents a weak imaging endpoint.
What is assessed: position, alignment, fixation, migration, radiolucency, osteolysis, loosening or implant integrity.
How it is assessed: an angle, a distance, a zone-based score, presence or absence, change over time, three-dimensional translation or rotation, or composite radiographic criteria.
A prespecified result, for example the change in femoral stem subsidence from baseline to twenty-four months.
The rule used to classify a finding as stable, progressive, clinically significant, indeterminate or unevaluable.
Key message. Radiographic assessment is not a complete endpoint. The view, the measurement, the reader, the timepoint and the interpretation rule must all be prespecified.
Arthroplasty imaging is not a single service. Three connected levels answer different questions at different levels of cost and complexity.
Images acquired at clinical sites under a common imaging charter. Suitable for routine assessment of position and alignment, monitoring of radiolucent lines, fracture or dislocation, implant integrity and standard clinical follow-up.
Central review by trained independent readers against prespecified criteria. Suitable when consistency across centers, blinded review, quantitative measurement, longitudinal comparison or regulatory-grade evidence is required.
Specialized measurement of three-dimensional implant migration relative to bone. Suitable for new fixation concepts, cementless implants, new coatings or porous structures, and small technically intensive early studies.
Traditional marker-based RSA uses simultaneous radiographs and a calibration cage to quantify implant or skeletal movement in three dimensions. CT-RSA is an evolving alternative that uses serial CT data rather than the conventional biplanar marker-based workflow.
Arthroplasty imaging is not a single service. Three connected levels answer different questions at different levels of cost and complexity.
The imaging charter is what turns an imaging intention into a repeatable process. It defines objectives, modality, required views and projections, patient positioning, side and anatomical coverage, equipment, acquisition parameters, calibration, the permitted acquisition window and the baseline definition — then file format, naming and de-identification, transfer, quality-control criteria, repeat-image rules, reading methodology, reader blinding, measurement definitions, adjudication rules, handling of missing or unevaluable images, and version control.
Small differences in positioning alter apparent alignment, component orientation, joint geometry and interface visibility. A technically correct but non-standardized image can be clinically useful and still be unsuitable for a quantitative endpoint. In practice, an illustrated acquisition manual is one of the highest-value documents in the study: it shows acceptable images, common positioning errors, correct landmark coverage, images requiring repeat acquisition, and images that remain clinically usable but are not endpoint-evaluable.
Important. The imaging charter should be finalized before site activation and reflected in the protocol, the CRF, the monitoring plan, the data-management plan and the statistical analysis plan.
The lists below are illustrative rather than prescriptive: the selection should follow the claim and the residual risk.
Acetabular inclination and anteversion, femoral stem alignment, femoral offset, leg-length difference, center of rotation, stem subsidence, cup migration, component seating, fracture, and component position in relation to dislocation.
Mechanical or anatomical limb alignment, femoral and tibial coronal alignment, femoral flexion, tibial slope, component rotation where CT is used, joint-line position, patellar position and tracking, component overhang, implant seating, planned versus achieved alignment.
Humeral component position, glenoid inclination and version, component seating, scapular notching, tuberosity healing, acromial or scapular spine fracture, radiolucency, migration or loosening, implant integrity.
Ankle component alignment, elbow loosening and implant integrity, small-joint implant migration or failure, and joint-specific radiographic classification systems where the device and its claims require them.
Particularly relevant for robotic arthroplasty, navigation-assisted implantation, patient-specific instruments and implants, 3D-printed components and preoperative CT-based planning systems.
Key message. Data generated by the surgical system should not automatically be treated as independent confirmation of that system’s accuracy.
Important clinical distinction. A radiolucent line is a radiographic finding, not automatically proof of osteolysis or mechanical loosening. Early lines may relate to implantation technique, cement penetration, component position or interface characteristics, and their meaning depends on timing, progression and accompanying findings. Do not use loosening, radiolucency, osteolysis and revision as interchangeable terms.
Multicenter imaging only works when every center is qualified, trained and monitored against the same charter. Equipment, DICOM capability and local radiation review differ market by market.
Imaging capability, calibration equipment and radiation approvals are assessed center by center during feasibility.
The strategy should distinguish new findings, persistent findings, progressive findings, resolved findings, findings present at baseline, findings appearing after an intercurrent event, and findings assessed after revision or reoperation.
Key message. RSA studies may need more frequent early assessments than routine radiographic PMCF studies. The schedule should follow the process being measured, not convention.
Independent reading means more than sending images to a radiologist. A core laboratory provides imaging-charter development, site and equipment qualification, acquisition training, central image receipt, de-identification checks, image-quality review and eligibility confirmation — then standardized measurements, independent reading, reader training and certification, longitudinal image matching, query management, adjudication support, blinded data export, reader-performance monitoring, and the final imaging dataset and report.
Differences in positioning, projection and equipment can make an apparently simple radiographic endpoint impossible to compare across sites. The charter is what prevents that.
Blinding deserves the same discipline. The protocol should state whether readers are blinded to site, timepoint and clinical outcome; whether sequential images are read together or separately; whether the implant type is concealed; who can see prior measurements; and when an image is reread.
Key message. The core lab should operate under a documented process that is separate from investigator interpretation and from sponsor commercial influence. Clinical expertise alone does not guarantee consistency — readers must be trained in the study-specific definitions.
Adjudication is appropriate when readers disagree on implant loosening, osteolysis, migration classification, component fracture, revision-related findings, endpoint evaluability, major safety findings, or whether a finding is new or progressive.
Important distinction. Adjudication should resolve defined disagreements. It should not be used to revise results because the initial findings are inconvenient.
It becomes valuable where early implant migration or fixation is a material evidence question.
A new cementless fixation concept, a new method of cementation, or a component for which early migration is a material residual risk.
A new porous coating, a new implant geometry, a shortened or bone-preserving stem, a new material or interface.
A new tibial baseplate or a new glenoid component whose early behavior is undocumented.
A first-in-human or phased introduction program where migration is the principal open question.
ISO 16087 addresses the RSA assessment of orthopedic implant migration, and the International Radiostereometry Society published updated recommendations covering both RSA and CT-RSA execution and reporting in 2024. See also first-in-human arthroplasty studies in DACH, where the migration question usually first arises.
Radiostereometric analysis measures three-dimensional implant migration relative to bone markers. Its value comes from the operational discipline around it.
A study-specific RSA plan, a marker or model strategy, site and equipment qualification, surgical marker insertion where required, a postoperative reference examination, scheduled follow-up, double examinations for precision, central analysis, migration reporting and integration with clinical outcomes. Operationally: tantalum bone markers, implant markers or validated models, biplanar acquisition, a calibration cage, a defined coordinate system, condition number, mean error of rigid body fitting, quality thresholds and exclusion rules.
Reduces the need to place markers on the implant by using a three-dimensional representation of the component. Before selecting it, confirm the availability of accurate CAD models, permission to use them, implant-size coverage, component visibility, model symmetry, software compatibility, validation of the analysis method, accuracy and precision for the specific implant, and the handling of modular components.
Considered where a markerless method is preferable, where CT is already in the clinical pathway, where component geometry suits segmentation, where conventional RSA infrastructure is unavailable, or where three-dimensional bone and implant relationships matter. Address CT protocol, dose, metal artifact, field of view, slice thickness, reconstruction, segmentation, registration, scanner consistency, cross-site calibration, software, precision and comparability with conventional RSA.
Key message. A model-based approach does not remove the need for method validation and study-specific precision assessment.
Important. The clinically relevant movement and axis depend on the implant and the fixation concept — a generic migration endpoint should not be copied across hip, knee and shoulder studies. Early migration may say something important about fixation behavior, but it should be read alongside the full device-specific evidence package, not as an isolated pass-or-fail result.
Precision and repeatability. Assessment should cover double examinations, their number and timing, patient repositioning, translation and rotation precision, measurement error, exclusion criteria, reader or analyst variability and site-level performance. The RSA standard and subsequent guidelines emphasize reporting study-specific precision rather than relying on values from other studies.
Available modality, equipment model, DICOM capability, technologist availability, calibration equipment, image-transfer capability, local radiation requirements, experience with research imaging, and the ability to repeat technically inadequate images.
The imaging manual, a positioning guide, a training session, test-image submission, approval before the first patient, and a contact route for acquisition questions.
Image-quality failure rate, missing and late images, positioning errors, protocol deviations, repeat-image rate, site query rate, transfer failures, and the proportion of images deemed unevaluable by readers.
Reconciliation of expected, received and missing images, queries, adjudications, final image status, the analysis dataset and archive requirements.
Delivered alongside site feasibility and selection, study start-up and on-site and remote monitoring.
Imaging generates two study records that both need to be traceable: the image, and the imaging result.
We can work with the sponsor’s existing imaging and EDC systems, or use an agreed integrated workflow suited to the study — including the eCRF platform where that is more proportionate.
Key message. The image and the imaging result are separate study records, and both must be traceable.
The statistical questions follow from the categories. Is the endpoint missing because of a technical problem, a missed visit or clinical deterioration? Can another modality be substituted? Can a repeat image be obtained? Does revision make subsequent imaging irrelevant? Is an unevaluable image treated as missing or as a failure? Which sensitivity analyses are needed?
Key message. The analysis plan should reflect how the image is acquired and read — statistical complexity cannot correct an inconsistent acquisition process. Equally, additional imaging must be proportionate to the evidence question and justified in the clinical investigation plan, with the biostatistics input agreed before enrollment.
Important. The CER should explain not only what the images showed, but how consistently and independently the findings were generated — a recurring theme in MDR clinical evaluation non-conformities.
Where standard-of-care imaging already answers the question.
Where a defined cohort and charter are needed post-market.
Where a registry cohort can carry an imaging layer.
Where images already exist but were never read consistently.
Where durability is the open question.
Where migration remains material after CE marking.
Also relevant after a device change, where imaging review confirms that the modification behaves as expected. See arthroplasty PMCF, registries and real-world evidence in DACH, orthopedic PMCF strategy, device registries under EU MDR and real-world evidence.
Key message. The PMCF imaging method should be linked to a specific residual uncertainty. Imaging should not be collected simply because it is routinely available.
Early protocol development. Output: imaging gap and recommendation report.
Multicenter clinical investigation. Output: full acquisition and reading charter.
Independent imaging assessment. Output: reader, QC and adjudication model.
New fixation technology. Output: RSA or CT-RSA recommendation.
An active study. Output: site qualification, training and QC.
Existing inconsistent data. Output: inventory, evaluability and recovery plan.
Completed or legacy studies. Output: critical appraisal and CER integration.
End-to-end support. Output: strategy through to the final imaging report.
Imaging questions are reviewed in the context of implant design, surgical technique, clinical outcomes and residual risk, under Dr Nikhil Khadabadi.
Protocol, imaging, data management, biostatistics, clinical reporting, PMCF and CER requirements are planned together.
Core-lab and adjudication models can be structured to separate image interpretation from site and sponsor influence.
Commission a focused imaging-charter review, an RSA feasibility assessment, or full multicenter imaging operations.
We can work with existing EDC, imaging-transfer and core-lab systems, or design an integrated workflow suited to the study.
Timelines follow the number of modalities, joints and implant configurations, any need for CT or RSA, the number of sites, core-lab setup, reader availability, imaging-charter maturity, test-image qualification, local radiation review, software validation, data-integration requirements and adjudication complexity. Costs follow the number of expected images and timepoints, image type and file size, site training, the reading model, the number of readers, quantitative measurements, adjudication frequency, RSA cages, markers and CT-RSA segmentation, image storage and transfer, and the duration of long-term follow-up. We quote in defined work packages, on the same basis as our European clinical trial cost benchmarking.
Related Eclevar pages across the orthopedic program.
Component position and alignment, radiolucent lines, subsidence or migration, osteolysis, loosening, implant integrity and fracture. The selection should follow the claim and the residual risk rather than convention.
When consistency across sites, blinded review or quantitative measurement materially affects the endpoint — particularly when the imaging evidence will support a regulatory conclusion.
The imaging objectives, modality, views, positioning, equipment and acquisition parameters, the baseline definition, transfer and de-identification, quality control, the reading methodology, measurement definitions, adjudication rules and version control.
No. A radiolucent line is a radiographic finding. Its meaning depends on the zone, width, timing, progression and the accompanying clinical and radiographic findings.
When early migration or fixation is a material residual uncertainty — a new cementless design, coating, geometry or fixation concept — and when suitable centers and a clear interpretation plan are available.
Conventional RSA uses markers and biplanar radiographs with a calibration cage. Model-based RSA replaces implant markers with a validated three-dimensional model of the component. CT-RSA uses serial CT data instead of the biplanar marker-based workflow. Each requires its own validation and study-specific precision assessment.
Yes, provided the participating centers have the necessary equipment, training and qualification. Because RSA is technically demanding, it is often delivered through a small number of specialist centers within a larger study.
By defining the categories in advance, distinguishing unevaluable from missing, setting repeat-image rules, and prespecifying the primary assumption and the sensitivity analyses in the SAP.
Yes, and it is often valuable. Imaging and PROMs answer different questions, so linking them supports a fuller interpretation, provided the data flow and reconciliation are planned in advance.
Each imaging endpoint should trace from the claim or residual risk through the charter, acquisition, independent assessment and analysis, into the clinical study report and, where applicable, the PMCF evaluation report and the CER.
Yes. We can work with existing imaging, transfer and EDC systems, review an existing core-lab model, or design an integrated workflow where that is more proportionate.
Sometimes. It depends on acquisition consistency, evaluability, the availability of a usable baseline and the specific residual uncertainty. An imaging inventory and evaluability review will usually determine whether the data are suitable before a study is designed around them.
Share your device description, intended claims, study design and proposed imaging schedule. We review the imaging questions, modalities, acquisition standards, core-lab requirements and whether RSA or CT-RSA is justified.
Your documents are reviewed confidentially. An NDA can be put in place before we receive any technical or clinical information. You can also reach the team through the contact page.
