# Inducible displacement CT for implant loosening detection: a scoping review on methods, validation, and challenges

## TL;DR

This scoping review examines the use of inducible displacement CT for detecting implant loosening, highlighting methodological variations and challenges in validation.

## Contribution

The paper provides a comprehensive overview of ID-CT methods and challenges, emphasizing the need for standardized protocols and improved validation.

## Key findings

- ID-CT methods vary widely in loading protocols and CT acquisition parameters.
- Diagnostic validation is limited by inconsistent reference standards and incomplete reporting.
- Technical validation often lacks full pipeline testing, leaving measurement variance unaddressed.

## Abstract

Inducible displacement CT (ID-CT) is an emerging method for diagnosing implant loosening by (i) acquiring CT scans under different joint loading conditions, (ii) analyzing scans via segmentation and registration, and (iii) quantifying and visualizing relative implant–bone displacement. With multiple centers approaching these steps differently, this scoping review aimed to summarize current methodologies and key challenges.

PubMed, Cochrane, and Embase were searched for clinical and experimental ID-CT studies on spinal and arthroplasty implants. Data was extracted using a table based on updated CT-radiostereometric analysis (RSA) guidelines, including study characteristics, CT acquisition parameters, image analysis methods, validation approaches, outcomes, and loading protocols. Diagnostic studies were assessed with QUADAS-2.

22 studies were included concerning the hip (10), knee (7), spine (4), and wrist (1), covering clinical and experimental (phantom and cadaveric) designs. Loading protocols varied widely, applying compressive, rotational, and angular forces with external stabilization or loading devices. CT acquisition ranged from µCT to conventional and weight-bearing CT, with variation in kVp, mAs, reconstruction spacing, and metal artefact reduction. Image-analysis workflows were broadly similar, though reporting of algorithms and displacement metrics was often incomplete. Diagnostic validation was limited by variable reference standards, non-prespecified loosening thresholds, and non-blinded assessments, undermining true performance. Technical validation, although often precise, did not cover the full ID-CT pipeline (i.e., complete loading protocol), leaving key sources of measurement variance untested.

We showed that ID-CT is used with a wide variation in approach and limited reporting prevents the establishment of its true diagnostic accuracy.