Prosthetic Loosening

AI for reporting of prosthetic loosening.

Modality:

SPECT/CT

Pathology:

Prosthetic loosening – mostly hip and knees, occasional elbow and shoulder.

Status:

Concluded


CSC Lead: Dika


This project has been withdrawn. An exploratory data analysis demonstrated that the data for this project is too complex, too varied and too few to proceed at this point in time. This project may be revisited at a later date but it is not currently being actively worked on.


Bone scan and SPECT CT are effective ways of assessing loosening as the metal artefact reduces the clarity of other imaging modalities. Patients will also be referred to x-ray and CT, but SPECT/CT supersedes these for accuracy.

Clinical lead: Sanjay Vijayanathan

Rationale

Prosthetic loosening is definitively diagnosed with a SPECT/CT. Around 40% of cases are not obvious and in 20% of cases the report is changed with further consultation. An AI tool would help in accuracy and increase confidence in reporting.

Patient pathway

Patients are referred as a result of 1 – pain in the replacement site and 2 – x-ray investigations suggest loosening. Patients are referred from orthopaedics and surgery to consider replacement of a prosthesis or to revise the replacement. Next steps are decided by SPECT/CT reports. SPECT/CT is also used for surveillance so imaging is repeated in intervals.

Training data

250 SPECT/CT scans identified. Data to be gathered retrospectively.

Risks

When the diagnosis of prosthetic loosening is missed, the patient remains in chronic (severe) pain for a prolonged amount of time. When the diagnosis of prosthetic loosening is incorrect, patient undergoes unnecessary surgery.

Goals

Reducing patient pain and avoiding unnecessary surgery. Improving diagnosis speed and accuracy.

Success criteria

Increased diagnostic accuracy and thus reduction in unnecessary surgery, increase in speed with which loosening is diagnosed.

Alternatives

Currently no commercial products identified.