Introducing extended PET Quality Control (xQC): Fully automated evaluation of protocol compliance and quantifiability of individual datasets

Ziel/Aim With continuing technological and radiopharmaceutical progress in PET imaging as well as increased use of PET quantification, the demands on PET Quality Control (QC) are growing. However, actual PET QC guidelines focus on equipment, good practice and reference measurements outside clinical routine but do not include actual QC product inspection, i.e. a systematic validation of each of the resulting PET images. We introduce the term ‘extended PET quality control’ (xQC) for PET QC including fully automated QC Inspection of individual datasets and images. Our aim is Proof of Concept for xQC by an in-depth evaluation of PET data from a recent nationwide survey organized by the Swiss Association for Nuclear Medicine.

Methodik/Methods We approach the complexity of xQC by multi-paradigm software programming including multiple layers of abstraction up to an embedded rule based expert system allowing automated handling of diverse protocols and differing demands. The software was tested in a multi-center study by evaluating a total of 458 datasets originating from 18 different PET/CT systems at 14 Swiss sites. Each site provided a broad range of various image reconstructions from multiple acquisitions of two differing phantoms.

Ergebnisse/Results The xQC software successfully handled, evaluated and classifed the diverse datasets. Automated image inspection revealed a variety of violations of protocol or a mismatch of quantitative specifications and actual data in 57% of the data sets. Further 7% of the data sets were rejected as not evaluable because of serious violations of the study protocol. This, despite the fact that all sites and their equipment passed conventional PET QC. Only a fraction (36%) of the acquired PET datasets did actually meet the required standards.

Schlussfolgerungen/Conclusions The results vigorously prove the need, the feasibility and the benefits of automated PET xQC. The multi-paradigm software approach including an embedded expert system proves itself as enabling technology for xQC.