MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Testing and Characterization of An Automated, Cloud-Based Joint Commission Compliant PET Scanner Validation Program

J Sunderland1*, C Bauer1 , R Beichel1 , M Lodge2 , R Muzic3 , J Nye4 , P Christian5 , L Burrell5 , L Zimmermann3 , M Czachowski6 , P Wojtylak3 , (1) University of Iowa, Iowa City, IA,(2) Johns Hopkins University, Baltimore, MD, (3) Case Western Reserve University, Cleveland, OH, (4) Emory Univ, Atlanta, GA, (5) University of Utah, Salt Lake City, UT, (6) Children's Hospital of Pittsburgh of UPMC, Pittsburgh, PA

Presentations

(Tuesday, 7/31/2018) 11:00 AM - 12:15 PM

Room: Room 202

Purpose: To test robustness and performance characteristics of a fully-automated, cloud-based phantom analysis program designed to fulfill recent Joint Commission (JC) imaging requirements for the annual assessment of PET scanner performance.

Methods: The proposed JC-compliant program requires sites perform two phantom scans. One is a 20cm diameter uniform cylinder imaged to measure calibration accuracy, axial and radial uniformity, and axial and radial spatial resolution using an edge spread function (ESF)-based algorithm. The other is a 3D-printed anthropomorphic chest phantom (soon commercially available) with 12 spherical, 7 to 37 mm diameter lesions (or alternately the NEMA IQ phantom), imaged for contrast recovery assessment, lesion detectability, and clinical noise measurement. Images are reconstructed using the site’s clinical reconstruction. For analysis, data are uploaded to the cloud for archiving and fully-automated analysis including quality control checks, measurement of image features, and report generation against JC requirements. After review by a qualified physicist, the pdf report is emailed back to the clinical site or consulting physicist. Software robustness and efficiency were tested using data from PET/CT systems of range of vendors and vintage; using five clinically-relevant reconstructions, each, including PSF and regularized reconstructions; and using independent implementations of ESF analyses.

Results: Without exception software identified the correct phantom. All features were accurately located and segmented. Automated analyses results required only 1 to 2 minutes per data set, including report generation, and were consistent across implementations and with manual measurements for accuracy, resolution, uniformity, contrast recovery, contrast to noise, and low contrast image quality.

Conclusion: Fully automated, cloud-based phantom analysis software, designed as a cornerstone technology for a JC-compliant PET phantom quality control program, was developed and validated using a variety of scanner and reconstruction technologies in anticipation of imminent public availability.

Keywords

Phantoms, Quality Control, PET

Taxonomy

IM- PET : Quantitative imaging/analysis

Contact Email