Auto-PERCIST: Semi-automated response assessment of FDG-PET based on PERCIST 1.0 and other criteria.

Learning Objectives 1) To demonstrate the process and methods involved in the application of the PERCIST 1.0 criteria when assessing an imaging study. 2) To demonstrate that these methods can be automated in software, providing a reader-independent, objective and personalized quantitative assessment of disease and treatment response. 3) To demonstrate multiple assessment methods and techniques within the same application framework. 4) To demonstrate how a framework for an automated system can be extended to incorporate new and experimental methods, techniques and algorithms for PET image analysis.

The PERCIST 1.0 criteria describe a methodology for a personalized, quantitative assessment of an FDG-PET study, including response assessment when comparing follow-up to baseline studies. Our Auto-PERCIST™ program implements these methods in a semi-automated software application. It was developed to provide a quantitative and objective set of tools to allow the user to perform a PERCIST 1.0 analysis. It also supports multiple, user-selectable criteria for disease measurement and response assessment, permitting the user to compare methodologies from a single system. Program features include the automated extraction of key image criteria for both intra- and inter-study QC (i.e. radiotracer dosing, uptake duration, etc.); automatic identification and measurement of normal reference hepatic tissue for calculation of disease threshold values; the identification and characterization of disease objects through the use of user-selectable detectors (e.g. PEAK, MAX) and measurement statistics (e.g. PEAK, MAX, MEAN, TLG, Volume, Heterogeneity); and the automatic transmission of these results to an industry standard database where a variety of reporting templates have been developed which automatically compile, cross-reference and compare results providing immediate and objective reports of treatment response.

Research Support NCI P30CA006973 and NIH/NCI 1U01CA140204-01A2