Integrated image analysis solutions for PET datasets in damaged brain

Abstract

Objective. To identify and discuss the problems inherent in the processing of multiparametric functional imaging datasets from patients with acute brain injury, using "triple oxygen" positron emission tomography (PET) as an example. To present an integrated imaging solution for analysis of such datasets and report on its use in practice. Methods. Structured analysis of the steps in image analysis for triple oxygen studies in patients with acute brain injury, subarachnoid haemorrhage and carotid artey disease was undertaken. Systematic identification of the drawbacks of conventional manual processing techniques, which make use of modules from different commercially available software packages was carried out. This analysis informed the construction of an image analysis gateway that addressed these issues. Results. We describe an integrated image analysis gateway (PETAn) that provides for the automated integration of all processing steps, along with outputs that detail analysis stages, error logs and results. Implementation of PETAn allowed batch processing of clinical PET data sets, reduced the level of technical expertise required for analysis, improved the quality of results and achieved significant reductions in operator analysis time. Conclusions. While modules from several imaging software suites can be usefully combined to allow analysis of multi-parametric PET data sets from patients with acute brain injury, sequential use of these modules for various steps in image analysis is fraught with difficulty. Careful automation and integration of these steps greatly facilitates the interrogation of information-rich data sets and increases research productivity.