Data Loading and Saving

The PMOD tools are able to read and write different types of image data. It is important to understand that correct data units and acquisition timing information is required for quantitative processing steps to derive meaningful results. As an example: In quantitative PET studies a tracer is injected at time 0. At the same time a series of image acquisitions is started, and arterial blood samples are withdrawn in certain intervals. The images are needed to monitor the time-course of tracer distribution in tissue. The blood samples are analyzed to determine the unchanged tracer in plasma, the input curve during the acquisition. The tissue response and the input curve are fed into kinetic models to quantify certain properties of the tissue under investigation. For applying the model, the image and blood units must be calibrated and the times must be measured in a common time scale.

In some image formats the data units are saved with the data (eg. DICOM, Ecat). With these formats the units are automatically detected upon loading into PMOD. Otherwise the user must select the correct units for the data and specify the acquisition timing. During loading, PMOD converts activity concentration data to its internal units, ie. kBq/cc and seconds. The appropriate scale factor is derived from the specified input units. As a result, PET images are always displayed in kBq/cc, and images are always saved with the displayed units. This has the consequence that if PET images residing in an Analyze format are loaded with units nCi/cc specified, and then again saved in Analyze, the values in the file are different, namely calibrated in kBq/cc. If this file is loaded, the units must therefore be specified as kBq/cc. These considerations do not apply if the image data are in DICOM. Therefore we highly recommend using the DICOM format, preferably in combination with the PMOD database.

Multiple transformations of the image data are supported during loading. The pre-processing capabilities include re-orientations such as orthogonal reslicing, rotations, mirroring, data processing such as smoothing or averaging over time, as well as the selection of subsets of the available slices and time frames.

Besides actual image data files PMOD is able to read and save a information needed for kinetic modeling from column-delimited text files (component data). These files may hold Volume-of-Interest definitions, kinetic modeling data, the configuration of the PMOD tools etc. They are either directly written as files into a directory, or to the database which saves the actual data into an internal directory and adds access information to the database tables. Component data reading/writing uses a unified interface which looks identical for the different component types.

It is helpful to remember that PMOD maintains three paths which are related to data loading/saving:

▪DICOM path: the directory which PMOD scans in order to look for DICOM part 10 data files.

▪Data path: the directory to load other image data types.

▪Database path: the directory where to look for component data.

PMOD maintains a separate history for each of the three paths to rapidly switch between frequently used locations. If databases are used to manage the data, access to all types of data can be done through the database user interface.

Starting with version 4.4 PMOD implements a unified data loader for the Scientific Data Management System (SDMS). It seamlessly integrates DATABASE, standalone DICOM data and AUTODETECT data loader for other formats within the same interface.