<< Click to Display Table of Contents >> Navigation: PXMOD Data Processing > Overview |
Step-by-Step Processing
PXMOD data processing is based on a step-by-step approach, whereby each step is performed on a separate pane. Once a processing step has been performed, the user moves to the next step using a red action button located in the lower right workspace corner. To go back to a previous step, one of the prior panes can be selected by its tab. If the tasks for all steps have been configured appropriately, the button in the taskbar can be used for a complete re-processing.
The use of transient data is not supported. Rather, information pieces created during the processing session such as masks or volumes-of-interest (VOIs) need to be saved before they can be applied. Therefore, at the end of a processing session, it is always possible to save the whole processing configuration into a protocol file. The protocol allows to exactly recall all data processing elements for examination, or to repeat a processing with some changes.
Processing Overview
The processing of a data set in PXMOD consists of the following steps which are performed interactively:
Model selection |
The first step of a pixelwise processing is to select an adequate model from the model list. As a consequence, the user-interface elements are configured according to the model requirements. |
Image data loading |
As a next step the image data is defined and then loaded. During loading, image transformations such as smoothing may be applied. |
Image masking |
Optionally, a mask can be created interactively to restrict pixelwise processing to the area of interest. Masking mainly serves for removing background pixels which might result in disturbing outliers. The created mask needs to be saved and is automatically added to the protocol definition. If a mask is applied, threshold-based masking within the model is disregarded. |
VOI definition |
Often, dedicated time-activity curves (TACs) are required for the preprocessing and/or the pixelwise processing. In this case there is an optional step for outlining appropriate tissue VOIs interactively. The resulting VOI definitions or the related TACs need to be saved and are automatically added to the protocol definition. |
Blood preprocessing |
As a next step the tracer activity in arterial plasma (the input curve) is selected and loaded. It may require some corrections, for example to compensate for a delayed arrival and a dispersed shape at the site of activity measurement. For some models the whole-blood activity can also be supplied for blood spillover correction. The result of blood preprocessing is shown on a separate pane. For all subsequent processing steps the corrected blood curve is used. Note: Blood data is not required for reference models and some other models. In this case the blood-related panels are not active. |
Model preprocessing |
Depending on the model some preliminary calculations may be required, for example look-up tables or the specification of initial parameters for the pixelwise fits. These calculations are typically based on the TACs obtained in the VOI definition step. Usually the preprocessing results should be inspected to see that the model works properly with the prepared configuration. Therefore the results are shown on a separate panel. |
Pixelwise processing |
Once the preprocessing was successful, the user can specify which of the parametric maps are to be calculated, and whether they should be restricted to a certain (physiological) value range. If results are outside the restriction, a NaN ("not a number") is entered into the parametric map. For the rapid processing during a prototyping phase, e.g. for determining the adequate table look-up range, the pixelwise calculations can be limited to the current slice. |
Results explorations |
The results may be explored in many different ways such as: ▪Comparison of the different maps in the included fusion tool. ▪Pixelwise arithmetic among the maps in the included fusion tool. ▪Comparison of the pixelwise outcome with results when analyzing the TACs interactively in the kinetic modeling tool (PKIN, optional tool). |
Saving of the parametric maps |
Each fitted model parameter results in a separate functional map. These quantitative images can be saved in any of the different formats available. |
Saving of the protocol |
PXMOD only uses explicit information for the calculation. Therefore, at the end of a processing, a protocol file can be saved which will allow to exactly reproduce the processing. |
Working Mode using Initial Configuration
A suitable initial configuration can be used as an alternative to the step-wise processing outlined above. In this case all required data elements are specified beforehand with the button from the taskbar to the right, and then all processing steps executed with . This, however, is only possible if all the elements such as the VOIs are already existing. Therefore it is better applicable for repeating a processing with slightly changed parameters based on a protocol, rather than to the processing of a new data set.
How To Continue
The following sections describe the sequence of steps required for processing a data set.