> For the complete documentation index, see [llms.txt](https://medomicslab.gitbook.io/mediml-app-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://medomicslab.gitbook.io/mediml-app-docs/radiomics/data-processing/radiomics-pre-checks.md).

# Radiomics Pre-Checks

Radiomics pre-checks consists of identifying the optimal parameters for your radiomics study. While not mandatory for feature extraction, it is a recommended step.

{% hint style="success" %}
A tutorial video is available at the bottom of the page.
{% endhint %}

Since high variation in imaging parameters will impact the robustness of radiomic features, the following code will assist you in identifying optimal intensity ranges and voxel dimension re-scaling parameters for a specific dataset. For further insights, refer to the following [article](https://doi.org/10.1016/j.ejmp.2021.07.023), which examines how intensity windows affect the stability of radiomic features in CT data.

<figure><img src="https://314741292-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FWC0t7nEhhoOT73GRY714%2Fuploads%2F03pJYLEyE04XvK4XaE5O%2FDataManager-Pre-Checks.jpg?alt=media&amp;token=cbb763e8-a5cc-4179-957a-f7d231d14349" alt=""><figcaption><p><em>DataManager</em> Page - Radiomics Pre-checks section</p></figcaption></figure>

### How to Run Radiomics Pre-Checks

**Radiomics pre-checks** scan your dataset before extraction and report the **voxel dimensions** and **intensity** ranges across your scans. Use them to choose sensible extraction settings, such as voxel size and re-segmentation range.

#### Steps

1. **Use workspace data** *(optional)*: turn on to pick the dataset and ROI CSV from the current workspace instead of from your disk.
2. **Dataset format**: choose **DICOM**, **NIfTI** or **NPY**. NPY is the MEDscan output of **Process data**.
3. **Custom List of ROIs to Analyze** *(optional)*: a CSV listing the ROIs to check. Leave it empty to use the union of all ROIs.
4. **Dataset folder**: select the folder to check.
5. **Save results to**: select where the plots and JSON summaries will be written.
6. **Pre-checks options**: choose which scans to check (see below).
7. **Check types**: turn **Voxel checks (dimensions)** and/or **Window checks (intensity)** on. Both are on by default.
8. Click **RUN**, then **Show results** to view the plots.

#### Choosing which scans to check

{% tabs %}
{% tab title="Filters (NPY)" %}
When you select an NPY folder, the **studies**, **institutions** and **modalities** found in the file names fill the three lists. Pick one or more values in any list. Empty lists match everything.

{% hint style="info" %}
The lists are read from file names such as `STUDY-INSTITUTION-ID__scan.Modality.npy`. After **Process data**, its save folder is selected here automatically.
{% endhint %}
{% endtab %}

{% tab title="Custom wildcard" %}
Type a file-name pattern instead, for example `*CECT*` or `STS*CECT*.npy`. A custom wildcard **replaces** the filters. Use it for DICOM or NIfTI folders, where the lists stay empty.
{% endtab %}
{% endtabs %}

{% hint style="warning" %}
**RUN** stays disabled until at least one filter value or a custom wildcard is given and at least one check type is on.
{% endhint %}

### Tutorial video <a href="#tutorial-video" id="tutorial-video"></a>

{% embed url="<https://www.youtube.com/watch?ab_channel=MEDomicsLab&v=rkpRYdb18k4>" %}