Test phase of Task 1

Will we have registered pre-RT image during testing for Task 1?

Hello,

Thank you for your question.

Task 1 is the pre-RT segmentation task so you will only be given the original pre-RT image as input to your model. Providing the pre-RT segmentation during testing would be giving the answers to the test set so naturally we want to avoid that :)

I believe maybe you instead meant to ask about Task 2 (mid-RT segmentation) in your question. If so, then yes, you will be provided the registered pre-RT image and registered pre-RT segmentation as input to your model (these are registered to the mid-RT image). In total, for Task 2 you will be provided the following 5 input files during testing: original pre-RT image, original pre-RT segmentation, registered pre-RT image, registered pre-RT segmentation, and mid-RT image. The original pre-RT image and segmentation are provided in addition to the registered ones in case registration is part of your pipeline. In your inference script, you can specify exactly which inputs you want to use for your model.

We briefly go through an example using some data on the Zenodo download page (https://zenodo.org/records/11199559): "The goal of Task 2 is to generate a mid-RT segmentation mask (e.g., "2_midRT_mask.nii.gz" is the relevant label). During blind testing for Task 2, the mid-RT MRI (e.g., "2_midRT_T2.nii.gz"), original pre-RT MRI (e.g., "2_preRT_T2.nii.gz"), original pre-RT tumor segmentation mask (e.g., "2_preRT_mask.nii.gz"), registered pre-RT MRI (e.g., "2_preRT_T2_registered.nii.gz"), and registered pre-RT tumor segmentation mask (e.g., "2_preRT_mask_registered.nii.gz") will be provided to the participants algorithms. "

Hope this is helpful. Happy to answer any follow-up questions as needed.

-Kareem

Thank you very much for your answer. I am deeply sorry I asked my question in a really bad (wrong) way, Of course we don't have the segmentation. :) what I meant is that, during the test phase of Task 1, do we have access to both the registered preRT images and the original preRT images, or only the original images? Based on your previous answer, I can tell that we only have the original images, but I would appreciate your confirmation.

Ah yes, I see now. No worries! Yes, for Task 1 you will only have access to the original pre-RT images. The registered pre-RT images will only be provided for Task 2.

Best,

-Kareem

Edit: We have added a few figures to visually illustrate model inputs on the Tasks and Evaluations page. Hopefully this will add further clarity for participants.