search

rcuocolo / PROSTATEx_masks

Lesion and prostate masks for the PROSTATEx training dataset, after a lesion-by-lesion quality check.
https://rcuocolo.github.io/PROSTATEx_masks/
Creative Commons Attribution 4.0 International
78 stars 17 forks source link

Unmatched size of prostate masks #10

Closed hiyuhan closed 3 years ago

hiyuhan commented 3 years ago

Hi, very nice work. I found there are some unmatched size between 'mask_prostate', 'mask_pz', 'mask_tz' and raw T2 dicoms from ProstateX challenge. Would you please check? Part of the files is listed as following: error size: ProstateX-0102 pz error size: ProstateX-0102 tz error size: ProstateX-0141 pz error size: ProstateX-0141 tz error size: ProstateX-0160 pz error size: ProstateX-0160 tz error size: ProstateX-0165 pz error size: ProstateX-0165 tz error size: ProstateX-0166 pz error size: ProstateX-0166 tz error size: ProstateX-0168 pz error size: ProstateX-0168 tz error size: ProstateX-0173 pz error size: ProstateX-0173 tz error size: ProstateX-0181 pz error size: ProstateX-0181 tz error size: ProstateX-0191 pz error size: ProstateX-0191 tz error size: ProstateX-0197 pz error size: ProstateX-0197 tz error size: ProstateX-0199 pz error size: ProstateX-0199 tz error size: ProstateX-0201 pz error size: ProstateX-0201 tz error size: ProstateX-0202 pz error size: ProstateX-0202 tz error size: ProstateX-0203 pz error size: ProstateX-0203 tz

rcuocolo commented 3 years ago

I have checked the first 2 cases (as a sample). The issue is simply that multiple axial T2 sequences are available, and the anatomical masks in these were not taken on the same images as the lesion masks. This is already reflected in the image list CSV files in each folder. These specify the correct T2 sequence to use for each mask (lesion or anatomical). I understand this is not ideal, but the two processes were conducted in separate occasions and different files were employed. Please check you are referencing the correct image list for the mask set.

hiyuhan commented 3 years ago

Thanks for the fast reply. I have checked the files and you are right. May I ask how to choose between the different T2 images in a single patient scan, like for zonal segmentation. Or I am confused on why there can be two similar T2s.

rcuocolo commented 3 years ago

The sequences were repeated in the dataset for unknown reasons. For us, the choice was essentially due to chance. In some cases there might be artifacts (e.g., patient movement) that require a new acquisition of a series, but the images used for the segmentation process were all considered viable for the task at hand. In one of the papers we report the exclusion of 1 case for artifacts on the ADC map (for the purpose of lesion segmentation, T2 images were fine for gland/zonal segmentation).

hiyuhan commented 3 years ago

Very clear. I shall check your papers. Thanks a lot.