ADC-HBV spacing difference

Hello,

As we know that ADCs are derived from DWIs, thats why we expect spacing of them to be equal. In case "10009" slice thicknesses are different, is that normal? As a second question, original annotations were done on T2W or DWI. How to know which modality was used in each case for annotation ? (could be needed for proper registration/resampling of label map)

Thank you Ahmet Karagöz @HeviAI

Hi Ahmet,

Regarding the disparity in spatial resolution between the ADC and DWI/HBV scans of patient 1009, you're right. Well spotted! In-plane resolution seems identical, down to every decimal. But the through-plane resolution is off by 1 mm. We'll double-check this issue at our end and revert back to you, as soon as possible.

Regarding which modality was used for original annotations and which modality do they map to, please take a look at the ('Annotations' section of the 'Datasets: Imaging, Labels' page](https://pi-cai.grand-challenge.org/DATA/), which states the following:

Original Annotations (picai_labels/csPCa_lesion_delineations/human_expert/original): All axial bpMRI sequences (T2W, DWI, ADC) per case, were used to localize and annotate csPCa lesions. However, depending on the annotator/center and their preference, some annotations have been created at the spatial resolution and orientation of the T2W image, while others have been created at the resolution and orientation of the DWI/ADC images. Either way, for every annotation in this folder, all lesion delineations (if any) will always clearly map to observations in DWI/ADC imaging.

Hope this helps.

Hi,

Thank you, it helps! I also need to mention that cases "10114, 10238, 10320, 10362, 10447, 10493, 10557, 10573, 10591, 10605" have the same problem (if it is a problem), there could be more cases i am not sure.

Thanks again.

Hi Ahmet,

We have investigated the issue and found the root cause. I cannot confirm whether this applies exactly to the PI-CAI case numbers you provided due to the anonymisation process, but I'm certain at least most cases I found match.

The discordance was caused because the high b-value was acquired rather than calculated. As a result, the (through-plane) resolution is not necesarily the same between the apparent diffusion coefficient map (calculated based on the non-high b-values <= 800 for these cases), and the aquired high-bvalue scan (b-value = 2000 for these cases).

So, while it may not be the preferred way, this is what was used in clinical practice.

Hope this helps, Joeran

Hi Joeran,

Thank you so much, it helps. :)

Ahmet