annotation protocol about autopet HOT 6 CLOSED

Thank you for your question! This is indeed a more challenging case with liver metastases that have relatively low tracer uptake compared to the liver background. Still, these are metastases and should - in principle - be identified. Admittedly, this case might be challenging even for human experts.

from autopet.

Hi @sergiosgatidis ,

Thanks for your answer very much.

Would it be possible to provide more details on the annotation protocol? E.g., how do the radiologists identify the lesions from PET and CT, and how do they avoid FP and FN based on the images?

from autopet.

The annotation process relies on many years of experience in clinical hybrid imaging combined with relevant clinical knowledge. We will soon publish a technical paper describing the dataset and the annotation process in detail. Once published we will post it on the TCIA page.

from autopet.

Hi @sergiosgatidis ,

Would it be possible for you to share some general protocols? how do the radiologists identify the lesions from PET and CT, and how do they avoid FP and FN based on the images?

Since the testing submission deadline is very close, this information is very important for us to select postprocessing strategies.

from autopet.

A very good source that describes this process can be found here: https://cancerimagingjournal.biomedcentral.com/articles/10.1186/s40644-016-0091-3

from autopet.

Awesome!
Thanks:)

from autopet.