Clinical and/or scanner information on the tuning and testing dataset

Hello,

Thanks for the organisation of this challenge.

I found that on the training dataset, patient age and manufacturer information was given in the panorama_labels repository.

I was wondering clinical information such as patient age and possibly scanner information will be provided for the other phases of this challenge.

Thanks in advance for your answer.

Best, Alexandre

Hi Alexandre,

Thank you for your question. Yes the clinical information will be provided for all phases of the challenge. We are working on the best way to do this and will have an update for you in the next couple of days.

Let me know if you have any other questions.

Natalia

Hi Alexandre,

We have added the clinical information to the validation images using the Grand Challenge interface clinical-information-pancreatic-ct. The information is stored as a json for each case. An example of how to unpack this information can be found in our baseline algorithm: https://github.com/DIAGNijmegen/PANORAMA_baseline/blob/main/src/process.py (lines 88-94):

self.clinical_info_path = "/input/clinical-information-pancreatic-ct.json"
with open(self.clinical_info_path, 'r') as file:
            clinical_info = json.load(file)
print('Clinical Information:')
print('age:', clinical_info['age'])
print('sex:',clinical_info['sex'])
print('study date:',clinical_info['study_date'])
print('scanner:',clinical_info['scanner'])

Please let me know if there are any other questions.

Hi Natalia,

Thanks for you quick answers. The example usage is very clear. Should we handle cases where age could be missing? For instance, age is missing in clinical_information.xlsx for patients from public providers NIH / MSD.

Best, Alexandre

Hi Alexandre,

We have collected all clinical parameters for cases in the validation and testing data sets. However, we strongly advise submitted algorithms to be able to handle missing information as it increases their compatibility with other data sets and the practical value outside the context of our challenge.

Regards, Natalia Alves