Skip to content

Pancreas Cancer AI-Driven Diagnosis in CT scan and EUS Imaging: A protocol for an Observational Multicentric Ambispective Study

Notifications You must be signed in to change notification settings

Sdamirsa/PanCanAID

Repository files navigation

About PanCanAID

This is a 2024 updated version of the PanCanAID project repository. PanCanAID is a philanthropic initiative that strives to enhance the early detection, diagnosis, and treatment of pancreatic cancer patients. We are committed to broadening our collaboration with teams and institutes who share our vision of improving the diagnosis and prognosis of pancreatic cancer. Our ultimate goal is to develop or aid in developing machine learning-driven tools with a minimum cost for the patients. Join us in our mission to combat pancreatic cancer and positively impact people’s lives.

PanCanAID Current Aim

PanCanAID's current focus is improving the accuracy of CT scan imaging and helping radiologists diagnose pancreatic cancer. We proudly announce that six leading institutes are involved in the PanCanAID data pipeline, providing us with a robust and diverse dataset. Our multidisciplinary team comprises experts from 10 institutes in Iran and the USA, including radiologists, oncologists, computer scientists, and data analysts. Join us in our efforts to make a real impact on the lives of those affected by this disease.

PanCanAID Vision

We aim to collect the most comprehensive imaging, laboratory, and clinical dataset, as well as detailed treatment timelines and pancreatic cancer outcomes. We hope that this can facilitate ML-based solutions for this cancer.

PanCanAID Licence

We provide the opportunity to all researchers worldwide and not-for-profit organizations to use this dataset and models as long as they get a minimum free from the patients, covering the inference and maintenance costs. Commercial use of the PanCanAID dataset and models is possible, mandating the commercial use to allocate 10% of the revenue of this project to charity porpuses of pancreatic cancer, including prevention of pancreatic cancer, community engagement to prevent pancreatic cancer, pancreatic cancer research, and pancreatic cancer treatment cost. The cost of supervision of this mandate should also be covered by the interested agencies. We acknowledge the challenge of governing bodies to adhere to this mandate, but since now, we couldn't find any other way that lies within our patients' interest, as we are beholding their entrusted data, and they mandated this rule, which is promoted by the PanCanAID co-founder.

PanCanAID Team and Centers

PanCanAID Charity Sponsors and Acknowledgment

  • Hostiran: Covered 1-year cost of the cloud platform for segmentation (COI: Hostiran doesn't provide any medical services and is not involved in any process, including design, collection, development, or deployment)
  • XNAT: We used their service to maintain a segmentation platform for our 15 segmentors.

PanCanAID Output

PanCanAID Code

Code for RE-USE: extracting dicom meta, assigning pseudonymized numbers, renaming folders, and anonymizing dicom

I prepared a no-code Python script that you can run from the command line. Please go to the folder named Step1_SortingFiles for a detailed guide.

It will guide you step by step to: 1- extract the dicom meta and store it in an Excel and JSON file 2- Assing desired pseudonymized numbers to folder names, and then add this code to the previously created Excel file (optional) 3- It will comprehensively anonymize dicom metadata (just necessary dicom meta will remain), and assign folder names as patient names and id.

That's it. Remember that for using it you should locate each dicom study (which can contain many files and series) in one folder in your directory. However, it can identify dicom studies in one folder of the directory (I am using study_id, study_date, and folder name to find unique studies). I wish you all luck in your projects, and since I was almost dead while handling so many tasks from so many centres I created this code for you and myself :) cheers

Code for RE-USE: Reading segmentation dicom files generated by OHIF or XNAT platforms + find the corresponding slices

loading...

PanCanAID Model

loading...

PanCanAID Paper

Protocol Paper

PanCanAID Data

loading...

PanCanAID On-boarding

What is PanCanAID?

How to join PanCanAID + onboarding

Follow the instruction bellow to understand what is PanCanAID and how you can collaborate. If you decided to join and help us, follow the instructions to understand how you should finish you contribution.

  • As a medical center

  • As a pancreatic cancer physician

  • As a radiologist

    1. What is PanCanAID

    2. How you can contribute and what we can provide to compensate part of your efforts

      After joining PanCanAID, follow this instruction describing the instructional pathway for you at PanCanAID Education:

  • As a GP

    1. What is PanCanAID

    2. How you can contribute and what we can provide to compensate part of your efforts

      After joining PanCanAID, follow this instruction describing the instructional pathway for you at PanCanAID Education:

  • As a (bio)medical student

    • to help with data collection
    • to help with patients' follow-up interview
  • As a pancreatic cancer patient

Please visit PanCanAID Education to see the all-in-one board.

About

Pancreas Cancer AI-Driven Diagnosis in CT scan and EUS Imaging: A protocol for an Observational Multicentric Ambispective Study

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published